Peter H Bartels

Interests

No activities entered.

Courses

No activities entered.

Scholarly Contributions

Journals/Publications

• Bartels, P. H., & Bartels, H. G. (2013). Classification in karyometry performance testing and prediction error. Analytical and Quantitative Cytology and Histology, 35(4), 181-188.
  More info

  PMID: 24341120;Abstract: Classification plays a central role in quantitative histopathology. Success is expressed in terms of the accuracy of prediction for the classification of future data points and an estimate of the prediction error. The prediction error is affected by the chosen procedure, e.g., the use of a training set of data points, a validation set, an independent test set, the sample size and the learning curve of the classification algorithm. For small samples procedures such as the "jackknife," the "leave one out" and the "bootstrap" are recommended in order to arrive at an unbiased estimate of the true prediction error. All of the procedures rest on the assumption that the data set used to derive a classification rule is representative for the diagnostic categories involved. It is this assumption that in quantitative histopathology has to be carefully verified before a clinically generally valid classification procedure can be claimed. © Science Printers and Publishers, Inc.
• Montironi, R., Bartels, P. H., DeCensi, A., Puntoni, M., Hurle, R., Decobelli, O., Carmignani, G., Mazzucchelli, R., Bartels, H. G., Alberts, D. S., & Maffezzini, M. (2013). A randomized phase IIb presurgical study of finasteride vs. low-dose flutamide vs. placebo in men with prostate cancer. Efficacy monitored by karyometry. Urologic Oncology: Seminars and Original Investigations, 31(5), 557-565.
  More info

  PMID: 21783387;Abstract: Objective: Presurgical, window of opportunity trials have been proposed as a model to assess the activity of preventive and therapeutic interventions in a cost-effective manner in prostate cancer (CaP). The aim of the study was to explore karyometry as a method for monitoring the efficacy of intervention with preventive agents in patients with CaP. Materials and methods: The material used in this investigation was from the 2F study, i.e., an Italian prospective randomized phase IIb presurgical study of finasteride vs. low-dose flutamide vs. placebo in men with CaP. Image analysis was performed in 16 cases treated with finasteride, 24 with flutamide, and 20 with placebo. For all these cases, CaP and normal looking secretory epithelium were present in the pretreatment biopsies as well as the post-treatment ex-vivo biopsies obtained from the radical prostatectomy specimens. Results: To establish a direction of nuclear change from normal to malignancy, i.e., the so-called line of progression, a discriminant function was derived with the normal looking epithelium in the pretreatment biopsies as one endpoint, and the CaP in the pretreatment biopsies as the other. The discriminant function was then applied to the post-treatment groups. The increase in relative nuclear area was the dominant feature. In the placebo group, 15 out of 20 CaP (75%) cases had a higher discriminant function score at the end of study, with a significant increase of the mean score by 90%. The flutamide treated CaP cases had increased discriminant function scores in 19 out of 24 cases (79%) and an increase of the mean score by 43%; the 5 cases with lower scores involved only minor reductions. In contrast, the finasteride treated CaP cases had increased discriminant function scores for 8 out of 16 cases (50%), but the increase in the mean score was by only 8%. Conclusion: This exploratory study establishes that karyometric monitoring can track the results of subtle nuclear changes induced by preventive interventions in men with CaP, thus allowing assessment of agent activity in a cost-effective manner. © 2013 Elsevier Inc.
• Bartels, P. H., Bartels, H. G., Alberts, D. S., Yozwiak, M., Prasad, A. R., Glazer, E. S., & Krouse, R. S. (2012). Karyometry of nuclear phenotypes in cutaneous squamous cell cancer. Analytical and Quantitative Cytology and Histology, 34(1), 1-8.
  More info

  PMID: 22590813;Abstract: Objective: To establish the karyometric characteristics of the two main nuclear phenotypes in cutaneous squamous cell cancer (cSCC) lesions. Study Design: The clinical materials comprised 75 cases of cSCC, 38 with aggressive lesions and 37 with nonaggressive lesions. High-resolution images of 100 nuclei per case were recorded. Data were partitioned into four subgroups covering the range of lesion progression. Four discriminant functions were derived to distinguish aggressive from nonaggressive lesions. The most typical nuclei from the phenotype predominant in aggressive lesions and nonaggressive lesions were separated out by thresholding on the discriminant function score axes. For these homogeneous sets of nuclei the karyometric features were computed. Results: The nuclear populations in cSCC lesions are a very heterogeneous set. There are two axes of dispersion, along the line of lesion progression and between aggressive and nonaggressive lesions. The analysis faces the difficulty that lesions from both diagnostic categories contain nuclei of the same two phenotypes with the difference between categories consisting only of differences in proportion of the two phenotypes. Conclusion: The nuclei of the aggressive phenotype I and nonaggressive phenotype II have substantially different chromatin patterns and can be distinguished with > 90% correct recognition rate.
• Bartels, P. H., Garcia, F. A., Trimble, C. L., Kauderer, J., Curtin, J., Lim, P. C., Hess, L. M., Silverberg, S., Zaino, R. J., Yozwiak, M., Bartels, H. G., & Alberts, D. S. (2012). Karyometry in atypical endometrial hyperplasia: A Gynecologic Oncology Group study. Gynecologic Oncology, 125(1), 129-135.
  More info

  PMID: 22155796;Abstract: Objectives: Treatment for atypical endometrial hyperplasia (AEH) is based on pathologic diagnosis. About 40% of AEH is found to be carcinoma at surgery. This study's objective is to derive an objective characterization of nuclei from cases diagnosed as AEH or superficially invasive endometrial cancer (SIEC). Methods: Cases from GOG study 167A were classified by a central pathology committee as AEH (n = 39) or SIEC (n = 39). High resolution digitized images of cell nuclei were recorded. Features of the nuclear chromatin pattern were computed. Classification rules were derived by discriminant analysis. Results: Nuclei from cases of AEH and SIEC occupy the same range on a progression curve for endometrial lesions. Cases of AEH and SIEC both comprise nuclei of two phenotypes: hyperplastic characteristics and premalignant/neoplastic characteristics. The principal difference between AEH and SIEC is the percentage of premalignant/neoplastic nuclei. When this percentage approaches 50-60% superficial invasion is likely. SIEC may develop already from lesions at the low end of the progression curve. Conclusions: AEH comprises cases which may constitute a low risk group involving < 40% of AEH cases. These cases hold a percentage of < 20% of nuclei of a preneoplastic phenotype. AEH cases from the central and high end of progression have > 40% of nuclei of preneoplastic phenotype. Nuclei of the preneoplastic phenotype in AEH lesions are almost indistinguishable from nuclei in SIEC, where this percentage exceeds 60%. The percentage of nuclei of the preneoplastic phenotype in AEH esions might serve as criterion for assessment of risk for the development of invasive disease. © 2012 Elsevier Inc. All rights reserved.
• Lopez-Beltran, A., & Bartels, P. H. (2012). Update of Journal Name. Analytical and Quantitative Cytology and Histology, 34(4), 171-172.
• Glazer, E. S., Bartels, P. H., Prasad, A. R., Yozwiak, M. L., Bartels, H. G., Einspahr, J. G., Alberts, D. S., & Krouse, R. S. (2011). Nuclear morphometry identifies a distinct aggressive cellular phenotype in cutaneous squamous cell carcinoma. Cancer Prevention Research, 4(11), 1770-1777.
  More info

  PMID: 21636541;PMCID: PMC3181389;Abstract: By identifying aggressive cutaneous squamous cell carcinoma (cSCC) in patients who are at high risk for recurrences or second primaries after resection, intensive surveillance and therapy may decrease morbidity and mortality. We investigated the role of nuclear morphometry (karyometry) in differentiating between aggressive and nonaggressive cSCC. We retrospectively analyzed cSCC lesions from 40 male patients. Twenty-two patients had evidence of aggressive cSCC (local/regional recurrence or a second primary cSCC), and 18 patients were identified with similar ages and sites of disease as control patients with nonaggressive cSCC (no evidence of recurrence, metastasis, or second primary). We carried out karyometric analysis to identify nuclear features that discriminate between aggressive and nonaggressive cSCC nuclei. We used statistically significant differences (Kruskal-Wallis test, P < 0.0001) to compose a quantitative aggressive classification score (proportion of aggressive nuclei from 0% to 100%). For comparisons, we used Fisher's exact test or Student's t test. The mean age was 79 ± 7 years for aggressive cSCC and 80 ± 9 years for nonaggressive cSCC (P = 0.66). We analyzed a mean of 96 nuclei in each group. The mean classification score for aggressive cSCC was significantly higher (69% ± 6%) than for nonaggressive cSCC (28% ± 5%, P = 0.00002). Overall, the classification score accurately categorized 80% of our patients (P = 0.0004). In most patients, karyometry differentiated between aggressive and nonaggressive cSCC. We found that classification scores, which provide information on individual lesions, could be used for risk stratification. ©2011 AACR.
• Mazzucchelli, R., Scarpelli, M., Lopez-Beltran, A., Cheng, L., Bartels, H., Bartels, P. H., Alberts, D. S., & Montironi, R. (2011). Global acetylation and methylation changes predict papillary urothelial neoplasia of low malignant potential recurrence: A quantitative analysis. International Journal of Immunopathology and Pharmacology, 24(2), 489-497.
  More info

  PMID: 21658323;Abstract: Papillary urothelial neoplasia of low malignant potential (PUNLMP) recurs in approximately 35% of patients. Conventional histopathological assessment does not distinguish non-recurrent from recurrent PUNLMP. The aim of this study is to explore the differences in global histone acetylation and global DNA methylation between non-recurrent and recurrent PUNLMP. Acetylated histone H3 lysine 9 (AcH3K9) and 5-methylcytosine (5MeC) were investigated by immunohistochemistry (IHC) in 20 PUNLMP cases (10 non-recurrent and 10 recurrent), in 5 cases of normal urothelium (NU) and in 5 cases of muscle invasive pT2 urothelial carcinoma (UC). The total optical density of the nuclear staining was measured photometrically in at least 40 nuclei separately for the basal, intermediate and luminal positions in each case. Concerning the total optical density values for both acetylation and methylation, a decrease in staining is observed from non-recurrent PUNLMP to recurrent PUNLMP, at all nuclear locations. For acetylation the mean value in non-recurrent PUNLMP, intermediate between NU and UC, is closer to the former than to latter. The mean value in recurrent PUNLMP is closer to UC than to NU. In NU, non-recurrent and recurrent PUNLMP, the acetylation to methylation ratio decreased from the nuclei in basal position to those in the surface, the average for the above groups being 1.491, 1.611 and 1.746, respectively. Setting the observed values for NU at each sampling location to unity, acetylation shows a steady decrease, the percentages of changes in this nuclear location compared to NU being -5% in non-recurrent PUNLMP, -15% in recurrent PUNLMP and -24% in UC. Concerning methylation, there is a slight increase in non-recurrent PUNLMP (+5%), a decrease in recurrent PUNLMP (-19%) followed by a sharp rise for the UC (+61%). In conclusion, there are differences in global histone acetylation and DNA methylation patterns between non-recurrent and recurrent PUNLMP. Further studies are needed to elucidate the complex interplay between chromatin structure, its modifications and recurrence of PUNLMP. Copyright © by BIOLIFE, s.a.s.
• Stratton, S. P., Alberts, D. S., Einspahr, J. G., Sagerman, P. M., Warneke, J. A., Curiel-Lewandrowski, C., Myrdal, P. B., Karlage, K. L., Nickoloff, B. J., Brooks, C., Saboda, K., Yozwiak, M. L., Krutzsch, M. F., Chengcheng, H. u., Lluria-Prevatt, M., Dong, Z., Bowden, G. T., & Bartels, P. H. (2010). A phase 2a study of topical perillyl alcohol cream for chemoprevention of skin cancer. Cancer Prevention Research, 3(2), 160-169.
  More info

  PMID: 20103724;PMCID: PMC3270887;Abstract: The chemopreventive and antitumor properties of perillyl alcohol (POH) that were studied preclinically indicate that topical POH inhibits both UVB-induced murine skin carcinogenesis (squamous cell tumor models) and 7,12-dimethylbenz(a) anthracene-induced murine melanoma (transgenic models involving tyrosinase-driven Ras). A previous phase 1 clinical trial in participants with normal-appearing skin showed that topical POH cream was well tolerated at a dose of 0.76% (w/w). Here, we performed a 3-month, double-blind, randomized, placebo-controlled phase 2a trial of two different doses of topical POH in individuals with sun-damaged skin. Participants applied POH cream twice daily to each dorsal forearm. Baseline and end-of-study biopsies were taken from each participant to evaluate whether the topical application of POH was effective in reversing actinic damage as evidenced by normalization of quantitative skin histopathologic scores and change in nuclear chromatin pattern as measured by karyometric analysis. There was a borderline reduction in the histopathologic score of the lower-dose POH group compared with the placebo (P = 0.1), but this was not observed in the high-dose group. However, in the high-dose group, a statistically significant reduction in the proportion of nuclei deviating from normal was observed by the use of karyometric analysis (P < 0.01). There was no statistical significance shown in the lower-dose group. No changes were observed in p53 expression, cellular proliferation (by proliferating cell nuclear antigen expression), or apoptosis in either treatment group compared with the placebo group. These results suggest that whereas our karyometric analyses can detect a modest effect of POH in sun-damaged skin, improved delivery into the epidermis may be necessary. Cancer Prev Res; 3(2); 160-9. ©2010 AACR. ©2010 American Association for Cancer Research.
• Bartels, P. H., & Bartels, H. G. (2009). Tutorial article discriminant analysis. Analytical and Quantitative Cytology and Histology, 31(5), 247-254.
  More info

  PMID: 20701090;Abstract: OBJECTIVE: To present a fully worked numerical example for the derivation of a discriminant function in order to provide insight into the processing steps and origin of the function coefficients. STUDY DESIGN: The example begins with the reduction of a set of raw data to the values needed to calculate the variance/covariance matrix. Next the inversion of the covariance matrix by pivotal condensation is carried through. This is followed by the calculation of the coefficients. All calculations are carried out on a simple hand calculator. RESULTS: While discriminant analysis is routinely and widely used in the analysis of karyometric data, the process of deriving the discriminant function and its coefficients has not been demonstrated in detail, by a numerical example, in over 50 years. CONCLUSION: It is clearly not practical to conduct, by hand, a discriminant analysis on data sets as commonly encountered in karyometry. However, the use of a computer algorithm without a full understanding of the processing steps has always been deeply unsatisfactory. This tutorial article should remedy that situation. © Science Printers and Publishers, Inc.
• Bartels, P. H., Bartels, H. G., & Alberts, D. S. (2009). Karyometry correction algorithm for differences in staining. Analytical and Quantitative Cytology and Histology, 31(2), 63-73.
  More info

  PMID: 19402382;Abstract: OBJECTIVE: To describe an algorithm that allows the correction of differences in staining of histopathologic sections while preserving chromatin texture. STUDY DESIGN: In order to preserve the texture of the nuclear chromatin in the corrected digital imagery, it is necessary to correct the images pixel for pixel. This is accomplished by mapping each pixel's value onto the cumulative frequency distribution of the data set to which the image belongs, to transfer to the cumulative frequency distribution of the data set serving as standard and to project the intersection down onto the pixel optical density scale for the corrected value. RESULTS: Feature values in the corrected imagery, for the majority of features used in karyometry, are between < 1% and a few percent of the feature values in standard imagery. For some higher-order statistical features involving multiple pixels, sensitivity to a shift in the cumulative frequency distribution may exist, and a secondary small correction by a factor may be required. CONCLUSION: The correction algorithm allows the elimination of the effects of small staining differences on karyometric analysis. © Science Printers and Publishers, Inc.
• Bartels, P. H., Montironi, R., Scarpelli, M., Bartels, H. G., & Alberts, D. S. (2009). Knowledge discovery processing and data mining in karyometry. Analytical and Quantitative Cytology and Histology, 31(3), 125-136.
  More info

  PMID: 19634783;Abstract: OBJECTIVE: To present the rationale for applying different sequences of multivariate analysis algorithms to determine if and where, in the large and high-dimensional data space, events have led to change in karyometric features. STUDY DESIGN: Clinical materials and results from the analysis of 4 studies were used: the demonstration of chemopreventive efficacy of letrozole in a situation where only a small subset of cells is affected, the detection of a preneoplastic lesion in colorectal tissue, data processing to document clues that predict risk of recurrence of a bladder lesion and the use of metafeatures and secondorder discriminant analysis in a study of efficacy of vitamin A in the chemoprevention of skin lesions. RESULTS: Evidence for chemopreventive efficacy was demonstrated in the first example only after processing identified the small subpopulation of affected nuclei in a study of breast epithelial cells. Detection of a preneoplastic development is linked to a progression curve connecting nuclei from normal tissue to nuclei from premalignant colorectal lesions. The prediction of risk of recurrence of papillary bladder lesions is possible by detecting changes in nuclei of a certain phenotype. Efficacy of vitamin A as a chemopreventive agent for skin cancer could be demonstrated with a dose-response curve after a second-order discriminant analysis was employed. CONCLUSION: In none of these instances would the information of biologic interest have been revealed by a straightforward, single algorithmic analysis. © Science Printers and Publishers, Inc.
• Bartels, P., Yozwiak, M., Einspahr, J., Saboda, K., Liu, Y., Brooks, C., Bartels, H., & Alberts, D. S. (2009). Chemopreventive efficacy of topical difluoromethylornithine and/or triamcinolone in the treatment of actinic keratoses analyzed by karyometry. Analytical and Quantitative Cytology and Histology, 31(6), 355-366.
  More info

  PMID: 20698351;Abstract: OBJECTIVE: To determine whether low-dose topical applications of difluoromethylornithine (DFMO) with or without Triamcinolone (Fougena, Melville, New York, U.S.A.) to moderately sun-damaged skin with actinic skin keratoses are efficacious. STUDY DESIGN: There were 4 topically administered, 6-month treatments, DFMO + Eucerin (Beiersdorf Inc., Hamburg, Germany), DFMO + Triamcinolone, Triamcinolone + Eucerin and Eucerin + Eucerin (to serve as double placebo). Participant eligibility included evidence of at least 2 actinic keratoses on each posterolateral forearm as well as moderate to severe evidence of sun-damaged skin, as evaluated by a board certified dermatologist. High resolution digitized imagery of nuclei from histologic sections of 4-mm punch biopsies from sun-damaged skin on the posterolateral forearms was recorded, at baseline and at the end of 6 months of study. RESULTS: With 102 participants and 185 skin biopsies, a total of 16,395 skin cell nuclei were recorded. The nuclei were analyzed to assess the changes in the pattern of the nuclear chromatin. Two specific measures of end point evaluation were computed, including the percentage of nuclei with high values of nuclear abnormality and the reduction of the percentage of nuclei assigned by a discriminant function to the baseline data set. All 3 active interventions, including low-dose topical DFMO, topical Triamcinolone and topical DFMO + Triamcinolone, led to statistically significant reductions of both the number of nuclei with high nuclear abnormality as well as the number of nuclei assigned to the baseline data set. These reductions were found for all 3 treatments involving DFMO or Triamcinolone. For the placebo data sets only small, statistically insignificant increases or decreases of these percentages were observed. CONCLUSION: The low-dose, topical drug interventions were all effective in reducing skin biopsy nuclear abnormality by a statistically significant 15-20%, whereas there was no evidence of a double placebo effect by karyometric assessment. These effects were greater than the case-to-case sampling error.
• Frank, D. H., Kimler, B. F., Fabian, C. J., Ranger-Moore, J., Yozwiak, M., Bartels, H. G., Alberts, D. S., & Bartels, P. H. (2009). Digital image analysis of breast epithelial cells collected by random periareolar fine-needle aspirates (RPFNA) from women at high risk for breast cancer taking hormone replacement and the aromatase inhibitor, letrozole, for six months. Breast Cancer Research and Treatment, 115(3), 661-668.
  More info

  PMID: 19125322;Abstract: Aromatase inhibitors are currently being evaluated as preventive agents in post-menopausal women at high risk for breast cancer. A phase II trial of 42 women on hormone replacement therapy (HRT) treated with letrozole for 6 months showed Ki-67 was reduced by 66% but showed no change in cytomorphology or Masood score. Subsequent image analytical procedures (karyometry) conducted on a subset of the samples captured subvisual information that showed reduced cellular abnormality after 6 months of letrozole. In the present study we expanded on the preliminary karyometry study to determine if the change in karyometric measurements corresponded to changes in risk biomarkers quantified in the Phase II trial; and secondly, whether these biomarkers might be used together to serve as markers of response in individual cases. Pap stained slides from the Phase II trial were used. Epithelial cell images were digitized on a CCD video-microphotometer and the nuclei were segmented from the field using a semiautomatic algorithm. Nine out of 37 cases analyzed showed a numerical decrease in all three markers, although only three of these exhibited changes substantial enough to be considered as an improvement. However, 12 cases showed improvement by cytology (a decrease in Masood score of at least 2), an additional 13 cases demonstrated a reduction in Ki-67 expression by 50% of the median baseline value, and an additional five cases exhibited a decrease of at least 10% in abnormal cells by nuclear morphometry. Thus, a total of 30 of 37 cases (81%) showed improvement in at least one marker. There was no correlation between changes in Ki-67%, karyometric abnormality, and Masood score change other than specimens that exhibited an improvement in cytology also displayed greater decreases in nuclear morphometry abnormalities. Given the heterogeneity of mechanisms leading to malignancy, the quantitative analysis of nuclear chromatin patterns may be valuable as a global, or integrating, biomarker of change in chemoprevention studies in conjunction with additional markers. Correlation with long term clinical outcome is needed to validate meaningful combinations of informative biomarkers. © 2009 Springer Science+Business Media, LLC.
• Krouse, R. S., Alberts, D. S., Prasad, A. R., Yozwiak, M., Bartels, H. G., Liu, Y., & Bartels, P. H. (2009). Progression of skin lesions from normal skin to squamous cell carcinoma. Analytical and Quantitative Cytology and Histology, 31(1), 17-25.
  More info

  PMID: 19320189;Abstract: Objective: To assess the changes in the nuclear chromatin pattern concomitant with progressive sun damage in skin biopsies ranging from sun-exposed, normal-appearing skin to squamous cell carcinoma (SCC). Study design: Biopsies were taken from 140 cases with sun-exposed but histopathologically normal skin, from 20 cases visually assessed as pre-actinic keratosis (pre-AK) or early AK,from 30 cases of AK, and from 21 cases of SCC. A total of 21,094 nuclei were recorded from these biopsies. High-resolution digital imagery was recorded, and features descriptive of the nuclear chromatin pattern were computed. Both supervised learning and unsupervised learning algorithms were employed to derive progression plots. Results: With increased sun exposure, the proportion of nuclei exhibiting changes in the nuclear chromatin pattern rises notably. Using karyometry, no significant differences could be substantiated between nuclei collected from early AK sites and AK lesions. Cases of SCC fell into 2 distinct groups. A larger group (∼ 66.7% of cases) had characteristics similar to AK. A smaller group (∼33.3% of cases) represented much more progressed lesions. Conclusion: Karyometric assessment can provide a numeric measure of progression for sun damage and of the deviation from normal in both AK and SCC lesions. © Science Printers and Publishers, Inc.
• Montironi, R., Cheng, L., Lopez-Beltran, A., Mazzucchelli, R., Scarpelli, M., & Bartels, P. H. (2009). Decision support systems for morphology-based diagnosis and prognosis of prostate neoplasms: A methodological approach. Cancer, 115(SUPPL 13), 3068-3077.
  More info

  PMID: 19544548;Abstract: Recent advances in computer and information technologies have allowed the integration of both numeric and non-numeric data, that is, descriptive, linguistic terms. This has led at 1 end of the spectrum of technology development to machine vision based on image understanding and, at the other, to decision support systems. This has had a significant impact on our capability to derive diagnostic and prognostic information from histopathological material with prostate neoplasms. © 2009 American Cancer Society.
• Bartels, P. H. (2008). Light Optical Microscopy. Comprehensive Cytopathology, 1005-1012.
• Bartels, P. H., Krouse, R. S., Prasad, A. R., Yozwiak, M., Liu, Y., Bartels, H. G., & Alberts, D. S. (2008). Actinic damage in histopathologically normal skin. Analytical and Quantitative Cytology and Histology, 30(6), 316-322.
  More info

  PMID: 19160696;Abstract: OBJECTIVE: To establish measures of sun damage in histopathologically normal skin. STUDY DESIGN: Biopsies were taken from the upper inner arm, representing skin with presumably minimum sun exposure, from skin of the forearm with no visible sun damage, from skin of the forearm with visible sun damage and from normal-appearing skin from the forearm of individuals who had sun exposure that had resulted in actinic keratosis (AK) lesions. In addition, a data set of nuclei from AKs was recorded. RESULTS: In histopathologically normal skin, monotonically increasing damage was observed in individuals with increased exposure to solar radiation. CONCLUSION: Karyometry can detect and statistically secure changes in skin due to solar exposure at a stage at which the skin is histopathologically determined to be normal. © Science Printers and Publishers, Inc.
• Bartels, P. H., Yozwiak, M. L., Bartels, H. G., Liu, Y., Hess, L. M., & Alberts, D. S. (2008). Limits of detection of chemopreventive efficacy: Karyometry of skin biopsies. Cancer Epidemiology Biomarkers and Prevention, 17(7), 1689-1695.
  More info

  PMID: 18583468;PMCID: PMC2574734;Abstract: Objective: This study was designed to establish estimates of the smallest effects due to chemopreventive intervention detectable by karyometry in skin biopsies. Methods: Estimates of the smallest change of statistical significance and estimates of the power of the test were derived for several key features descriptive of the distribution of nuclear chromatin. Results from triplicate biopsies from the same case were used to provide estimates of the within-case, biopsy-to-biopsy variance. Results: Generally, a change in feature value due to chemopreventive intervention can be statistically secured when it amounts to 5% to 10%. In clinical trials where matched baseline and end of study biopsies from the same cases are available, paired comparison ANOVA can detect a 2% change on samples of 25 cases. Establishing efficacy in individual cases requires a change in feature values on the order of 10% to 15%. Conclusions: Karyometry provides a sensitive, quantitative method for the assessment of efficacy of chemoprevention. The effects of within-case, biopsy-to-biopsy variance need to be considered only in the evaluation of individual cases and are on the order of 5% in skin biopsies. Copyright © 2008 American Association for Cancer Research.
• Alberts, D. S., Einspahr, J. G., Krouse, R. S., Prasad, A., Ranger-Moore, J., Hamilton, P., Ismail, A., Lance, P., Goldschmid, S., Hess, L. M., Yozwiak, M., Bartels, H. G., & Bartels, P. H. (2007). Karyometry of the colonic mucosa. Cancer Epidemiology Biomarkers and Prevention, 16(12), 2704-2716.
  More info

  PMID: 18086777;Abstract: Objective: The study summarizes results of karyometric measurements in epithelial cells of the colorectal mucosa to document evidence of a field effect of preneoplastic development amongpatients with colorectal adenocarcinoma or adenoma. Methods: Karyometric analyses were done on high-resolution images of histologic sections from 48 patients with colorectal adenocarcinomas and 44 patients with adenomas and on images from matching normal-appearing mucosa directly adjacent to such lesions, at a 1-cm and 10-cm distance from the lesions or from the rectal mucosa of adenoma patients, as well as from 24 healthy normal controls with no family history of colonic disease. Results: The nuclei recorded in the histologically normal-appearing mucosa of patients with either colorectal adenoma or adenocarcinoma exhibited differences in karyometric features in comparison with nuclei recorded in rectal mucosa from patients who were free of a colonic lesion. These differences were expressed to the same extent in tissue adjacent to the lesions and in normal-appearing tissue as distant as the rectum. Conclusions: The nuclear chromatin pattern may serve as an integrating biomarker for a preneoplastic development. The field effect might provide an end point in chemopreventive intervention trials. Copyright © 2007 American Association for Cancer Research.
• Bartels, P. H., Fabian, C. J., Kimler, B. F., Ranger-Moore, J. R., Frank, D. H., Yozwiak, M. L., & Alberts, D. S. (2007). Karyometry of breast epithelial cells acquired by random periareolar fine needle aspiration in women at high risk for breast cancer. Analytical and Quantitative Cytology and Histology, 29(2), 63-70.
  More info

  PMID: 17484269;Abstract: OBJECTIVE: To establish whether karyometry was likely to detect change in the proportion of abnormal cells in random periareolar fine needle aspiration (RPFNA) specimens from high-risk women in a 6-month prevention trial with an aromatase inhibitor. STUDY DESIGN: Papanicolaou-stained ThinPrep slides of RPFNA samples from 11 of 42 women were digitally recorded at high resolution, with 200 cells measured per slide, at baseline (BL) and at the end of study (ES) after 6 months. The nuclear chromatin pattern characteristics were assessed by multivariate analytic techniques; determination of nuclear abnormalities was performed and cells that showed expression of abnormality were identified. RESULTS: The BL FNA samples contain ∼90% cells with a chromatin pattern as expected in a normal cell population. A small subpopulation of cells had deviations from normal. At ES the proportion of these cells was reduced, to a statistically significant degree, from < 10% to 2-5%. CONCLUSION: Nuclear karyometry is a promising technique for characterizing the proportion of cells deviating from normal in cytologic specimens and should be explored further as an intermediate endpoint in prevention trials. © Science Printers and Publishers, Inc.
• Montironi, R., Scarpelli, M., Lopez-Beltran, A., Mazzucchelli, R., Alberts, D., Ranger-Moore, J., Bartels, H. G., Hamilton, P. W., Einspahr, J., & Bartels, P. H. (2007). Chromatin phenotype karyometry can predict recurrence in papillary urothelial neoplasms of low malignant potential. Cellular Oncology, 29(1), 47-58.
  More info

  PMID: 17429141;Abstract: Background: A preceding exploratory study (J. Clin. Pathol. 57(2004), 1201-1207) had shown that a karyometric assessment of nuclei from papillary urothelial neoplasms of low malignant potential (PUNLMP) revealed subtle differences in phenotype which correlated with recurrence of disease. Aim of the study: To validate the results from the exploratory study on a larger sample size. Materials: 93 karyometric features were analyzed on haematoxylin and eosin-stained sections from 85 cases of PUNLMP. 45 cases were from patients who had a solitary PUNLMP lesion and were disease-free during a follow-up period of at least 8 years. The other 40 were from patients with a unifocal PUNLMP, with one or more recurrences in the follow-up. A combination of the previously defined classification functions together with a new P-index derived classification method was used in an attempt to classify cases and identify a biomarker of recurrence in PUNLMP lesions. Results: Validation was pursued by a number of separate approaches. First, the exact procedure from the exploratory study was applied to the large validation set. Second, since the discriminant function 2 of the exploratory study had been based on a small sample size, a new discriminant function was derived. The case classification showed a correct classification of 61% for non-recurrent and 74% for recurrent cases, respectively. Greater success was obtained by applying unsupervised learning technologies to take advantage of phenotypical composition (correct classification of 92%). This approach was validated by dividing the data into training and test sets with 2/3 of the cases assigned to the training sets, and 1/3 to the test sets, on a rotating basis, and validation of the classification rate was thus tested on three separate data sets by a leave-k-out process. The average correct classification was 92.8% (training set) and 84.6% (test set). Conclusions: Our validation study detected subvisual differences in chromatin organization state between non-recurrent and recurrent PUNLMP, thus allowing a very stable method of predicting recurrence of papillary urothelial neoplasms of low malignant potential by karyometry. © 2007 - IOS Press and the authors. All rights reserved.
• Weinstein, R. S., López, A. M., Barker, G. P., Krupinski, E. A., Descour, M. R., Scott, K. M., Richter, L. C., Beinar, S. J., Holcomb, M. J., Bartels, P. H., McNeely, R. A., & Bhattacharyya, A. K. (2007). The innovative bundling of teleradiology, telepathology, and teleoncology services. IBM Systems Journal, 46(1), 69-84.
  More info

  Abstract: Teleradiology, telepathology, and teleoncology are important applications of telemedicine. Recent advances in these fields include a preponderance of radiology PACS (Picture Archiving and Communications System) users, the implementation of around-the-clock teleradiology services at many hospitals, and the invention of the first ultrarapid whole-slide digital scanner based on the array microscope. These advances have led to the development of a new health-care-delivery clinical pathway called the ultrarapid breast care process (URBC), which has been commercialized as the UltraClinics® process. This process bundles telemammography, telepathology, and teleoncology services and has reduced the time it takes for a woman to obtain diagnostic and therapeutic breast-care planning services from several weeks to a single day. This paper describes the UltraClinics process in detail and presents the vision of a network of same-day telemedicine-enabled UltraClinics facilities, staffed by a virtual group practice of teleradiologists, telepathologists, and teleoncologists. © Copyright 2007 by International Business Machines Corporation.
• Bartels, P. H., Ranger-Moore, J., Alberts, D., Hess, L., Scarpelli, M., & Montironi, R. (2006). Carcinogenesis and the hypothesis of phylogenetic reversion. Analytical and Quantitative Cytology and Histology, 28(5), 243-252.
  More info

  PMID: 17067006;Abstract: Chemoprevention must target early molecular events involved in malignant transformation. The sequence of events leading from a normally functioning interphase cell to an uncontrolled tumor cell is only partially understood, impeding systematic design of chemopreventive agents. The respective roles of mutagenic and epigenetic mechanisms have not been definitively established. Also, traditional models do not appear to incorporate cellular response to events leading to carcinogenesis. A perspective on system response offered by complexity science elucidates the roles of feedback and control in maintaining functional stability during carcinogenesis. Carcinogenesis is seen as a process of epigenetic redifferentiation resulting in a cell behaving like an archetypal karyocyte free of growth restraints (phylogenetic reversion). Genes that evolved during the development of multicellular organisms, restraining uncontrolled growth and regulating intercell communication may be systematically silenced during carcinogenesis. The formation of heterochromatin, which results in epigenetic silencing by hypermethylation in CpG-dense islands, finds expression in the nuclear chromatin pattern. Karyometry is an integrating biomarker of chromatin pattern information that accommodates the possibility of multiple, differently ordered pathways and provides exquisite sensitivity, allowing detection of very early transformation events. Its use can monitor the impact of chemopreventive agents on the earliest events in progression to cancer. © Science Printers and Publishers, Inc.
• Bartels, P. H., Descour, M., Alberts, D. S., & Ranger-Moore, J. (2005). The evolution of scanning microphotometry: A historical review. Analytical and Quantitative Cytology and Histology, 27(6), 323-336.
  More info

  PMID: 16450789;Abstract: This article traces the evolution of scanning microphotometry from its beginning, with an emphasis on the developing technologies that have made feasible the rapid and detailed capture of high-resolution image information. Consideration is given to future directions that may prove fruitful to clinical practice. © Science Printers and Publishers, Inc.
• Bartels, P. H., Ranger-Moore, J., Bartels, H. G., & Alberts, D. S. (2005). Second order discriminant analysis in chemopreventive efficacy measurement. Analytical and Quantitative Cytology and Histology, 27(1), 15-26.
  More info

  PMID: 15794448;Abstract: OBJECTIVE: To describe the use of second order discriminant analysis as a classification methodology along with the underlying assumptions and sampling requirements, with special emphasis on the use of this analysis in chemopreventive efficacy studies. STUDY DESIGN: The discriminant function score distributions derived in an analysis of 2 diagnostic groups may show such overlap that a statistically significant difference in mean values cannot be shown and, more important, that a useful case-based classification cannot be attained. By using the discriminant function score distributions from each case, it is frequently possible to derive a second order discriminant function based on case-specific characteristics, rather than characteristics of nuclei, thereby attaining improved case classification. RESULTS: Second order discriminant analysis has proven very useful in the documentation of case-level efficacy in chemopreventive trials. In a study of orally administered vitamin A, a first order discriminant analysis did not achieve a statistically significant difference in the score distributions for nuclei, but a second order discriminant analysis allowed a correct recognition of intervention effects in 85% of submitted cases. In a chemopreventive study of triamcinolone, a similarly inadequate discrimination based on discriminant function scores for nuclei resulted. After a second order discriminant analysis, a reduction in solar-actinic damage could be shown in 14/15, or 93%, of treated cases. CONCLUSION: Second order discriminant analysis can be highly effective when the discriminating information offered at the nuclear level is inadequate due to high dispersion and small differences in mean values of discriminant function scores for the diagnostic groups. Second order analysis utilizes case-specific characteristics of the discriminant function score distributions to document diagnostic group separation and/or efficacy of chemopreventive intervention by a reduction in case discriminant function scores. © Science Printers and Publishers, Inc.
• Brewer, M. A., Ranger-Moore, J., Baruche, A., Alberts, D. S., Greene, M., Thompson, D., Liu, Y., Davis, J., & Bartels, P. H. (2005). Exploratory study of ovarian intraepithelial neoplasia. Cancer Epidemiology Biomarkers and Prevention, 14(2), 299-305.
  More info

  PMID: 15734950;Abstract: Purpose: This was an exploratory study to test two hypotheses related to potential epithelial precursors to ovarian cancer: (a) histologically normal ovarian surface epithelium exhibited changes in the nuclear chromatin pattern, which indicate an ovarian abnormality, and (b) such changes were detectable in the ovarian surface epithelium of cancer-free subjects who were at high risk for ovarian cancer. Experimental Design: Ovaries were carefully collected to avoid damage to the surface epithelium. Five-micron-thick histologic sections were cut and stained with H&E. High-resolution images were recorded from the ovarian surface epithelium and from the underlying stroma of ovaries from normal women (10 cases), women at high risk of developing ovarian cancer (7 cases), and histologically normal areas adjacent to ovarian cancer (3 cases). Karyometric features and measurements of nuclear abnormality were computed for 3,390 epithelial nuclei. Discriminant function analyses and unsupervised learning algorithms were employed to define deviations from normal and to identify the subpopulations of nuclei exhibiting these changes. Results: Epithelium from ovaries harboring a malignant lesion had changes in the nuclear chromatin pattern consistent with a second phenotype, which were not visually detected with histopathologic surveillance. This phenotype was also present in the ovaries obtained from women at increased risk of ovarian cancer, suggesting that it may represent a premalignant abnormality. These changes were statistically significant. Conclusion: The observed changes in karyometric features were sufficiently distinct to warrant further study as both diagnostic and prognostic biomarkers for early detection and prevention of ovarian cancer.
• Frank, D. H., Davis, J. R., Ranger-Moore, J., Liu, Y., Bartels, H. G., Alberts, D. S., & Bartels, P. H. (2005). Karyometry of infiltrating breast lesions. Analytical and Quantitative Cytology and Histology, 27(4), 195-201.
  More info

  PMID: 16220830;Abstract: OBJECTIVE: To characterize nuclei from well-differentiated, moderately differentiated and poorly differentiated lesions of invasive breast cancer by karyometry and to test the hypothesis that these diagnostic categories form homogeneous sets. STUDY DESIGN: Histopathologic sections from 6 cases of well-differentiated, 11 cases of moderately differentiated and 17 cases of poorly differentiated ductal carcinomas were digitally recorded. From each case 100 nuclei were segmented and analyzed by karyometry. A discriminant analysis was performed, and nuclear and lesion signatures were computed. The nonsupervised learning algorithm P-index was applied. A progression curve per diagnostic category based on mean nuclear abnormality and a discriminant function score was derived. RESULTS: The well-differentiated lesions formed a homogeneous set, but both the moderately and poorly differentiated lesions showed 2 significantly different sub-populations with nuclei of substantially different nuclear abnormality and progression. CONCLUSION: The visual histopathologic diagnostic assessment of these lesions was based on an evaluation of both tissue architectural criteria and nuclear criteria. Here, only the pattern of nuclear chromatin was evaluated. Cases belonging to the same diagnostic category as assessed by their differentiation may be further characterized by the extent to which the nuclei deviate from normal. There was substantial case-to-case heterogeneity in these invasive lesions. © Science Printers and Publishers, Inc.
• Ranger-Moore, J., Alberts, D. S., Montironi, R., Garcia, F., Davis, J., Frank, D., Brewer, M., Mariuzzi, G. M., Bartels, H. G., & Bartels, P. H. (2005). Karyometry in the early detection and chemoprevention of intraepithelial lesions. European Journal of Cancer, 41(13), 1875-1888.
  More info

  PMID: 16087328;Abstract: The ideal chemopreventive agent targets pre-neoplastic changes and intraepithelial neoplasia, preventing progression over time without notable side effects. Assessment of success of chemopreventive intervention in the short and medium term remains a challenge, and in this review the suggestion is investigated that karyometric measurements constitute suitable markers of chemopreventive efficacy. Karyometry provides the sensitivity required to detect small differences amidst relatively high biological variability. It can help establish progression curves of intraepithelial neoplasia (IEN) to invasive cancer, and thus detect chemopreventive effects. Such effects can be observed in two ways, at the group level (intervention vs. placebo), and at the case (or patient) level. The latter is more difficult to establish, necessitating the development of specialised statistical methods. Analysis of between-case and within-case heterogeneity can reveal useful information about cancer progression and prevention. We suggest that karyometry can objectively quantify IEN progression, providing a framework for statistically securing chemopreventive effects. It can act as an integrating biomarker by detecting chemopreventive activity even when the mechanism for a given progression pathway is unknown, or when multiple pathways exist. The sensitivity of karyometric detection can help optimise the design of clinical trials of novel chemopreventive agents by decreasing trial duration and/or sample size. © 2005 Elsevier Ltd. All rights reserved.
• Ranger-Moore, J., Frank, D., Lance, P., Alberts, D., Yozwiak, M., Bartels, H. G., Einspahr, J., & Bartels, P. H. (2005). Karyometry in rectal mucosa of patients with previous colorectal adenomas. Analytical and Quantitative Cytology and Histology, 27(3), 134-142.
  More info

  PMID: 16121634;Abstract: OBJECTIVE: To determine whether karyometric measurements taken in biopsies from histologically normal-appearing rectal mucosa could serve as a biomarker for the risk of recurrence of polyps. MATERIALS AND METHODS: Biopsies were taken from the rectal mucosa of cases with a prior history of colonic polyps at the baseline of the study. In 57 cases recurrent polyps occurred (R cases); in 72 cases no recurrent disease was found at the end of the study (NR cases). From each biopsy 100 nuclei were recorded at high resolution. After segmentation, feature extraction and selection of a discriminating subset of features, a number of discriminant functions were derived. Also, measures of nuclear abnormality were computed. RESULTS: The differences in karyometric feature values for nuclei from biopsies of cases with recurrent or nonrecurrent disease were very small and not notably expressed in the majority of nuclei. It was possible by focusing on nuclei showing clear deviations from normal to derive a discriminant function that exhibited a shift for the NR and R data sets. The distributions of discriminant function scores were then subjected to a second-order discriminant analysis to separate cases according to recurrence status. This function showed a statistically highly significant correlation with recurrence. At one extreme of its score distribution were 11 of 57 cases that had a recurrence, and at the other extreme were 8-10 of 72 cases that had no recurrence. The distributions of nuclear abnormality values for these subsets of cases were drastically different, with an average value of 1.72 for the group that may be at high risk for another recurrence and 1.02 for the group possibly at low risk. All cases with a prior history of colonic polyps showed a nuclear abnormality deviating from normal. CONCLUSION: Measurement of a sample of 100 nuclei from the rectal mucosa will suggest, for approximately 10% of cases, that a high risk for recurrence of adenomatous polyps exists and, for a slightly lower proportion, confirm that the nuclei deviate only slightly from those from individuals with no history of colonic polyps. For the majority of cases with a prior history of adenoma, the nuclei in the biopsy show a notable deviation from normal, but the deviation is practically the same for cases that had a recurrence and those that did not. However, a tentative discriminant function (DF I,3) derived from the characteristics of the extreme cases correctly classified approximately 64% of nonrecurrent and 83% of recurrent cases using a Bayesian decision boundary. © Science Printers and Publishers, Inc.
• A., F., Ranger-Moore, J., Barker, B., Davis, J., Brewer, M., Lozevski, J., Vinyak, S., Liu, Y., Yemane, J., Hatch, K. D., Alberts, D. S., Bartels, H. G., & Bartels, P. H. (2004). Karyometric image analysis for intraepithelial and invasive cervical lesions. Analytical and Quantitative Cytology and Histology, 26(3), 141-150.
  More info

  PMID: 15218690;Abstract: OBJECTIVE: To derive an objective, numeric measure for the progression of intraepithelial and invasive squamous cell cervical lesions. STUDY DESIGN: Thin-layer cervical cytology preparations from colposcopically confirmed normal cervix, low grade squamous intraepithelial lesions, high grade squamous intraepithelial lesions and carcinoma were identified from a cross-sectional study. Fifty-nine cases representing 4 diagnostic categories were selected, and 2,375 nuclei from epithelial cells representative of the diagnostic category were randomly selected for imaging and measurement from these cases. Additionally, 1,378 visually normal appearing intermediate cells from low and high grade squamous intraepithelial lesions, as well as from carcinoma cases, were identified for analysis. The nuclei were quantitatively characterized, and discriminant analyses were performed to derive a progression curve from normal cytology to carcinoma. RESULTS: The lesion signatures show a clear increase in nuclear abnormality with increasing progression. A progression curve was derived based on mean discriminant function scores for each diagnostic category and on the mean nuclear abnormality values for the nuclei in each category, as expressed by their deviation in feature values from normal reference nuclei. CONCLUSION: A numeric assessment of lesion progression for cervical precancerous and cancerous lesions based on karyometric measurements is possible and may provide an objective, precise characterization of each lesion as well as a basis for improved performance in automated cytology-based cervical cancer screening.
• Alberts, D. S., Ranger-Moore, J., Einspahr, J., Saboda, K., Bozzo, P., Liu, Y., Xu, X., Lotan, R., Warneke, J., Salasche, S., Stratton, S., Levine, N., Goldman, R., Islas, M., Duckett, L., Thompson, D., & Bartels, P. (2004). Safety and Efficacy of Dose-Intensive Oral Vitamin A in Subjects with Sun-Damaged Skin. Clinical Cancer Research, 10(6), 1875-1880.
  More info

  PMID: 15041701;Abstract: Purpose: Previously, we reported the results of a Phase III, placebo-controlled trial in 2,297 randomized participants with moderately severe actinic keratoses wherein 25,000 IU/ day vitamin A caused a 32% risk reduction in squamous cell skin cancers. We hypothesized that dose escalation of vitamin A to 50,000 or 75,000 IU/day would be both safe and more efficacious in skin cancer chemoprevention. Experimental Design: One hundred and twenty-nine participants with severely sun-damaged skin on their lateral forearms were randomized to receive placebo or 25,000, 50,000, or 75,000 IU/day vitamin A for 12 months. The primary study end points were the clinical and laboratory safety of vitamin A, and the secondary end points included quantitative, karyometric image analysis and assessment of retinoid and rexinoid receptors in sun-damaged skin. Results: There were no significant differences in expected clinical and laboratory toxicities between the groups of participants randomized to placebo, 25,000 IU/day, 50,000 IU/day, and 75,000 IU/day. Karyometric features were computed from the basal cell layer of skin biopsies, and a total of 22,600 nuclei from 113 participants were examined, showing statistically significant, dose-response effects for vitamin A at the 25,000 and 50,000 IU/day doses. These karyometric changes correlated with increases in retinoic acid receptor α, retinoic acid receptor β, and retinoid X receptor α at the 50,000 IU/day vitamin A dose. Conclusions: The vitamin A doses of 50,000 and 75,000 IU/day for 1 year proved safe and equally more efficacious than the 25,000 IU/day dose and can be recommended for future skin cancer chemoprevention studies.
• Bartels, P. H. (2004). In memoriam George L. Wied, M.D., D.Sc. (hon), F.I.A.C. February 7, 1921-July 25, 2004.. Analytical and quantitative cytology and histology / the International Academy of Cytology [and] American Society of Cytology, 26(6), 301-303.
  More info

  PMID: 15852560;
• Brewer, M. A., Ranger-Moore, J., Greene, M. H., Alberts, D. S., Liu, Y., Bartels, H. G., Baruch, A. C., & Bartels, P. H. (2004). Preneoplastic changes in ovarian tissues. Analytical and Quantitative Cytology and Histology, 26(4), 207-216.
  More info

  PMID: 15457673;Abstract: OBJECTIVE: To test whether histologically normal epithelium within ovarian inclusion cysts and stroma exhibit changes in nuclear chromatin pattern that indicate the presence of occult ovarian lesions. STUDY DESIGN: Ovaries were collected from 10 lowrisk women, from 7 high-risk women and from 3 women with ovarian cancer. Histologic sections were cut at 5 μm and hematoxylin and eosin stained. High-resolution images were recorded from the epithelium lining inclusion cysts and from the underlying stroma of ovaries from these 20 subjects. A total of 2,860 epithelial nuclei and 3,610 stromal nuclei were recorded. Karyometric features and nuclear abnormality were computed. Discriminant analyses and unsupervised learning algorithms defined deviations from normal that were designated "above threshold" and used to compute average nuclear abnormality of a second nuclear phenotype. RESULTS: Histologically normal epithelium from inclusion cysts of ovaries harboring a malignant lesion was shown to exhibit changes in the nuclear chromatin pattern that were statistically significant using quantitative image analysis procedures. Similar changes were seen in the inclusion cyst epithelia of high-risk ovaries. A sub-population of cells representing a new phenotype was detected in the underlying stroma of women harboring a malignant ovarian lesion and in women at high risk of ovarian cancer. CONCLUSION: The karyometric changes observed in the epithelia lining inclusion cysts and in the underlying stroma of ovaries either with ovarian cancer or at high risk of ovarian cancer suggest the presence of preneoplastic changes in histologically normal tissue.
• Diamond, J., Anderson, N. H., Bartels, P. H., Montironi, R., & Hamilton, P. W. (2004). The use of morphological characteristics and texture analysis in the identification of tissue composition in prostatic neoplasia. Human Pathology, 35(9), 1121-1131.
  More info

  PMID: 15343515;Abstract: Quantitative examination of prostate histology offers clues in the diagnostic classification of lesions and in the prediction of response to treatment and prognosis. To facilitate the collection of quantitative data, the development of machine vision systems is necessary. This study explored the use of imaging for identifying tissue abnormalities in prostate histology. Medium-power histological scenes were recorded from whole-mount radical prostatectomy sections at × 40 objective magnification and assessed by a pathologist as exhibiting stroma, normal tissue (nonneoplastic epithelial component), or prostatic carcinoma (PCa). A machine vision system was developed that divided the scenes into subregions of 100 × 100 pixels and subjected each to image-processing techniques. Analysis of morphological characteristics allowed the identification of normal tissue. Analysis of image texture demonstrated that Haralick feature 4 was the most suitable for discriminating stroma from PCa. Using these morphological and texture measurements, it was possible to define a classification scheme for each subregion. The machine vision system is designed to integrate these classification rules and generate digital maps of tissue composition from the classification of subregions; 79.3% of subregions were correctly classified. Established classification rates have demonstrated the validity of the methodology on small scenes; a logical extension was to apply the methodology to whole slide images via scanning technology. The machine vision system is capable of classifying these images. The machine vision system developed in this project facilitates the exploration of morphological and texture characteristics in quantifying tissue composition. It also illustrates the potential of quantitative methods to provide highly discriminatory information in the automated identification of prostatic lesions using computer vision. © 2004 Elsevier Inc. All rights reserved.
• Montironi, R., Thompson, D., Scarpelli, M., Mazzucchelli, R., Peketi, P., Hamilton, P. W., Bostwick, D. G., & Bartels, P. H. (2004). Karyometry detects subvisual differences in chromatin organization state between cribriform and flat high-grade prostatic intraepithelial neoplasia. Modern Pathology, 17(8), 928-937.
  More info

  PMID: 15105811;Abstract: This digital texture analysis-based study evaluates the chromatin organization state in flat and cribriform high-grade prostatic intraepithelial neoplasia (PIN), in the adjacent normal looking secretory epithelium and in the co-occurring adenocarcinoma. Digital texture analysis (karyometry) was carried out on hematoxylin and eosin-stained sections from 24 radical prostatectomy specimens with high-grade PIN (12 with flat and 12 with cribriform architectural pattern, respectively) and cancer. Quantification was also conducted on the normal looking secretory epithelium. Discriminant analysis and the nonsupervised learning algorithm P-index were used to identify suitable subsets of features useful for the discrimination and classification of pathological groups and to explore multivariate data structure in the pathological subgroups. The average nuclear abnormality increases monotonically from the histologically normal appearing secretory epithelium to high-grade PIN and to adenocarcinoma. The nuclei from the so-called perimeter compartment of the flat high-grade PIN lesions show a higher nuclear abnormality compared to the nuclei of the cribriform high-grade PINs. Discriminant analysis shows that flat and cribriform high-grade PINs fall into two populations. Processing by the nonsupervised learning algorithm P-index revealed the existence of three well-defined, distinct subpopulations of nuclei of different chromatin phenotype. In the flat high-grade PIN lesions the proportions of nuclei in the three subpopulations are 16.5% (low abnormality), 25.0% (mid abnormality) and 58.5% (high abnormality), respectively. In the cribriform high-grade PIN lesions, 100% of the nuclei are in the mid-abnormality subpopulation. These differences are also discernible in the co-occurring adenocarcinoma and the histologically normal appearing secretory epithelium. To conclude, karyometry and statistical analysis detect the existence of distinct cell subpopulations of different chromatin packaging and phenotype, with the nuclei from the flat high-grade PIN lesions, adjacent normal looking epithelium and co-occurring adenocarcinoma expressing a greater nuclear abnormality than in the specimens with cribriform high-grade PIN.
• Ranger-Moore, J., Bartels, P. H., Bozzo, P., Einspahr, J., Liu, Y., Saboda, K., & Alberts, D. S. (2004). Karyometric analysis of actinic damage in unexposed and sun-exposed skin and in actinic keratoses in untreated individuals. Analytical and Quantitative Cytology and Histology, 26(3), 155-165.
  More info

  PMID: 15218692;Abstract: OBJECTIVE: To assess the ability of karyometric analysis to demonstrate progression of actinic damage as a function of sun exposure in individuals with actinic keratoses (AKs) and to evaluate the stability of that assessment over a 3-month period. STUDY DESIGN: Biopsies from subjects with AKs were obtained from unexposed skin, sun-exposed skin and AK lesions. Subjects used an SPF 50 sunscreen, and 3 months later additional biopsies were taken from sun-exposed and AK sites. A total of 13,300 nuclei were recorded from 31 subjects. RESULTS: Measures of nuclear abnormality (NA) and effects of sun damage based on discriminant function (DF) scores were derived. Actinic damage levels varied significantly across biopsy site, demonstrating the method's sensitivity. Accrual of actinic damage was demonstrated in sun-exposed skin and AK lesions when all nuclei were examined over 3 months but only for sun-exposed skin when the worst-damaged nuclei were examined. This suggests a ceiling effect of nuclear damage in the progression to abnormality. Within-subject variability was similar for both NA and DF when all nuclei were considered. Among the worst-damaged nuclei (as defined by high DF), DF showed lower within-case variability than NA, perhaps due to a reduction in nuclear heterogeneity. CONCLUSION: Karyometry's ability to detect subtle levels of actinic damage in nuclear chromatin patterns may make it useful in screening agents for possible use in cancer chemoprevention.
• Scarpelli, M., Montironi, R., Salvolini, U., Messori, A., & Bartels, P. H. (2004). Karyometric Features of Low and High Grade Glial Tumors as Compared with Their MRI Appearance. Analytical and Quantitative Cytology and Histology, 26(2), 87-96.
  More info

  PMID: 15131896;Abstract: OBJECTIVE: To compare the results of a magnetic resonance imaging (MRI) grading designed to identify low and high grade gliomas with karyometry used as a tool to grade primary brain tumors. STUDY DESIGN: A consecutive series of 23 primary brain tumors was selected for this study. The neuroradiologist, not knowing the histologic diagnoses, divided the cases into low and high grade categories on the basis of the following 7 features: border sharpness, heterogeneity without contrast, cavitation, contrast enhancement, hypervascularity, mass effect and perifocal T2 hyperintensity. To each feature was given a numerical value, ranging from 1 to 3. All the cases were reviewed and classified by the same pathologist, blinded to the MRI diagnosis. Two hundred nuclei per case were recorded, and 93 karyometric features related to nuclear area, total optical density and chromatin distribution were analyzed for each nucleus. Statistical analysis included discriminant analysis, Kruskal-Wallis test, nonsupervised learning algorithm P-index and Beale statistic. RESULTS: Ten cases were classified as low grade on the basis of their MRI features. The corresponding histopathologic diagnoses were: grade 2 astrocytoma in 2 cases and grade 2 oligodendroglioma in 8 cases. An MRI diagnosis of high grade tumor was made in 13 cases. In 10 cases it was confirmed by the histopathologic diagnosis (3 grade 3 astrocytomas, 1 grade 3 oligodendroglioma and 6 glioblastomas). In the remaining 3 cases the histologic examination revealed a low grade tumor, 1 grade 2 astrocytoma and 2 grade 2 oligodendrogliomas. For the purposes of the karyometric analysis the cases were allocated to the low or high grade category according to their histologic diagnoses (13 cases low grade and 10 cases high grade). Nuclei from low and high grade tumors showed clearly different karyometric characteristics. The oligodendroglioma nuclei had abnormality values close to the low grade standard, while the astrocytoma nuclei were a highly dispersed group with characteristics indicative of a higher degree of nuclear abnormality than the oligodendroglioma nuclei. The results of karyometric analysis showed that grade 2 tumors, corresponding to the low grade group, form a rather distinct category from grade 3 and 4 tumors belonging to the high grade group. CONCLUSION: The results of MRI grading based on a series of features that are routinely assessed by the neuroradiologist to reach a final diagnosis correlate highly with the histopathologic diagnosis. Karyometry can be a useful adjunct to histologic grading.
• Scarpelli, M., Montironi, R., Tarquini, L. M., Hamilton, P. W., Beltran, A. L., Ranger-Moore, J., & Bartels, P. H. (2004). Karyometry detects subvisual differences in chromatin organisation state between non-recurrent and recurrent papillary urothelial neoplasms of low malignant potential. Journal of Clinical Pathology, 57(11), 1201-1207.
  More info

  PMID: 15509685;PMCID: PMC1770490;Abstract: Aim: To analyse nuclear chromatin texture in non-recurrent and recurrent papillary urothelial neoplasms of low malignant potential (PUNLMPs). Materials: Ninety three karyometric features were analysed on haematoxylin and eosin stained sections from 20 PUNLMP cases: 10 from patients with a solitary PUNLMP lesion, who were disease free during at least eight years' follow up, and 10 from patients with unifocal PUNLMP, one or more recurrences being seen during follow up. Results: Kruskal-Wallis analysis was used to search for features showing significant differences between recurrent and non-recurrent cases. Significance was better than p
• Weinstein, R. S., Descour, M. R., Liang, C., Barker, G., Scott, K. M., Richter, L., Krupinski, E. A., Bhattacharyya, A. K., Davis, J. R., Graham, A. R., Rennels, M., Russum, W. C., Goodall, J. F., Zhou, P., Olszak, A. G., Williams, B. H., Wyant, J. C., & Bartels, P. H. (2004). An array microscope for ultrarapid virtual slide processing and telepathology. Design, fabrication, and validation study. Human Pathology, 35(11), 1303-1314.
  More info

  PMID: 15668886;Abstract: This paper describes the design and fabrication of a novel array microscope for the first ultrarapid virtual slide processor (DMetrix DX-40 digital slide scanner). The array microscope optics consists of a stack of three 80-element 10 x 8-lenslet arrays, constituting a "lenslet array ensemble." The lenslet array ensemble is positioned over a glass slide. Uniquely shaped lenses in each of the lenslet arrays, arranged perpendicular to the glass slide constitute a single "miniaturized microscope." A high-pixel-density image sensor is attached to the top of the lenslet array ensemble. In operation, the lenslet array ensemble is transported by a motorized mechanism relative to the long axis of a glass slide. Each of the 80 miniaturized microscopes has a lateral field of view of 250 microns. The microscopes of each row of the array are offset from the microscopes in other rows. Scanning a glass slide with the array microscope produces seamless two-dimensional image data of the entire slide, that is, a virtual slide. The optical system has a numerical aperture of N.A. = 0.65, scans slides at a rate of 3 mm per second, and accrues up to 3,000 images per second from each of the 80 miniaturized microscopes. In the ultrarapid virtual slide processing cycle, the time for image acquisition takes 58 seconds for a 2.25 cm2 tissue section. An automatic slide loader enables the scanner to process up to 40 slides per hour without operator intervention. Slide scanning and image processing are done concurrently so that post-scan processing is eliminated. A virtual slide can be viewed over the Internet immediately after the scanning is complete. A validation study compared the diagnostic accuracy of pathologist case readers using array microscopy (with images viewed as virtual slides) and conventional light microscopy. Four senior pathologists diagnosed 30 breast surgical pathology cases each using both imaging modes, but on separate occasions. Of 120 case reads by array microscopy, there were 3 incorrect diagnoses, all of which were made on difficult cases with equivocal diagnoses by light microscopy. There was a strong correlation between array microscopy vs. "truth" diagnoses based on surgical pathology reports. The kappa statistic for the array microscopy vs. truth was 0.96, which is highly significant (z = 10.33, p < 0.001). There was no statistically significant difference between rates of agreement with truth between array microscopy and light microscopy (z = 0.134, p > 0.05). Array microscopy and light microscopy did not differ significantly with respect to the number/percent of correct decisions rendered (t = 0.552, p = 0.6376) or equivocal decisions rendered (t = 2.449, p = 0.0917). Pathologists rated 95.8% of array microscopy virtual slide images as good or excellent. None were rated as poor. The mean viewing time for a DMetrix virtual slide was 1.16 minutes. The DMetrix virtual slide processor has been found to reduce the virtual slide processing cycle more than 10 fold, as compared with other virtual slide systems reported to date. The virtual slide images are of high quality and suitable for diagnostic pathology, second opinions, expert opinions, clinical trials, education, and research. © 2004 Elsevier Inc. All rights reserved.
• A., F., Davis, J. R., Alberts, D. S., Liu, Y., Thompson, D., & Bartels, P. H. (2003). A Karyometric Approach to the Characterization of Atypical Endometrial Hyperplasia with and Without Co-occurring Adenocarcinoma. Analytical and Quantitative Cytology and Histology, 25(6), 339-346.
  More info

  PMID: 14714300;Abstract: OBJECTIVE: To develop a karyometric image analysis approach to distinguishing atypical endometrial hyperplasia with and without co-occurring adenocarcinoma. STUDY DESIGN: Six cases of atypical hyperplasia without and 6 cases with co-occurring adenocarcinoma, 4 cases of normal endometrium and 3 cases of adenocar cinoma were identified. From each case 100 nuclei were measured in representative diagnostic areas identified by an experienced pathologist. Discriminant analyses were performed. An unsupervised learning algorithm was applied to define and characterize different nuclear phenotypes, and those data were used to identify cases with cooccurring adenocarcinoma. RESULTS: Discriminant analysis showed that nuclei from atypical hyperplasia and atypical hyperplasia with co-occurring adenocarcinoma are statistically different. The unsupervised learning algorithm revealed differences in nuclear subpopulations that can be used to correctly identify an estimated 85% of individual cases. CONCLUSION: Nuclei from atypical hyperplasia without and with co-occurring adenocarcinoma have statistically different karyometric characteristics that may facilitate case classification.
• Anderson, N., Houghton, J., Kirk, S. J., Frank, D., Ranger-Moore, J., Alberts, D. S., Thompson, D., & Bartels, P. H. (2003). Malignancy-associated changes in lactiferous duct epithelium. Analytical and Quantitative Cytology and Histology, 25(2), 63-72.
  More info

  PMID: 12746974;Abstract: OBJECTIVE: To determine whether cells from histologically normal appearing epithelium of the lactiferous duct from women with a remote ductal lesion in the breast provide any clues indicating the existence of such a lesion. STUDY DESIGN: Tissue sections cut to 4 μm and stained with hematoxylin and eosin were prepared from duct tissue of 20 women with breast lesions and of 20 women free of any such lesion who had undergone mammoplastic procedures or resection for benign reasons. One hundred nuclei were measured from each case. Measures of nuclear deviation from normal were computed, discriminant functions were derived, and multivariate significance tests were conducted. RESULTS: Nuclei from histologically normal appearing regions of lactiferous duct epithelium from women harboring distant lesions exhibited changes in the distribution pattern of their nuclear chromatin, indicating the presence of these lesions. The statistical significance of these changes was documented. The changes were clearly evident in all 20 subjects with lesions and were not observed in 19 of the 20 subjects without lesions. CONCLUSION: The results suggest that studies aimed at detecting malignancy-associated changes in cells collected by ductal lavage might lead to a minimally invasive screening procedure for breast lesions.
• Montironi, R., Scarpelli, M., Mazzucchelli, R., Hamilton, P. W., Thompson, D., Ranger-Moore, J., Bostwick, D. G., & Bartels, P. H. (2003). Subvisual changes in chromatin organization state are detected by karyometry in the histologically normal urothelium in patients with synchronous papillary carcinoma. Human Pathology, 34(9), 893-901.
  More info

  PMID: 14562285;Abstract: This study analyzed the chromatin organization state in histologically normal urothelium in patients with synchronous papillary carcinoma using digital texture analysis. The quantitative evaluation was carried out on hematoxylin and eosin-stained sections from 17 cases of urothelial papillary carcinoma in which a simultaneous biopsy specimen featuring histologically normal urothelium was available. Five bladder biopsy specimens of histologically normal urothelium from patients with prostate pathology in whom cystoscopy revealed a normal bladder mucosa were also analyzed. Karyometry showed that the 17 cases of papillary carcinoma, morphologically classified according to the 1973 World Health Organization scheme, belonged to a continuous spectrum or trend curve spanning grade 1 to grade 3. An abnormal pattern and distribution of the nuclear chromatin was seen in the normal4ooking urothelium from the 17 bladders with papillary lesions. When this population was plotted along the trend curve, it occupied an intermediate position between the normal samples and samples from grade 1 carcinoma. When the nuclei were considered individually, the changes were detected only in a subpopulation of nuclei with chromatin alteration pointing toward that seen in grade 1 cases, even though distinct from them. In conclusion, karyometry can detect an abnormal chromatin pattern and distribution in the normal-looking urothelium adjacent to papillary carcinoma. Such alterations correspond to the so-called "malignancy-associated change." © 2003 Elsevier Inc. All rights reserved.
• Price, G. J., McCluggage, W. G., Morrison, M. L., McClean, G., Venkatraman, L., Diamond, J., Bharucha, H., Montironi, R., Bartels, P. H., Thompson, D., & Hamilton, P. W. (2003). Computerized Diagnostic Decision Support System for the Classification of Preinvasive Cervical Squamous Lesions. Human Pathology, 34(11), 1193-1203.
  More info

  PMID: 14652822;Abstract: Previous studies have revealed considerable interobserver and intraobserver variation in the histological classification of preinvasive cervical squamous lesions. The aim of the present study was to develop a decision support system (DSS) for the histological interpretation of these lesions. Knowledge and uncertainty were represented in the form of a Bayesian belief network that permitted the storage of diagnostic knowledge and, for a given case, the collection of evidence in a cumulative manner that provided a final probability for the possible diagnostic outcomes. The network comprised 8 diagnostic histological features (evidence nodes) that were each independently linked to the diagnosis (decision node) by a conditional probability matrix. Diagnostic outcomes comprised normal; koilocytosis; and cervical intraepithelial neoplasia (CIN) I, CIN II, and CIN III. For each evidence feature, a set of images was recorded that represented the full spectrum of change for that feature. The system was designed to be interactive in that the histopathologist was prompted to enter evidence into the network via a specifically designed graphical user interface (i-Path Diagnostics, Belfast, Northern Ireland). Membership functions were used to derive the relative likelihoods for the alternative feature outcomes, the likelihood vector was entered into the network, and the updated diagnostic belief was computed for the diagnostic outcomes and displayed. A cumulative probability graph was generated throughout the diagnostic process and presented on screen. The network was tested on 50 cervical colposcopic biopsy specimens, comprising 10 cases each of normal, koilocytosis, CIN I, CIN II, and CIN III. These had been preselected by a consultant gynecological pathologist. Using conventional morphological assessment, the cases were classified on 2 separate occasions by 2 consultant and 2 junior pathologists. The cases were also then classified using the DSS on 2 occasions by the 4 pathologists and by 2 medical students with no experience in cervical histology. Interobserver and intraobserver agreement using morphology and using the DSS was calculated with κ statistics. Intraobserver reproducibility using conventional unaided diagnosis was reasonably good (κ range, 0.688 to 0.861), but interobserver agreement was poor (κ range, 0.347 to 0.747). Using the DSS improved overall reproducibility between individuals. Using the DSS, however, did not enhance the diagnostic performance of junior pathologists when comparing their DSS-based diagnosis against an experienced consultant. However, the generation of a cumulative probability graph also allowed a comparison of individual performance, how individual features were assessed in the same case, and how this contributed to diagnostic disagreement between individuals. Diagnostic features such as nuclear pleomorphism were shown to be particularly problematic and poorly reproducible. DSSs such as this therefore not only have a role to play in enhancing decision making but also in the study of diagnostic protocol, education, self-assessment, and quality control. © 2003 Elsevier Inc. All rights reserved.
• Ranger-Moore, J., Bozzo, P., Alberts, D., Einspahr, J., Liu, Y., Thompson, D., Stratton, S., Stratton, M. S., & Bartels, P. (2003). Karyometry of Nuclei from Actinic Keratosis and Squamous Cell Cancer of the Skin. Analytical and Quantitative Cytology and Histology, 25(6), 353-361.
  More info

  PMID: 14714302;Abstract: OBJECTIVE: To use karyometric analysis methods to compare actinic keratoses (AKs) to squamous cell carcinomas (SCCs) to determine if SCCs showed a logical progression beyond that seen in AKs and to explore variability within and between lesion types to better understand distinctions between the 2. STUDY DESIGN: Biopsies from 31 subjects with AKs were obtained from upper inner arm skin, forearm skin and AK lesions. Biopsies from 23 different subjects in a related subproject provided SCC biopsies for comparison. RESULTS: Karyometric measures of nuclear abnormality and sun damage were derived. Mean actinic damage levels progressed logically from inner arm to sunexposed skin, to AK, to SCC. Considerable heterogeneity existed at the case level. Unsupervised learning methods revealed 2 distinct clusters of progressed lesions with different nuclear signatures, reflecting differing levels of actinic damage. Number of AKs and SCCs and invasiveness and differentiation of SCCs were distributed across both clusters in roughly equivalent proportions. CONCLUSION: Karyometric methods, shown previously to be capable of sensitively detecting subtle nuclear changes, revealed the possibility of 2 progression pathways, each containing AKs and SCCs. This finding may have prognostic implications.
• A., F., Davis, J. R., Hatch, K., Alberts, D. S., Thompson, D., & Bartels, P. H. (2002). Karyometry in endometrial adenocarcinoma of different grades. Analytical and Quantitative Cytology and Histology, 24(2), 93-102.
  More info

  PMID: 12026057;Abstract: OBJECTIVE: To characterize nuclei from adenocarcinoma of the endometrium of different grades in order to establish whether significantly different nuclear subpopulations exist. STUDY DESIGN: A total of 1,400 nuclei from 14 cases of adenocarcinoma of the endometrium of grades 1, 2 and 3 were recorded, and 600 nuclei from normal, secretory phase glandular endometrial epithelium were used as a reference data set. Karyometric features were computed and a discriminant function derived to define a trend in feature values for nuclei from lesions of different grades. The nuclei from lesions of different grades were processed by a nonsupervised learning algorithm in an effort to detect and define subpopulations. RESULTS: Nuclei from grade 1 lesions represented a near-diploid stemline, whereas nuclei from grade 3 lesions formed an aneuploid set with a mode around 3N. The existence of three subpopulations with statistically different nuclear chromatin patterns could be shown for nuclei from each grade. In each grade these three subpopulations occupied the same relative position in feature space. For grade 1 lesions all three subpopulations were near diploid, for grade 3 lesions all three were approximately triploid. CONCLUSION: The nuclei in adenocarcinoma of the endometrium form a heterogeneous set, with subpopulations of distinctly different chromatin patterns and different ploidy. This should be taken into consideration in assessing the efficacy of chemopreventive intervention.
• Bartels, P. H., Ranger-Moore, J., Stratton, M. S., Bozzo, P., Einspahr, J., Liu, Y., Thompson, D., & Alberts, D. S. (2002). Statistical analysis of chemopreventive efficacy of vitamin A in sun-exposed, normal skin. Analytical and Quantitative Cytology and Histology, 24(4), 185-197.
  More info

  PMID: 12199319;Abstract: OBJECTIVE: To develop numeric, statistically secured measures of chemopreventive efficacy and to derive procedures with high sensitivity of detection. STUDY DESIGN: Karyometric features were computed for nuclei from the basal cell layer of biopsies taken from sun-exposed but histologically "normal" skin. Biopsies were collected from placebo-treated subjects and subjects treated for one year with daily, oral doses of 25,000, 50,000 and 75,000 IU of vitamin A. A total of 22,600 nuclei were recorded from 113 cases, at baseline and after one year. RESULTS: Two numeric measures of chemopreventive efficacy were applied: a measure of nuclear abnormality and a measure based on discriminant function scores. Both showed statistically significant chemopreventive effects of vitamin A. Dose-response curves were derived. A novel procedure, second order discriminant analysis, resulted in very high sensitivity for the detection of change in nuclear chromatin patterns. CONCLUSION: Karyometric analysis has increased in sensitivity such that changes on the order of 10%, found in only a low percentage of nuclei in a biopsy specimen, can be reliably documented. The methodology lends itself to cost-efficient screening of compounds for chemopreventive efficacy.
• Mariuzzi, L., Mombello, A., Granchelli, G., Rucco, V., Tarocco, E., Frank, D., Davis, J., Thompson, D., Bartels, H., Sie, M., Mariuzzi, G. M., & Bartels, P. H. (2002). Quantitative study of breast cancer progression: Different pathways for various In Situ cancers. Modern Pathology, 15(1), 18-25.
  More info

  PMID: 11796837;Abstract: The chromatin pattern in nuclei from breast ductal proliferative lesions was quantitatively evaluated with the objective of deriving measures of tumor progression. A total of 110 cases were analyzed. There were 38 cases of normal tissue or benign proliferative lesions, 41 cases of ductal carcinoma in situ (DCIS), and 31 cases of microinfiltrating DCIS and of infiltrating cancer. A total of 9424 nuclei were analyzed. High-resolution images were digitally recorded. For each nucleus, 93 karyometric features descriptive of the spatial and statistical distribution of the nuclear chromatin were computed. Data analysis included establishing a profile of relative deviations of each feature from "normal," called the nuclear signature, and of lesion signatures as well as of trends of lesion progression. Two trends of evolution could be discerned: One from normal to hyperplasia, atypical hyperplasia, and comedo DCIS as representative of high-grade lesions; and the other from normal to hyperplasia to cribriform DCIS, solid DCIS, and infiltrating cancer, representing lower grade lesions. The nuclei in microinfiltrating foci are distinctly different from nuclei in high-grade comedo DCIS. The nuclei in microinfiltrating foci have a statistically significantly lower nuclear abnormality. They may represent outgrowing clones.
• Montironi, R., Bartels, P. H., Thompson, D., Scarpelli, M., Bartels, H. G., Hamilton, P. W., Silva, V. D., Sakr, W. A., Weyn, B., & Daele, A. V. (2002). Transcontinental communication and quantitative digital histopathology via the internet; with special reference to prostate neoplasia. Journal of Clinical Pathology, 55(6), 452-460.
  More info

  PMID: 12037030;PMCID: PMC1769673;Abstract: Objective: To describe practical experiences in the sharing of very large digital data bases of histopathological imagery via the Internet, by investigators working in Europe, North America, and South America. Materials: Experiences derived from medium power (sampling density 2.4 pixels/μm) and high power (6 pixels/μm) imagery of prostatic tissues, skin shave biopsies, breast lesions, endometrial sections, and colonic lesions. Most of the data included in this paper were from prostate. In particular, 1168 histological images of normal prostate, high grade prostatic intraepithelial neoplasia (PIN), and prostate cancer (PCa) were recorded, archived in an image format developed at the Optical Sciences Center (OSC), University of Arizona, and transmitted to Ancona, Italy, as JPEG (joint photographic experts group) files. Images were downloaded for review using the Internet application FTP (file transfer protocol). The images were then sent from Ancona to other laboratories for additional histopathological review and quantitative analyses. They were viewed using Adobe Photoshop, Paint Shop Pro, and Imaging for Windows. For karyometric analysis full resolution imagery was used, whereas histometric analyses were carried out on JPEG imagery also. Results: The three applications of the telecommunication system were remote histopathological assessment, remote data acquisition, and selection of material. Typical data volumes for each project ranged from 120 megabytes to one gigabyte, and transmission times were usually less than one hour. There were only negligible transmission errors, and no problem in efficient communication, although real time communication was an exception, because of the time zone differences. As far as the remote histopathological assessment of the prostate was concerned, agreement between the pathologist's electronic diagnosis and the diagnostic label applied to the images by the recording scientist was present in 96.6% of instances. When these images were forwarded to two pathologists, the level of concordance with the reviewing pathologist who originally downloaded the files from Tucson was as high as 97.2% and 98.0%. Initial results of studies made by researchers belonging to our group but located in others laboratories showed the feasibility of making quantitative analysis on the same images. Conclusions: These experiences show that diagnostic teleconsultation and quantitative image analyses via the Internet are not only feasible, but practical, and allow a close collaboration between researchers widely separated by geographical distance and analytical resources.
• Montironi, R., Mazzucchelli, R., Colanzi, P., Streccioni, M., Scarpelli, M., Thompson, D., & Bartels, P. H. (2002). Improving inter-observer agreement and certainty level in diagnosing and grading papillary urothelial neoplasms: Usefulness of a Bayesian belief network. European Urology, 41(4), 449-457.
  More info

  PMID: 12074818;Abstract: Background and Objective: A Bayesian belief network (BBN), as diagnostic decision support system, enables the processing of our knowledge of histopathology expressed in descriptive terms, words and concepts. The aim of this study was to evaluate the contribution of a BBN in the improvement of inter-observer agreement and certainty level in the diagnosis and grading of papillary urothelial neoplasms. Materials: Inter-observer agreement and certainty level were investigated on 40 cases of non-invasive papillary urothelial neoplasms subdivided according to the WHO 1973 classification. There were 10 urothelial papillomas (UPs), 10 grade 1 papillary carcinomas (G1), 10 grade 2 papillary carcinomas (G2) and 10 grade 3 papillary carcinomas (G3). Five consecutive sessions were held with three observers (RMa, PC and MSt). Sessions A, B and D were based on the morphological evaluation of the specimens with a conventional light microscope only. In sessions C and E, a BBN was used in addition to the microscope. The BBN output was represented by four belief values for four possible diagnostic outcomes. These values ranged from 0.0 to 1.0, with the sum of the belief values being 1.0. Concerning the certainty level, a two-tier system of assessment was adopted in sessions A, B and D: certain versus less certain. In sessions C and E, a belief value equal to or greater than 0.65 was considered as equivalent to "certain". Results: In session A, an all-encompassing or synthetic approach to decision-making was adopted. Agreement with the gold standard was seen in 60% (RMa), 55% (PC) and 65% (MSt) of cases, respectively. The level of subjective confidence was "certain" in 35%, 40% and 35% of cases, respectively. Better agreement - 70% (RMa), 68% (PC) and 72% (MSt) of cases - was present in session B where an analytical approach based on the evaluation of a series of morphological features was used. The level of subjective confidence was "certain" in 45%, 50% and 55% of cases, respectively. In session C, where a BBN was utilised, a further increase in degree of agreement with the gold standard was observed, e.g. 85% (RMa), 80% (PC) and 86% (MSt) of cases, respectively. Levels of certainty or belief values were high. Decrease in both the level of agreement - 60% (RMa), 62% (PC) and 65% (MSt) of cases - and certainty was seen in session D where the observers were left free to evaluate the cases morphologically without the constrain of either a synthetic or analytical approach. In session E, where the BBN was used again, the percentage of cases in agreement with the gold standard increased to 83% (RMa), 81% (PC) and 84% (MSt), respectively. Increase in certainty or belief was also seen. The difference of the results obtained in the sessions A, B and D with those seen in the BBN-based sessions (C and E) is statistically significant. Conclusions: Conventional morphological evaluation of papillary urothelial neoplasms is affected by inter-observer variability and, in many instances, by diagnostic uncertainty. The greatest difficulties are found with G1 and G2 cases. Improvement in inter-observer agreement and certainty level can be achieved with a BBN. © 2002 Elsevier Science B.V. All rights reserved.
• Morrison, M. L., McCluggage, W. G., Price, G. J., Diamond, J., Sheeran, M. R., Mulholland, K. M., Walsh, M. Y., Montironi, R., Bartels, P. H., Thompson, D., & Hamilton, P. W. (2002). Herpesviruses in brain and Alzheimer's disease. Journal of Pathology, 197(3), 395-402.
  More info

  PMID: 12115887;Abstract: It has been established, using polymerase chain reaction (PCR), that herpes simplex virus type 1 (HSV1) is present in a high proportion of brains of elderly normal subjects and Alzheimer's disease (AD) patients. It was subsequently discovered that the virus confers a strong risk of AD when in brain of carriers of the type 4 allele of the apolipoprotein E gene (apoE-ε4). This study has now sought, using PCR, the presence of three other herpesviruses in brain: human herpesvirus 6 (HHV6)-types A and B, herpes simplex virus type 2 (HSV2) and cytomegalovirus (CMV). HHV6 is present in a much higher proportion of the AD than of age-matched normal brains (70% vs. 40%, p = 0.003) and there is extensive overlap with the presence of HSV1 in AD brains, but HHV6, unlike HSV1, is not directly associated in AD with apoE-ε4. In 59% of the AD patients' brains harbouring HHV6, type B is present while 38% harbour both type A and type B, and 3% type A. HSV2 is present at relatively low frequency in brains of both AD patients and normals (13% and 20%), and CMV at rather higher frequencies in the two groups (36% and 35%); in neither case is the difference between the groups statistically significant. It is suggested that the striking difference in the proportion of elderly brains harbouring HSV1 and HSV2 might reflect the lower proportion of people infected with the latter, or the difference in susceptibility of the frontotemporal regions to the two viruses. In the case of HHV6, it is not possible to exclude its presence as an opportunist, but alternatively, it might enhance the damage caused by HSV1 and apoE-ε4 in AD; in some viral diseases it is associated with characteristics brain lesions and it also augments the damage caused by certain viruses in cell culture and in animals. Copyright © 2002 John Wiley & Sons, Ltd.
• Bartels, P. H., Garcia, F. A., Davis, J., Silva, V. D., Bartels, H. G., Thompson, D., & Alberts, D. S. (2001). Progression curves for endometrial lesions. Analytical and Quantitative Cytology and Histology, 23(1), 1-8.
  More info

  PMID: 11233737;Abstract: OBJECTIVE: To derive a numeric measure for the progression of endometrial lesions as a baseline study for an eventual assessment of chemopreventive intervention efficacy. STUDY DESIGN: Tissue sections from normal endometrium at the proliferative and secretory phase, simple hyperplasia, atypical hyperplasia from cases free of concomitant adenocarcinoma and adenocarcinoma of the endometrium were recorded at high spatial resolution. Six cases from each diagnostic category were chosen as "typical," and 60 epithelial nuclei were randomly selected for measurement for each case. Discriminant analyses were carried out to derive a direction of progressive change in feature space and to correct the progression curve for the presence of cells not expressing progressive change among the random sample of nuclei. RESULTS: A well-conditioned progression curve was derived based on the mean discriminant function scores for each diagnostic category and the mean nuclear abnormality of the nuclei in each category, as expressed by their deviation in feature values from normal reference nuclei. The lesion signatures showed a clear trend toward extension into the range of higher nuclear abnormalities with increasing progression. There was an indication that abnormal endometrial lesions may comprise cases with distinctly different degrees of nuclear abnormality. CONCLUSION: A numeric assessment of lesion progression for endometrial lesions, based on karyometric measurements, is possible. The data suggest that additional analysis may provide further characterizing information for individual lesions.
• Bartels, P. H., Montironi, R. M., Bostwick, D., Marshall, J., Thompson, D., Bartels, H. G., & Kelley, D. (2001). Karyometry of secretory cell nuclei in high-grade PIN lesions. Prostate, 48(3), 144-155.
  More info

  PMID: 11494330;Abstract: BACKGROUND. The goal of this study was a karyometric characterization of secretory cell nuclei in high-grade prostatic intraepithelial neoplasia (PIN) lesions. Specifically, the hypothesis is tested that distinctly different subgroups of nuclei exist in these lesions. METHODS. High-resolution images of 1,713 nuclei from high-grade PIN lesions were recorded. Karyometric features were computed. Discriminant function scores against normal reference nuclei, and nuclear abnormality values were derived. Data sets were processed by a nonsupervised learning algorithm to establish the presence of subgroups of nuclei with statistically different nuclear chromatin distributions. RESULTS. Three sets of nuclei were formed, facing an intact basal cell layer, a near vanishing basal cell layer, and a gap in the basal cell layer. For each set, a nonsupervised learning algorithm formed three statistically different subgroups of approximately equal sizes. Each subgroup is found in every one of the three sampling locations. The total optical density distribution of nuclei in two subgroups suggests an aneuploid distribution, the third subgroup has a near diploid distribution. CONCLUSION. Secretory cell nuclei in high-grade PIN lesions are a heterogeneous population, forming statistically different subgroups. Studies aimed at characterizing the progression of such lesions should consider the inhomogeneous nature of these nuclei. © 2001 Wiley-Liss, Inc.
• Boone, C. W., Lieberman, R., Mairinger, T., Palcic, B., Bacus, J., & Bartels, P. (2001). Computer-assisted image analysis-derived intermediate endpoints. Urology, 57(4), 129-131.
  More info

  PMID: 11295610;Abstract: The development of prostatic lesions undergoes a slow progression. To establish efficacy of chemopreventive intervention it is therefore necessary to define surrogate endpoint biomarkers. Such biomarkers should be sensitive in their ability to indicate response. They should be objective, ie, the result of measurement, and numerically defined so that a statistical validation of response is possible. They should be able to indicate not only a halt of progression of a lesion, but also a reversal of progression. The spatial and statistical distribution of nuclear chromatin in the secretory and luminal cells in prostatic intraepithelial neoplastic lesions has been shown to be well defined. It can be represented by a set of features. These have been used to define a progression curve along which progression or regression of a lesion can be assessed. One could define a fixed endpoint, or one might choose to accept a statistically significant regression along the progression curve as criterion for chemopreventive efficacy. Expected difficulties could arise from lesion heterogeneity, as it would affect the sampling, and from multifocal lesions of differing progressions. Lesion heterogeneity thus limits the precision with which regression could be detected. These problems might be partially overcome by observations taken in histologically normal appearing regions of the prostate. The nuclear chromatin pattern of secretory cell nuclei measured in such tissue regions from prostates harboring intraepithelial or malignant lesions has been shown to exhibit distinctive changes from the chromatin pattern seen in secretory cell nuclei from prostates free from any such lesions. These changes appear to be expressed in the tissue up to a substantial distance from a lesion. The expression of changes in the nuclear chromatin suggests the existence of an intraepithelial preneoplastic lesion that can be detected by biomarkers, but which is not apparent from visual microscopic inspection. Since chemoprevention might be expected to be most effective at the earliest stages of lesion development, the assessment of such early alterations is seen as highly relevant to efforts to validate the efficacy of chemopreventive intervention. © 2001, Elsevier Science Inc.
• Bozzo, P., Alberts, D. S., Vaught, L., Silva, V. D., Thompson, D., Warnecke, J., Miller, R. C., Einspahr, J., & Bartels, P. H. (2001). Measurement of chemopreventive efficacy in skin biopsies. Analytical and Quantitative Cytology and Histology, 23(4), 300-312.
  More info

  PMID: 11531145;Abstract: OBJECTIVE: To explore methods suitable for quantitative assessment of the efficacy of chemopreventive intervention. STUDY DESIGN: High-resolution imagery of nuclei from the suprabasal and basal cell layers of sun-damaged skin were recorded. There were 10 cases. A shave biopsy was taken from an area of clearly evident solar keratosis before and after treatment with 2-difluoromethyl-dl-ornithine (DFMO) and from the colateral forearm, treated with a placebo. A number of karyometric variables were computed and combined to derive marker features that provided a numeric measure of the degree of nuclear deviation from normal. RESULTS: DFMO treatment was effective overall in reducing the degree of nuclear abnormality seen in the biopsies; in 8 of the 10 cases there was a significant improvement. The placebo-treated arm did not show a statistically different abnormality from the untreated arm. CONCLUSION: Karyometric analysis can provide numeric measures that allow documentation of statistically significant regression of actinic keratotic lesions following treatment with DFMO.
• Frank, D. H., Davis, J. R., Alberts, D. S., Thompson, D., Liu, Y., & Bartels, P. H. (2001). Nuclear chromatin characteristics of breast solid pattern ductal carcinoma in situ. Analytical and Quantitative Cytology and Histology, 23(6), 418-426.
  More info

  PMID: 11777277;Abstract: OBJECTIVE: To characterize nuclei from breast solid pattern ductal carcinoma in situ (DCIS) by their karyometric features and to search for the presence of statistically significantly different subsets of nuclei. STUDY DESIGN: One hundred nuclei from each of 6 normal, 13 solid DCIS, (9 low and intermediate grade and 4 high grade DCIS) histopathologic samples of breast tissue were digitally recorded. Karyometric features were computed and subjected to a nonsupervised learning algorithm (P-index) to identify significantly different subgroups. RESULTS: Nuclei in low grade lesions displayed a diploid/near diploid pattern, while the majority of intermediate grade lesions fell into a range beyond 5N. The high grade lesions showed substantial genomic instability and represented three statistically different subsets or phenotypes. CONCLUSION: There is a progression of nuclear abnormality from low grade to high grade DCIS. The nuclei from high grade DCIS form a heterogeneous set that represents three phenotypes. One of these phenotypes shows a nuclear chromatin pattern that more closely resembles poorly differentiated, infiltrating disease. The observation of such a phenotype may have prognostic implications.
• Garcia, F. A., Davis, J. R., Alberts, D. S., Hatch, K., Weyn, B., Thompson, D., & Bartels, P. H. (2001). Nuclear chromatin patterns in normal, hyperplastic and atypical endometrium. Analytical and Quantitative Cytology and Histology, 23(2), 144-150.
  More info

  PMID: 11332081;Abstract: OBJECTIVE: To characterize the nuclei of endometrial lesions for the diagnostic categories of normal glandular tissue, simple hyperplasia, atypical hyperplasia and adenocarcinoma of the endometrium, with the specific goal of probing for heterogeneity. STUDY DESIGN: For each diagnostic category the images of 360 nuclei were recorded on a high-resolution video microphotometer. Features descriptive of the statistical and spatial distribution of nuclear chromatin were computed for each nucleus. A nonsupervised learning algorithm, P-index, was employed to establish subsets of nuclei within each diagnostic category and to determine whether these subsets were statistically significantly different in the nuclear chromatin pattern. RESULTS: Lesions from cases of hyperplasia, atypical hyperplasia and adenocarcinoma of the endometrium each contained several subsets of nuclei with statistically significantly different chromatin patterns. For one such subset from each diagnostic category, a clear trend of progression toward adenocarcinoma could be demonstrated. CONCLUSION: The nuclei in endometrial lesions represent a highly heterogeneous set. Any measure of lesion progression or regression due to chemopreventive intervention, in an individual case, will have to examine the proportion of nuclei in each of these subsets as well as measures of deviation from normal for each subset.
• Mombello, A., Mariuzzi, L., Morelli, L., Granchelli, G., Rucco, V., Tarocco, E., Silva, V. d., Thompson, D., Bartels, H. G., Bartels, P. H., & Mariuzzi, G. (2001). Quantitative study of ductal breast cancer progression: nuclear signatures for evaluation of progression grade.. Advances in clinical pathology : the official journal of Adriatic Society of Pathology, 5(3), 59-70.
  More info

  PMID: 11753877;Abstract: The evaluation of progressive morphological changes, with 93 morphometric parameters in tissue lesions representative of ductal breast cancer progression, has been performed in order to define in great detail the profile of chromatin texture (nuclear signature) changes. A gradual, distinctive increase in nuclear signature alterations from hyperplasia to infiltrating carcinoma has been found. The nuclear signatures' analysis of microinfiltrating foci in comedo DCIS showed sharp differences compared with those of comedo DCIS they derived from: these foci consist of cells with smaller and also more homogeneous nuclei. Opposite to the prominent heterogeneity of those of comedo DCIS: they appear to express a reduced clonality in the new, more progressed, cell population. Digital analysis of chromatin patterns seems to be useful, beyond mere extraction of individual features of value, in getting objective data for individual grading and prognosis of breast cancer.
• Silva, V. D., Prolla, J. C., Sharma, P., Sampliner, R., Thompson, D., & Bartels, P. H. (2001). Karyometry in Barrett's esophagus. Analytical and Quantitative Cytology and Histology, 23(1), 40-46.
  More info

  PMID: 11233742;Abstract: OBJECTIVE: To derive a progression curve for lesions in Barrett's esophagus based on karyometric features. STUDY DESIGN: High-resolution imagery of 900 nuclei from normal gastric tissue, Barrett's metaplasia, Barrett's high grade dysplasia and adenocarcinoma of the esophagus was recorded. Karyometric features were computed, and nuclear signatures and lesion signatures for these lesions were derived. A progression curve was defined. RESULTS: Esophageal lesions were distinctly different from the normal gastric fundus tissue, with nuclei from Barrett's metaplasia deviating from normal almost as much as nuclei from high grade dysplasia and adenocarcinoma. There was considerable case-to-case variability and overlap between lesions histologically assigned to different diagnostic categories. CONCLUSION: The karyometric data suggest that Barrett's metaplasia is a more developed lesion than previously assumed.
• Tarocco, E., Mariuzzi, L., Bettini, R., Bartels, P. H., & Mariuzzi, G. (2001). Early stromal invasion in cervical cancer: immunocytochemical changes occurring in infiltrating neoplastic cells.. Advances in clinical pathology : the official journal of Adriatic Society of Pathology, 5(4), 133-138.
  More info

  PMID: 17582937;Abstract: Early Stromal Invasion (ESI) in cervical cancer progression should be considered as a separate histological diagnostic category for its morphological characters very different from those of both carcinoma in situ (CIS) and microcarcinoma (MIC). To have some more microscopical details on these differences we performed immunocytochemical investigation addressed to evaluate, in cervical cancer malignancy progression, the evolutionary changes in the expression of some proteins involved in cell differentiation and cell cycle regulation. The results provide data improving the knowledge about ESI and supporting, with objective proofs, the nosological autonomy of ESI, with respect to CIS and MIC.
• Thompson, D., Richards, D., Bartels, H., Montironi, R., Scarpelli, M., Hamilton, P. W., & Bartels, P. H. (2001). Multimegapixel images in histopathology. Analytical and Quantitative Cytology and Histology, 23(3), 169-177.
  More info

  PMID: 11444185;Abstract: OBJECTIVE: To describe methods and procedures for the assembly of very large scale microscopic image arrays. STUDY DESIGN: Microscopic imagery was recorded on different video microphotometers, equipped either with a three-chip CCD Sony MD 760 (Parkridge, New Jersey, U.S.A.), a COHU vidicon (San Diego, California, U.S.A.) or a PROGRES camera (JenOptik, Jena, Germany), yielding image tiles of 512×470, 512×470 or 1,496×1,120 pixels, respectively. The slide was moved while mounted on a Maerzheuser scanning stage with 0.1-μm precision, under computer control. The MERGE software (Optical Sciences Center, University of Arizona, Tucson, Arizona, U.S.A.) was written in C and currently implemented on a Sun Ultra Sparc 2 computer (Sun Microsystems, Palo Alto, California, U.S.A.). RESULTS: The MERGE program allows the assembly of very large scale digitized image arrays preserving exact tile alignment such that even within a single nucleus, highly precise registration is maintained. Images up to 150 megapixels have been assembled, although most practical applications required assembly of only 60-300 tiles. CONCLUSION: The single limiting effect of assembling very large image arrays is the problem of angular misalignment between CCD scan line orientation and scanning stage travel direction. For misalignment of even less than 1°, very large arrays need substantial tile overlap. For object areas extending over only 5-10 mm, the effects can be controlled.
• Bartels, P. H. (2000). Automated primary screening devices: Expectations for the next generation. Acta Cytologica, 44(5), 703-708.
  More info

  PMID: 11015970;
• Bartels, P. H. (2000). Future directions in quantitative pathology: Digital knowledge in diagnostic pathology. Journal of Clinical Pathology, 53(1), 31-37.
  More info

  PMID: 10767853;PMCID: PMC1731072;
• Bartels, P. H., Thompson, D., Montironi, R., Hamilton, P. W., Mariuzzi, G. M., & Silva, V. D. (2000). Digital knowledge and diagnostic information. Journal of Histotechnology, 23(3), 183-190.
  More info

  Abstract: The long term objective of this research was the development of objective, digitally defined procedures for histopathologic assessment. The development of procedures based on digital knowledge had as its first aim the design of a machine vision system with image understanding capability (ie, capable of autonomous processing and analyses of histopathologic imagery). Next, histometric and karyometric diagnostic information extraction led to highly specific characterization of nuclei and lesions. Based on such detailed characterizations, we were able to derive progression curves for prostatic, colonic, breast epithelial, and esophageal lesions. The specific signatures of nuclei and lesions revealed substantial diversity among lesions of the same visual-diagnostic grade; profiles of deviation of nuclei from a normal standard were derived to provide a novel, additional level of diagnostically discriminating features. Knowledge guided machine vision opens the way to an extremely specific characterization of nuclei and lesions, which may allow better prediction of biological behavior and, thus, more accurate individual patient targeted prognosis.
• Keenan, S. J., Diamond, J., McCluggage, W. G., Bharucha, H., Thompson, D., Bartels, P. H., & Hamilton, P. W. (2000). An automated machine vision system for the histological grading of cervical intraepithelial neoplasia (CIN). Journal of Pathology, 192(3), 351-362.
  More info

  PMID: 11054719;Abstract: The histological grading of cervical intraepithelial neoplasia (CIN) remains subjective, resulting in inter- and intra-observer variation and poor reproducibility in the grading of cervical lesions. This study has attempted to develop an objective grading system using automated machine vision. The architectural features of cervical squamous epithelium are quantitatively analysed using a combination of computerized digital image processing and Delaunay triangulation analysis; 230 images digitally captured from cases previously classified by a gynaecological pathologist included normal cervical squamous epithelium (n = 30), koilocytosis (n = 46), CIN 1 (n = 52), CIN 2 (n = 56), and CIN 3 (n=46). Intra- and inter-observer variation had kappa values of 0.502 and 0.415, respectively. A machine vision system was developed in KS400 macro programming language to segment and mark the centres of all nuclei within the epithelium. By object-oriented analysis of image components, the positional information of nuclei was used to construct a Delaunay triangulation mesh. Each mesh was analysed to compute triangle dimensions including the mean triangle area, the mean triangle edge length, and the number of triangles per unit area, giving an individual quantitative profile of measurements for each case. Discriminant analysis of the geometric data revealed the significant discriminatory variables from which a classification score was derived. The scoring system distinguished between normal and CIN 3 in 98.7% of cases and between koilocytosis and CIN 1 in 76.5% of cases, but only 62.3% of the CIN cases were classified into the correct group, with the CIN 2 group showing the highest rate of misclassification. Graphical plots of triangulation data demonstrated the continuum of morphological change from normal squamous epithelium to the highest grade of CIN, with overlapping of the groups originally defined by the pathologists. This study shows that automated location of nuclei in cervical biopsies using computerized image analysis is possible. Analysis of positional information enables quantitative evaluation of architectural features in CIN using Delaunay triangulation meshes, which is effective in the objective classification of CIN. This demonstrates the future potential of automated machine vision systems in diagnostic histopathology. Copyright (C) 2000 John Wiley and Sons, Ltd.
• Mariuzzi, L., Mombello, A., Rucco, V., Morelli, L., Zamò, A., Thompson, D., Vaught, L., Bartels, H. G., Mariuzzi, G., & Bartels, P. H. (2000). Quantitative study of ductal breast cancer progression: signatures of nuclei in proliferating breast lesions and in situ cancer.. Advances in clinical pathology : the official journal of Adriatic Society of Pathology, 4(2), 87-97.
  More info

  PMID: 11080789;Abstract: AIMS: The objective of this study is to derive highly specific nuclear signatures (NS's) for the characterization of nuclei of ductal breast epithelium in proliferative lesions and in situ cancers in order to evaluate if nuclear structural changes are able to describe the main events of ductal cancer progression and if the method can be used for objective grading. METHODS: A total of 82 different features descriptive of the nuclear chromatin patterns were computed in nuclei from normal glandular breast tissue, florid hyperplasia, and ductal carcinoma in situ (DCIS) and of DCIS with microinfiltration. The feature values were arranged to form a profile or signature. Measures of difference to a standard profile derived from normal glandular breast tissue were defined. One may then compute a standardized distance measure for a nucleus from "normal". Lesions can be characterized in the same manner, on the basis of the mean profile for all of their nuclei, and on the basis of the distribution of distances of their constituent nuclei from normal. RESULTS: The selected histopathologic patterns on which the diagnostic categories for DCIS are based were found to have corresponding distinctive patterns in the chromatin of the lesion's nuclei. A monotonic trend of ductal neoplastic progression was found. In addition, lesions histologically assessed as belonging to the same diagnostic category were found to offer substantially different distribution patterns. CONCLUSIONS: The full utilization of nuclear texture features allows the derivation of highly specific signatures for nuclei so that a reproducible grading can be performed for prognostic purposes.
• Montironi, R., Mazzucchelli, R., Stramazzotti, D., Pomante, R., Thompson, D., & Bartels, P. H. (2000). Expression of π-class glutathione s-transferase: Two populations of high grade prostatic intraepithelial neoplasia with different relations to carcinoma. Journal of Clinical Pathology - Molecular Pathology, 53(3), 122-128.
  More info

  PMID: 10897330;PMCID: PMC1186917;Abstract: Background/Aims: Patients with high grade prostatic intraepithelial neoplasia of the transition zone appear to be at increased risk of developing prostatic carcinoma, although not to the same degree as patients with high grade prostatic intraepithelial neoplasia of the peripheral/central zone. Previous investigations have shown loss of expression of π-class glutathione S-transferase (GST-π; an enzyme that protects against electrophilic carcinogens) in prostatic carcinoma and in high grade prostatic intraepithelial neoplasia. The aim of this study was to compare the expression of GST-π in high grade prostatic intraepithelial neoplasia of the transition zone with that in high grade prostatic intraepithelial neoplasia of the peripheral/central zone (that is, non-transition zone). Methods: Immunostaining with the anti-GST-π antibody was performed on 20 high grade prostatic intraepithelial neoplasia samples of the transition zone, either isolated or associated with prostatic carcinoma (groups 1 and 2, respectively; 10 cases each) and on 20 high grade prostatic intraepithelial neoplasia samples of the non-transition zone, either isolated or associated with prostatic carcinoma (groups 3 and 4, respectively; 10 cases each). This study also included six samples of high grade prostatic intraepithelial neoplasia simultaneously present in the transition and non-transition zones and not associated with prostatic carcinoma (group 5). The presence of immunostaining, staining intensity, and the distribution of immunostaining were evaluated in the high grade prostatic intraepithelial neoplasia lesions and in the normal tissue and cancer areas. Results: The GST-π antibody stained the cytoplasm of the cells lining the ducts and acini of normal prostate tissue. Staining was stronger and more diffuse in the basal cell layer than in the luminal (or secretory) cell layer. Immunohistochemical staining with anti-GST-π antibodies failed to detect the enzyme in all prostatic carcinoma foci but one. Two patterns were detected in high grade prostatic intraepithelial neoplasia. One was represented by GST-π staining similar to that of the normal tissue (pattern A). The other deviated from it and was characterised by absence of GST-π expression in the secretory cells and abundant expression in scattered basal cells (pattern B). Pattern A staining was seen more frequently in the transition than in the non- transition zone. Pattern B staining was seen mainly in high grade prostatic intraepithelial neoplasia of non-transition zone associated with cancer. Conclusions: The differential expression of GST-π in the transition and non- transition zones indicates the existence of two populations with the morphological appearance of high grade prostatic intraepithelial neoplasia that might have different associations with carcinoma.
• Protopapa, E., Delides, G., Miaoulis, G., Thompson, D., & Bartels, P. H. (2000). Image analysis of mesothelioma II. Discrimination of mesothelioma from metastatic serous ovarian adenocarcinoma. Analytical and Quantitative Cytology and Histology, 22(4), 338-345.
  More info

  PMID: 10965411;Abstract: OBJECTIVE: To examine the utility of karyometric measurements in the differentiation of mesothelioma from metastatic serous ovarian adenocarcinoma. STUDY DESIGN: High-resolution images of 1,631 nuclei from 32 cases of mesothelioma and 742 nuclei from 15 cases of ovarian adenocarcinoma were recorded. A stepwise discriminant analysis and nonparametric classifier were applied. RESULTS: Nuclei from these two diagnostic categories appear very similar and occupy feature space with significant overlap. A nonparametric classification procedure provided acceptable correct classification. CONCLUSION: For certain regions in feature space, cases could be unequivocally classified.
• Protopapa, E., Delides, G., Miaoulis, G., Thompson, D., & Bartels, P. H. (2000). Image analysis of mesothelioma: I. Differentiation of mesothelioma from adenocarcinoma of the lung. Analytical and Quantitative Cytology and Histology, 22(2), 114-122.
  More info

  PMID: 10800612;Abstract: OBJECTIVE: To explore the usefulness of nuclear micromorphometric analysis for the differentiation between epithelial mesothelioma and metastatic adenocarcinoma in the chest wall. STUDY DESIGN: High-resolution images of 2,100 nuclei from 27 cases of epithelial mesothelioma and 15 cases of adenocarcinoma of the lung were recorded. Stepwise discriminant analysis and a nonparametric classifier were applied to derive estimates for a case diagnosis correct classification rate. RESULTS: Nuclei from epithelial mesothelioma and adenocarcinoma of the lung showed statistically significantly different properties, but there was a region of overlap in feature space such that approximately 15-20% of cases could not be correctly classified. The lesion signatures derived from the mesothelioma cases with discriminant function scores that might result in case misclassification and the cases of adenocarcinoma of the lung spanned a similar range of degree of nuclear abnormality. However, the distribution of nuclear abnormality values for the mesothelioma cases has a mode at 0.87 SD from normal, whereas the distribution seen in lung adenocarcinoma cases had a mode at about 3.7 SD. CONCLUSION: Cases of epithelial mesothelioma and adenocarcinoma of the lung have nuclei with a wide range of deviation from normal in the spatial and statistical distribution of their nuclear chromatin. For approximately 80% of cases, correct case classification can be provided by nuclear micromorphometric analysis. Cases of epithelial mesothelioma with highly abnormal nuclei overlap in feature space with nuclei from adenocarcinoma of the lung. However, it is possible that characterization by a lesion signature may allow correct assignment for those cases.
• Scarpelli, M., Baccarini, M. G., Colanzi, P., Arnaldi, G., Montironi, R., Thompson, D., & Bartels, P. H. (2000). Chromatin texture analysis of cortical adrenal gland adenomas, including incidentalomas, and adjacent normal-appearing cortical tissue. Analytical and Quantitative Cytology and Histology, 22(3), 235-243.
  More info

  PMID: 10872041;Abstract: OBJECTIVE: To describe, by morphometric and chromatin texture analysis, a series of adrenal gland lesions, including Cushing's and Conn's adenomas and incidentalomas. STUDY DESIGN: The material for the study consisted of five consecutive cases of incidentaloma, three cases of Conn's adenoma and three cases of Cushing's adenoma. Also included were five cases of adrenal carcinoma. Sections were stained according to the Feulgen procedure. Measurements were taken from the nodules and from two different zones, identified as outer and inner parts, of the normal-appearing adrenal cortex adjacent to the tumor. Data on approximately 50 nuclei were recorded for each of these three sites (tumor and outer and inner normal-appearing adrenal cortex). The nuclei were subjected to feature extraction and were analyzed by identification procedures - i.e., establishing nuclear and lesion signatures. RESULTS: The total optical density (OD) distributions of the nuclei from the normal-appearing adrenal cortex pointed to their diploid or near-diploid nature. In incidentalomas there was a very small increase in the number of nuclei, with increased total OD. In Conn's adenoma there was a noticeable but modest extension of the total OD distribution into the higher OD range. This trend continued for Cushing's adenoma. The pixel OD histograms for nuclei from normal-appearing tissue and from incidentalomas were hardly distinguishable. Starting with nuclei from Conn's adenoma, a shift toward lower pixel OD values began. The trend continued for nuclei from Cushing's adenoma and was very pronounced for nuclei from carcinoma. The nuclear signatures showed no appreciable difference between nuclei from normal-appearing cortex and from incidentaloma. Nuclei from Conn's adenoma were more similar to those from normal tissue in their signatures than nuclei from Cushing's adenoma. In fact, the nuclear signatures from Cushing's adenoma were almost identical to those of carcinoma. The lesion signatures for normal tissue, incidentaloma and Conn's adenoma confirmed the results seen in the nuclear signatures. There was a very modest increase in the number of nuclei with greater deviation from normal in incidentalomas, and the trend was more obvious in Conn's adenoma. However, in Cushing's adenoma there was a very substantial increase in the number of nuclei, with large deviations of their nuclear chromatin texture from normal. CONCLUSION: Computer-assisted analysis of nuclear characteristics proved useful in identifying and describing differences between groups of tumors arising in the adrenal cortex and highlighted the similarity between incidentalomas and adjacent normal-appearing cortex and between Cushing's adenoma and adrenal carcinoma.
• Weyn, B., Jacob, W., D., V., Montironi, R., Hamilton, P. W., Thompson, D., Bartels, H. G., Daele, A. V., Dillon, K., & Bartels, P. H. (2000). Data representation and reduction for chromatin texture in nuclei from premalignant prostatic, esophageal, and colonic lesions. Cytometry, 41(2), 133-138.
  More info

  PMID: 11002269;Abstract: Background: To identify nuclei and lesions with great specificity, a large set of karyometric features is arranged in the form of a linear profile, called a nuclear signature. The karyometric feature values are normalized as z-values. Their ordering along the profile axis is arbitrary but consistent. The profile of the nuclear signature is distinctive; it can be characterized by a new set of variables called contour features. A number of data reduction methods are introduced and their performance is compared with that of the karyometric features in the classification of prostatic, colonic, and esophageal lesions. Methods: Contour characteristics were reduced to descriptive statistics of the set of z-values in the nuclear signature and to sequence information. The contour features derived were (1) relative frequencies of occurrence of z-values and of their differences and (2) co-occurrence statistics, run lengths of z-values, and statistics of higher-order dependencies. Performance was evaluated by comparing classification scores of diagnostic groups. Results: Rates for correct classification by karyometric features alone and contour features alone indicate equivalent performance. Classification by a combined set of features led to an increase in correct classification. Conclusions: Image analysis and subsequent data reduction of nuclear signatures of contour features is a novel method, providing quantitative information that may lead to an effective identification of nuclei and lesions. (C) 2000 Wiley-Liss, Inc.
• Anand, S. S., Smith, A. E., Hamilton, P. W., Anand, J. S., Hughes, J. G., & Bartels, P. H. (1999). An evaluation of intelligent prognostic systems for colorectal cancer. Artificial Intelligence in Medicine, 15(2), 193-214.
  More info

  PMID: 10082181;Abstract: In this paper we describe attempts at building a robust model for predicting the length of survival of patients with colorectal cancer. The aim of the research is to study the effective utilisation of artificial intelligence (AI) techniques in the medical domain. We suggest that an important research objective of proponents of intelligent prognostic systems must be to evaluate the additionality that AI techniques can bring to an already well-established field of medical prognosis. We compare a number of different AI techniques that lend themselves to the task of predicting survival in colorectal cancer patients. We describe the pros and cons of each of these methods, present the notion of intelligent hybrid systems and evaluate the role that they may potentially play in developing robust prognostic models. We describe a number of innovations used within this hybrid paradigm. In keeping with our objective of studying the additionality that AI techniques bring to building prognostic models, we use Cox's regression as a standard and compare each AI technique with it.
• Bartels, P. H., & Vooijs, G. P. (1999). Automation of primary screening for cervical cancer. Sooner or later?. Acta Cytologica, 43(1), 7-12.
  More info

  PMID: 9987443;
• Bartels, P. H., Montironi, R., Duval, V., Hamilton, P. W., Thompson, D., Vaught, L., & Bartels, H. G. (1999). Tissue architecture analysis in prostate cancer and its precursors: An innovative approach to computerized histometry. European Urology, 35(5-6), 484-491.
  More info

  PMID: 10325510;Abstract: Aims: It is the aim of these studies to derive a numerically defined progression index for prostatic intraepithelial neoplasia (PIN) lesions. Methods: Histometric and karyometric features were automatically extracted from images of histopathologic sections by a machine vision system. Results: Both histometric and karyometric measures lend themselves to the defining of a progression index. Karyometric features were found to be more sensitive. They allow the detection of very early change. Conclusions: It is possible to measure progression of PIN lesions with precision. The methodology would lend itself for measurement of regression due to chemopreventive intervention.
• D., V., Montironi, R., Thompson, D., Bartels, H. G., Vaught, L., Hamilton, P. W., & Bartels, P. H. (1999). Chromatin texture in high grade prostatic intraepithelial neoplasia and early invasive carcinoma. Analytical and Quantitative Cytology and Histology, 21(2), 113-120.
  More info

  PMID: 10560478;Abstract: OBJECTIVE: To evaluate individual nuclei from high grade prostatic intraepithelial neoplasia (PIN) lesions with early invasive carcinoma foci in the area of microinvasion and in the gland in which the microinvasion originated. STUDY DESIGN: High-resolution, digitized images of nuclei from defined locations were recorded and segmented, and karyometric variables were computed. These included a set of 93 features, which form a nuclear signature characterizing the spatial and statistical distribution of the nuclear chromatin. Nuclei in the glandular epithelium were recorded sequentially, along the basal cell layer, at increasing distances from the point of microinvasion and by random selection in the region of microinvasion. RESULTS: At a distance >60 nuclear locations from the point of microinvasion, the nuclear signatures corresponded to those seen in high grade PIN. Between 40 and 20 nuclear locations removed from the microinvasion focus the signatures began to change gradually until at a distance of 15-5 locations they strongly resembled the signatures seen in adenocarcinoma. The total optical density decreased to values seen in adenocarcinoma, and the nuclear chromatin had finer granularity. While nuclei in high grade PIN followed a widely dispersed total optical density distribution suggestive of wide- ranging aneuploidy, the nuclei in the region of microinvasion exhibited a less dispersed and bimodal total optical density distribution. CONCLUSION: The chromatin texture signatures showed a clear trend: there was an obvious attenuation as the measured nuclei approached the microinvasion area. The decrease in total optical density at the microinvasion might suggest the emergence of one or two clones that can be responsible for the invasive phenotype.
• Hamilton, P. W., Bartels, P. H., Anderson, N., Thompson, D., Montironi, R., & Sloan, J. M. (1999). Case-based prediction of survival in colorectal cancer patients. Analytical and Quantitative Cytology and Histology, 21(4), 283-291.
  More info

  PMID: 10560505;Abstract: OBJECTIVE: To develop an approach to the prediction of survival in patients with colorectal cancer using nearest neighbor analysis and case- based reasoning. STUDY DESIGN: A total of 216 patients with full clinicopathologic records and five-year follow-up were the subjects of this study. They were divided into a core database of 162 cases and a test group of 54 cases, with follow-up on all patients. When the patient was still alive at the end of the follow-up period, censored survival time was used. For each of the test cases, the four closest neighbors from the database were retrieved and their median survival time recorded and used as the predicted estimate of survival. Case matching was based on a Euclidean multivariate distance measure for the three best predictor variables: patient age, Dukes stage and tubule configuration. Cases with the smallest distance from the test case were considered to be the most similar. The predicted survival times for the test cases were compared with the actual, observed survival in the test cases to determine the success of this approach. RESULTS: The results showed reasonable concordance between observed and predicted survival figures, although there was a large degree of spread. Classification of cases into ≤ 60 and > 60 months' survival showed a correct classification rate of 63%. For the prediction of survival time, the distribution of differences between observed and predicted survival times for the uncensored test cases had a median value of -5 months but also showed a wide dispersion of values. Correlation of observed and predicted survival times, while not reaching statistical significance at P < .05, did show a strong positive association. CONCLUSION: Case-based approaches to the prediction of survival times in cancer patients are important. The results of the current study illustrate the difficulties in applying this approach to survival data and highlight the complexity of patient information and the inability to accurately predict patient outcome on a small subset of clinicopathologic features. While extensive work needs to be carried out to improve prediction power, this study illustrates the potential for case-based analyses. The ability to retrieve feature-matched cases from hospital patient databases has clear, independent advantages in patient management, but the ability to provide reliable, targeted prognostic estimates on individual cases should be a common goal in medical research.
• Montironi, R., Hamilton, P. W., Scarpelli, M., Thompson, D., & Bartels, P. H. (1999). Subtle morphological and molecular changes in normal-looking epithelium in prostates with prostatic intraepithelial neoplasia or cancer. European Urology, 35(5-6), 468-473.
  More info

  PMID: 10325507;Abstract: Background: Prostate cancer develops over an extended period of time. Until recently, the events initiating the process and the developments concomitant with the evolution towards invasive disease were largely unknown. Methods: Analytical and quantitative methods are applied to provide insights into certain individual molecular events and their effects on the complex multiple feedback system of cellular metabolism and regulation in prostate neoplasia. Results: Prostatic intraepithelial neoplasia (PIN) and prostate cancer (PCa) are associated with or possibly preceded by changes in the chromatin of secretory cell nuclei. The changes are detectable with a Bayesian belief network and quantifiable by computer image analysis in prostatic tissue that still appears histologically normal. In addition, normal-looking prostate epithelium shows some molecular changes similar to those present in the associated preneoplastic and neoplastic lesions. Such changes are also occasionally present in normal prostate glands without PIN and cancer. Conclusions: The subtle morphological and molecular changes of normal-looking epithelium might be seen as the onset of the development of prostatic neoplasia.
• Montironi, R., Mazzucchelli, R., Marshall, J. R., & Bartels, P. H. (1999). Prostate cancer prevention: Review of target populations, pathological biomarkers, and chemopreventive agents. Journal of Clinical Pathology, 52(11), 793-803.
  More info

  PMID: 10690166;PMCID: PMC501588;
• Montironi, R., Mazzucchelli, R., Pomante, R., Thompson, D., Duval, V., Vaught, L., & Bartels, P. H. (1999). Immunohistochemical expression of π class glutathione S-transferase in the basal cell layer of benign prostate tissue following chronic treatment with finasteride. Journal of Clinical Pathology, 52(5), 350-354.
  More info

  PMID: 10560354;PMCID: PMC1023070;Abstract: Background-Glutathione S-transferases (GST) may prevent carcinogenesis through inactivation of reactive electrophiles by conjugation to reduced glutathione. Treatment directed at the induction or preservation of GST-π expression in normal epithelium could have a profound impact on the prevention of prostate neoplasia. Finasteride, a 5-α-reductase inhibitor, is used as a chemopreventive agent because it blocks the conversion of testosterone to its byproduct which promotes prostate tumour growth. Objective-To investigate GST-π expression immunohistochemically in benign prostate tissue from untreated patients and from patients chronically treated with finasteride. Materials-Immunostaining with anti-GST-π antibody was performed on 10 (cysto-) prostatectomy, eight simple prostatectomy, and three transurethral prostatectomy specimens. The first set of 10 prostates was from untreated patients operated on for bladder cancer. The other cases were from patients with benign prostatic hyperplasia and chronically treated with finasteride. None of the specimens in either group showed prostatic cancer, prostatic intraepithelial neoplasia, urothelial carcinoma, or chronic prostatitis. Specimens were evaluated for the presence, intensity, and distribution of immunostaining. Results-Diffuse cytoplasmic immunostaining was observed in the basal cell layer of the untreated specimens. Some variability in the expression of GST-π was seen within each zone and also between the prostate zones. Only a minority of the secretory cells was stained weakly, mainly in the subnuclear region of the cells facing an uninterrupted basal cell layer. Staining was more homogeneously diffuse in the cytoplasm of the luminal cells facing the basement membrane directly. In the benign epithelium of the finasteride treated specimens the circumferential staining of the basal cells appeared to be more continuous than in the untreated cases, the gaps in the stained basal cell layer being fewer, shorter, or even absent in some ducts and acini. There was no variability in the intensity of staining of the basal cell layer, all the cells being intensely stained in a uniform way. The intensity of staining of the secretory cells was not influenced by finasteride treatment. Conclusions- Following chronic treatment with finasteride the immunohistochemical expression of π class glutathione S-transferase in the benign prostate ducts and acini is upregulated in relation to an expanded basal cell layer. This could indicate that finasteride acts as a GST-π inducer.
• Montironi, R., Santinelli, A., Pomante, R., Hamilton, P. W., Thompson, D., & Bartels, P. H. (1999). Inference network-based analyses of the histopathological effects of androgen deprivation on prostate cancer. Journal of Pathology, 187(4), 462-468.
  More info

  PMID: 10398107;Abstract: The evaluation of prostate cancer histology following hormonal therapy often represents a diagnostic problem for the pathologist. Previous studies have shown that an inference or Bayesian belief network (BBN) offers a descriptive classifier useful for the accurate analysis of morphological changes in individual cases of prostate neoplasia. Three different BBNs were evaluated in 94 cancer foci present in 20 radical prostatectomy (RP) specimens and in the matching biopsies in which the initial diagnosis of prostatic adenocarcinoma was made. Ten PR specimens were from patients treated with total androgen ablation or combination endocrine therapy (CET) before surgery. The first and second BBN allowed the identification with high certainly of the cancer foci present in the biopsies and RP specimens, as well as their Gleason grade, the belief value often being close to 1·0. The results of the second BBN showed a good correspondence between the Gleason grade given in the biopsies and that in the RP specimens, except in the surgical material of the treated patients, in which upgrading was always present. The third BBN showed the existence of three subgroups in treated RP specimens, one with morphological effect, another with poor effect, and the third with the histology of untreated (i.e. unaffected) cancer. In conclusion, an inference network-based analysis allows the characterization of treated prostate cancers according to the degree of histopathological change.
• Scarpelli, M., Montironi, R., Mazzucchelli, R., Thompson, D., & Bartels, P. H. (1999). Distinguishing cortical adrenal gland adenomas from carcinomas by their quantitative nuclear features. Analytical and Quantitative Cytology and Histology, 21(2), 131-138.
  More info

  PMID: 10560481;Abstract: OBJECTIVE: To explore data from a set of cases of adrenal cortical adenomas with different endocrine syndromes and carcinomas to determine whether quantitative image analysis of nuclear features might be used to separate the groups. STUDY DESIGN: Fifteen adrenal cortical tumors in which clinical information and optimally preserved, paraffin-embedded tissue were available were used. There were 10 adenomas and 5 carcinomas. Among the adenomas, five were associated with primary hyperaldosteronism (Conn's syndrome) and five with Cushing's syndrome. Five-micrometer-thick sections were stained with hematoxylin and eosin. In each case 50 nuclei were measured, and a number of morphometric and densitometric features were extracted. The data were subjected to multivariate analysis. RESULTS: Quantitative analysis showed that nuclei from adrenal carcinomas are larger than those from adenomas. Total optical density (OD) had a near-diploid distribution in the adenomas, while it was clearly aneuploid in the carcinomas. The pixel OD histograms were almost identical for all groups. Differences in nuclear chromatin texture were found between adenomas and carcinomas and also between the two adenoma categories. Multivariate analysis showed good discrimination between carcinomas and adenomas (Wilks lambda = .628, P < .0001) and between adenomas. However, based on Bayesian decision boundaries, 20-25% of carcinoma nuclei could be expected to be in the range of adenoma, and about 12% of Cushing's adenoma nuclei and 15% of Conn's adenoma nuclei would be classified as carcinoma. CONCLUSION: Computer- assisted analysis of nuclear characteristics proved useful in identifying and describing differences between groups of tumors arising in the adrenal cortex.
• Bartels, P. H., Bartels, H. G., Montironi, R., Hamilton, P. W., & Thompson, D. (1998). Machine vision in the detection of prostate lesions in histologic sections. Analytical and Quantitative Cytology and Histology, 20(5), 358-364.
  More info

  PMID: 9801753;Abstract: OBJECTIVE: To explore the utility of N-gram encoding for the automated detection and delineation of regions of histologic abnormality in tissue sections of prostate. STUDY DESIGN: Digitized imagery of tissue sections from normal prostate glandular tissue, stroma and regions of well- and poorly differentiated lesions was recorded and successively subdivided into square subregions of 256 x 256 to 16 x 16 pixels. N-grams of N = 2 to N = 6 were computed, with each element assuming a value representing an optical density interval 0.30 units wide, covering the range from optical density = 0.0 to 1.80. Then, from a large database, prototype frequency histograms of the different N-grams were established. For each subregion the Euclidean distances to the different prototype histograms were computed and defined as 'distance to prototype' features. Standard discriminant analyses and a nonparametric classifier were used to assign subregions to the different tissue categories. RESULTS: Classification of subregions was achieved for most discrimination tasks at a correct recognition rate ranging from 85% to 100% on both training set and test set data, with a few exceptions. N-grams of N > 4 had considerable discriminatory power. CONCLUSION: N-gram encoding has the potential to provide highly discriminating, texture-based characterization of subregions of digitized imagery of prostate lesions and may be very useful in the development of decision procedures for the automated detection of prostate lesions by a machine vision system.
• Bartels, P. H., Bibbo, M., Hutchinson, M. L., Gahm, T., Grohs, H. K., Gwi-Mak, E., Kaufman, E. A., Kaufman, R. H., Knight, B. K., Koss, L. G., Magruder, L. E., Mango, L. J., McCallum, S. M., Melamed, M. R., Peebles, A., Richart, R. M., Robinowitz, M., Rosenthal, D. L., Sauer, T., , Schenck, U., et al. (1998). Computerized screening devices and performance assessment: Development of a policy towards automation: IAC task force summary. Acta Cytologica, 42(1), 59-68.
  More info

  PMID: 9479324;Abstract: Issues: The extension of automation to the diagnostic assessment of clinical materials raises issues of professional responsibility, on the part of both the medical professional and designer of the device. The International Academy of Cytology (IAC) and other professional cytology societies should develop a policy towards automation in the diagnostic assessment of clinical cytologic materials. Consensus Position: The following summarizes the discussion of the initial position statement at the International Expert Conference on Diagnostic Cytology Towards the 21st Century, Hawaii, June 1997. 1. The professional in charge of a clinical cytopathology laboratory continues to bear the ultimate medical responsibility for diagnostic decisions made at the facility, whether automated devices are involved or not. 2. The introduction of automated procedures into clinical cytology should under no circumstances lead to a lowering of standards of performance. A prime objective of any guidelines should be to ensure that an automated procedure, in principle, does not expose any patient to new risks, nor should it increase already-existing, inherent risks. 3. Automated devices should provide capabilities for the medical professional to conduct periodic tests of the appropriate performance of the device. 4. Supervisory personnel should continue visual quality control screening of a certain percentage of slides dismissed at primary screening as within normal limits (WNL), even when automated procedures are employed in the laboratory. 5. Specifications for the design of primary screening devices for the detection of cervical cancer issued by the IAC in 1984 were reaffirmed. 6. The setting of numeric performance criteria is the proper charge of regulatory agencies, which also have the power of enforcement. 7. Human expert verification of results represents the 'gold standard' at this time. Performance characteristics of computerized cytology devices should be determined by adherence to defined and well-considered protocols. Manufacturers should not claim a new standard of care; this is the responsibility of the medical community and professional groups. 8. Cytology professionals should support the development of procedures that bring about an improvement in diagnostic decision making. Advances in technology should be adopted if they can help solve problems in clinical cytology. The introduction of automated procedures into diagnostic decision making should take place strictly under the supervision and with the active participation and critical evaluation by the professional cytology community. Ongoing Issues: Guidelines should be developed for the communication of technical information about the performance of automated screening devices by the IAC to governmental agencies and national societies. Also, guidelines are necessary for the official communication of IAC concerns to industry, medicolegal entities and the media. Procedures and guidelines for the evaluation of studies pertaining to the performance of automated devices, performance metrics and definitions for evaluation criteria should be established.
• Bartels, P. H., D., V., Montironi, R., Hamilton, P. W., Thompson, D., Vaught, L., & Bartels, H. G. (1998). Chromatin texture signatures in nuclei from prostate lesions. Analytical and Quantitative Cytology and Histology, 20(5), 407-416.
  More info

  PMID: 9801759;Abstract: OBJECTIVE: To characterize nuclei from prostatic lesions in a highly specific manner by developing a nuclear chromatin texture signature and to characterize lesions by means of their composition of nuclei with diverse degrees of deviation from normal. STUDY DESIGN: High-resolution digitized imagery of nuclei from normal prostates, from prostatic neoplastic lesions of low and high grade and from histoIogically normal appearing regions of prostates with low and high grade prostatic intraepithelial neoplasia (PIN) lesions were recorded. A set of 65 features descriptive of the spatial and statistical distribution of nuclear chromatin was computed for each nucleus. These features were arranged and processed to form a distinctive signature. A distance metric from 'normal' was defined and computed for each nucleus. RESULTS: Profiles of feature values can, after suitable scaling, be presented as distinctive feature value signatures. For many practical applications, profiles based on a standardized distance from normal nuclei may be more useful. Such profiles allow the derivation of a progression curve, showing increasing distances for diagnostic groups with increasing lesion progression up to high grade PIN lesions. Within each diagnostic group different cases show distinctive distributions of nuclei with differing degrees of deviation from normal, allowing the derivation of a lesion signature. CONCLUSION: Nuclear chromatin texture signatures may be of value for the characterization of both nuclei and lesions. They are based on a more comprehensive use of information offered by the nuclear chromatin pattern than that included in classification methods. While these signatures offer a more specific characterization of a clinical sample, they also are subject to more variability within a diagnostic category. This may not be due to randomness but may reflect some actual differences between lesions.
• Bartels, P. H., Montironi, R., Hamilton, P. W., Thompson, D., Vaught, L., & Bartels, H. G. (1998). Nuclear chromatin texture in prostatic lesions: I. Pin and adenocarcinoma. Analytical and Quantitative Cytology and Histology, 20(5), 389-396.
  More info

  PMID: 9801757;Abstract: OBJECTIVE: To document changes in the chromatin pattern in secretory cell nuclei from prostates with prostatic intraepithelial neoplasia (PIN) or adenocarcinoma. STUDY DESIGN: High-resolution images of nuclei were recorded, and a set of features descriptive of the chromatin texture and spatial distribution was computed. From this data set, features undergoing a monotonic trend of progression were selected and plotted to reveal trends in lesion progression. RESULTS: The nuclear chromatin in secretory cells in prostates with either PIN or malignant adenocarcinoma undergoes distinct and statistically significant changes in its texture and spatial distribution. Two trends of progressive change were observed. First, the values of a number of features descriptive of the clumpiness of the chromatin increase from values found in normal prostates to those recorded for nuclei from low grade to high grade PIN lesions. The second trend is a decrease in the values of the same features from those found in nuclei from high grade PIN still facing an intact basal cell layer to those no longer facing such a layer. This may be the first detectable step in progression towards development of a malignant lesion. There is a further decrease in nuclei in glands immediately adjacent to adenocarcinoma and in malignant lesions themselves. CONCLUSION: The described changes may lend themselves to the monitoring of lesion progression or of response to treatment or to chemopreventive intervention.
• Bartels, P. H., Montironi, R., Hamilton, P. W., Thompson, D., Vaught, L., & Bartels, H. G. (1998). Nuclear chromatin texture in prostatic lesions: II. Pin and malignancy associated changes. Analytical and Quantitative Cytology and Histology, 20(5), 397-406.
  More info

  PMID: 9801758;Abstract: OBJECTIVE: The objective of this study was to identify and document prostatic intraepithelial neoplasia (PIN) and malignancy associated changes in secretory cell nuclei from visually normal appearing tissue regions of prostates harboring PIN or adenocarcinoma. STUDY DESIGN: High-resolution digitized images of nuclei were recorded in histologically normal appearing tissue regions at defined distances from the margin of PIN or malignant lesions. Features descriptive of nuclear chromatin texture were computed and used to derive a discriminant function score for each nucleus. RESULTS: Secretory cell nuclei in prostates harboring either PIN or adenocarcinoma were shown to have statistically significantly different chromatin texture from secretory cell nuclei recorded in prostates free from any such lesion. The expression of PIN or malignancy associated changes was documented for distances up to 10 mm from the margin of a lesion. CONCLUSION: The finding of characteristic changes in nuclear chromatin texture of nuclei from histologically normal appearing tissue in prostates with PIN or adenocarcinoma offers the potential for higher sensitivity of detection of such lesions and for earlier detection of changes potentially preceding the development of clinically significant disease.
• Bartels, P. H., Montironi, R., Thompson, D., Vaught, L., & Hamilton, P. W. (1998). Statistical histometry of the basal cell/secretory cell bilayer in prostatic intraepithelial neoplasia. Analytical and Quantitative Cytology and Histology, 20(5), 381-388.
  More info

  PMID: 9801756;Abstract: OBJECTIVE: To delineate the sampling requirements for a histometric assessment of progression in low grade and high grade prostatic intraepithelial neoplasia (PIN) lesions. STUDY DESIGN: Images of whole glands from normal prostates, low grade PIN lesions and high grade PIN lesions were digitized. The images were processed by a machine vision system and automatically segmented, and a number of histometric characteristics descriptive of the disruption of the basal cell layer were extracted. Next, high-resolution images of secretory cell nuclei still facing or no longer facing intact segments of the basal cell layer were recorded and karyometrically analyzed. RESULTS: For the characterization of an individual lesion a minimum of 20-30 glands should be analyzed to provide an estimate of a progression index. Then, a change in progression, or due to regression, of approximately 16% can be documented. The disruption of the basal cell layer is accompanied by statistically highly significant changes in the chromatin texture and spatial distribution in secretory cell nuclei no longer facing an intact segment of that layer. CONCLUSION: Automated histometry by machine vision can provide valuable quantitative data for diagnostic assessment and for monitoring the efficacy of chemopreventive treatment.
• Bozzo, P. D., Vaught, L. C., Alberts, D. S., Thompson, D., & Bartels, P. H. (1998). Nuclear morphometry in solar keratosis. Analytical and Quantitative Cytology and Histology, 20(1), 21-28.
  More info

  PMID: 9513688;Abstract: OBJECTIVE: To carry out a feasibility study for the development of procedures for the objective characterization and grading of solar keratotic skin lesions. STUDY DESIGN: Imagery from sections of skin shave biopsies from 12 light-skinned individuals were digitized. A minimum of 25 nuclei from a solar keratotic lesion and 25 nuclei from a location in histologically normal appearing skin adjacent to the lesion were recorded for each case. Values of karyometric features were computed, and a discriminant function distinguishing normal nuclei from nuclei exhibiting solar irradiation damage was derived. RESULTS: Approximately 50% of nuclei in solar keratotic lesions were markedly affected by solar irradiation, but even in biopsies from histologically normal appearing skin, 3-30% of nuclei showed signs of such damage. Nuclei from solar keratotic lesions exhibiting such damage had numerous morphometric and karyometric features commonly found in malignant cells. The state of progression of a solar keratotic lesion can be graded by a plot of proportion of nuclei exhibiting solar damage versus the average discriminant function score of the most affected nuclei. This plot provides a monotonically rising progression curve and a numeric grading score. CONCLUSION: Karyometry of nuclei from skin biopsies allows objective assessment of the progression of solar keratotic lesions. Similarity of feature values in nuclei from solar keratotic lesions to those in malignant lesions was noted. The progression curve derived in this study could serve to measure the efficacy of chemopreventive or therapeutic intervention.
• Hamilton, P. W., Bartels, P. H., Montironi, R., Anderson, N. H., Thompson, D., Diamond, J., Trewin, S., & Bharucha, H. (1998). Automated histometry in quantitative prostate pathology. Analytical and Quantitative Cytology and Histology, 20(5), 443-460.
  More info

  PMID: 9801763;Abstract: OBJECTIVE: To review progress on the development of machine vision and image understanding in prostate tissue histology and to discuss the problems and opportunities afforded to pathology through the use of these techniques. STUDY DESIGN: A variety of concepts in machine vision are explored, and methodologies are described that have been developed to deal with the complexities of histologic imagery. The theory of human vision and its impact on machine vision are discussed. Software has been specifically developed for the analysis of prostate histology, allowing accurate gland segmentation, basal cell identification and measurement of vascularization within lesions. RESULTS: Image interpretation can be achieved using knowledge-based image analysis and the application of local object-oriented processing. This successfully allows an automated quantitative analysis of histologic morphology in the diagnosis of prostate intraepithelial neoplasia and invasive prostatic cancer. The use of low-power image scanning, based on textural or n-gram mapping, permits the development of fully automated devices for the rapid detection of tissue abnormalities. High-power, knowledge-guided scene segmentation can be carried out for the quantitative analysis of cellular features and the objective grading of the lesion. CONCLUSION: Automated tissue section scanning and image interpretation is now possible and holds much promise in prostate pathology and other diagnostically demanding areas. Issues of standardization still need to be addressed, but the development of such systems will undoubtedly enhance our diagnostic capabilities through the automation of time-consuming procedures and the quantitative evaluation of disease processes.
• Montironi, R., Bartels, P. H., & Bostwick, D. G. (1998). Quantitative methods in prostate pathology: Recent findings and future directions. Introduction to the symposium. Analytical and Quantitative Cytology and Histology, 20(5), 321-.
• Montironi, R., Mazzucchelli, R., Santinelli, A., Hamilton, P. W., Thompson, D., & Bartels, P. H. (1998). Case diagnosis as positive identification in prostatic neoplasia. Analytical and Quantitative Cytology and Histology, 20(5), 424-436.
  More info

  PMID: 9801761;Abstract: OBJECTIVE: To apply a distance measure and Bayesian belief network- based methodology to the positive identification of case diagnosis in prostatic neoplasia. STUDY DESIGN: Eight morphologic and cellular features were analyzed in 20 cases of normal prostate, 20 of low grade prostatic intraepithelial neoplasia (PIN), 20 of high grade PIN, 20 of prostatic adenocarcinoma with a cribriform pattern and 20 of prostatic adenocarcinoma with an acinar pattern. The diagnostic distance was evaluated to measure the 'extent' to which the feature outcomes of the individual cases differed from the expected profile of outcomes in typical cases of normal prostate, low and high grade PIN, and cribriform and large acinar adenocarcinoma. Belief values were evaluated with a Bayesian belief network (BBN). RESULTS: A bivariate representation of the cumulative absolute diagnostic distances of all the cases from the prototypes of normal prostate and cribriform adenocarcinoma was made. Three separate groups of cases were observed, corresponding to normal prostate, low grade PIN and cribriform adenocarcinoma. An additional group was formed by the cases of high grade PIN and acinar adenocarcinoma - i.e., there was complete overlap between the diagnostic distance values of cases belonging to these two categories. However, these cases showed differences in clue outcomes. To explore the contribution of such observations to case identification, a bivariate representation of the diagnostic distances from high grade PIN and acinar adenocarcinoma was made. The cases then formed five separate groups corresponding to the five diagnostic categories. When the individual cases were considered, their shortest distance was from the prototype of the category into which they were originally diagnosed. The BBN gave these diagnostic categories the highest belief values. CONCLUSION: The combined evaluation of diagnostic distance and belief represents an identification procedure. The numeric value of certainty characterizes individual cases according to the level of progression from PIN toward cancer.
• Anderson, N. H., Hamilton, P. W., Bartels, P. H., Thompson, D., Montironi, R., & Sloan, J. M. (1997). Computerized scene segmentation for the discrimination of architectural features in ductal proliferative lesions of the breast. Journal of Pathology, 181(4), 374-380.
  More info

  PMID: 9196433;Abstract: The distinction between ductal hyperplasia (DH) and ductal carcinoma in situ (DCIS) still remains a problem in the histological diagnosis of non- invasive breast lesions. In this study, a method was developed for the automatic segmentation and quantitative analysis of breast ducts using knowledge-guided machine vision. This permitted duct profiles and intraduct lumina to be identified and their shape, size, and number computed. These were used to derive measures of duct cribriformity and architectural complexity which were examined as an objective tool in the characterization of duct pattern in proliferative lesions. A total of 215 images of ducts were digitally captured from 22 cases of DCIS and 21 cases of DH diagnosed independently by two pathologists. The cribriformity index proved to be a useful measure of duct architecture, showing a monotonic increase with increasing duct complexity. The number of lumina also increased with increasing overgrowth of ductal epithelium until the duct was filled. Discriminant analysis of the duct characteristics for benign and malignant groups selected the lumen area/duct area ratio and the duct area as significant discriminatory variables and they were combined into a discriminant function. Of the lumen features, the mean area of the lumen and the polar average (mean of the distribution of the number of events with an increasing spiral from the centre of the duct) were combined into a second discriminant function. Plotting cases against these two functions provided good separation of DH and DCIS groups, with correct classification estimated on the training sample as being over 80 per cent. With an increasing incidence of complex proliferative lesions arising from mammography, the ability to diagnose these lesions correctly is more important than ever. The use of expert system-guided machine vision facilitates the quantitative evaluation of breast duct architecture; along with established histological and cytological criteria, it is hoped that this will lead to a more objective means of diagnosis and disease classification.
• Bartels, P. H., & Wied, G. L. (1997). Automated screening for cervical cancer: diagnostic decision procedures.. Acta cytologica, 41(1), 6-10.
  More info

  PMID: 9022719;
• Bartels, P. H., Gahm, T., & Thompson, D. (1997). Automated microscopy in diagnostic histopathology: From image processing to automated reasoning. International Journal of Imaging Systems and Technology, 8(2), 214-223.
  More info

  Abstract: A machine vision system for diagnostic histopathology offers five modules: 1) for the automated detection of regions of abnormality in histopathologic sections; 2) for fully automated image segmentation and diagnostic information extraction by a knowledge-guided procedure; 3) for the derivation of histometric indices, such as a progression index or grade for a lesion; 4) for diagnostic evidence evaluation by Bayesian inference networks; and 5) for individual patient targeted prognosis based on a case-based reasoning process. The system has been in operation for several years. Correct segmentation for even complex scenes such as cribriform glands has been achieved with a high success rate for histopathologic sections from prostate, colon, and breast. The lesion search module and prognostic module have passed feasibility testing and are still undergoing development. © 1997 John Wiley & Sons, Inc.
• Hamilton, P. W., Anderson, N., Diamond, J., Montironi, R., Trewin, S., Thompson, D., & Bartels, P. H. (1997). Computer-based decision support systems in diagnostic pathology. Electronic Journal of Pathology and Histology, 3(3), 23-33.
  More info

  Abstract: When faced with a difficult diagnostic decision, pathologists often seek support. This can come from information in the current medical literature including pathological journals and textbooks and through consultation with other pathologists in their own institute and experts from a different centre. Problems arise when the information available is poor or when consultation with specialists in the field is expensive. Here access to information or knowledge stored within a computer system would prove to be a very valuable tool. In their simplest form, computers can extend the conventional text book scenario through the digital storage of numerous microscopic images representing a given condition thus providing the pathologist with better decision support. An alternative form of decision support methodology is where attempts are made to 'capture' the knowledge used in diagnostic decision making and where the machine actively suggests diagnoses based on information entered into it. Given that experience is often lost when experts in a given field retire or die, the ability to represent knowledge in an expert system is an extremely attractive option. Capturing knowledge can be carried out through the use of formal logic methodology or by the conversion of linguistic terms and their relationships into numerical equivalents, thus allowing us to make mathematical classifications based on descriptive terminology rather than from formal measurement. Bayesian belief networks in simple and more complex architectures provide a robust method for representing and handling uncertainty in morphological diagnosis. The use of distance metrics based on descriptive assessment enhances the capabilities of conventional diagnostic practice, providing objective data on the certainty of the diagnostic decision and a quantitative insight into the decision making process. Case-based reasoning uses the advantages of database management and case retrieval to assist diagnosis. These varied approaches provide an extremely rich environment for the investigation of pathological decision making, its shortfalls and its complexities. Such tools could have a major impact on diagnostic accuracy, consistency and reproducibility and are likely to be widely applied in the next few years.
• Hamilton, P. W., Bartels, P. H., Thompson, D., Anderson, N. H., Montironi, R., & Sloan, J. M. (1997). Automated location of dysplastic fields in colorectal histology using image texture analysis. Journal of Pathology, 182(1), 68-75.
  More info

  PMID: 9227344;Abstract: Automation in histopathology is an attractive concept and recent advances in the application of computerized expert systems and machine vision have made automated image analysis of histological images possible. Systems capable of complete automation not only require the ability to segment tissue features and grade histological abnormalities, but must also be capable of locating diagnostically useful areas from within complex histological scenes. This is the first stage of the diagnostic process. The object of this study was to develop criteria for the automatic identification of focal areas of colorectal dysplasia from a background of histologically normal tissue. Fields of view representing normal colorectal mucosa (n = 120) and dysplastic mucosa (n = 120) were digitally captured and subjected to image texture analysis. Two features were selected as being the most important in the discrimination of normal and adenomatous colorectal mucosa. The first was a feature of the co-occurrence matrix and the second was the number of low optical density pixels in the image. A linear classification rule defined using these two features was capable of correctly classifying 86 per cent of a series of training images into their correct groups. In addition, large histological scenes were digitally captured, split into their component images, analysed according to texture, and classified as normal or abnormal using the previously defined classification rule. Maps of the histological scenes were constructed and in most cases, dysplastic colorectal mucosa was correctly identified on the basis of image texture: 83 per cent of test images were correctly classified. This study demonstrates that abnormalities in low-power tissue morphology can be identified using quantitative image analysis. The identification of diagnostically useful fields advances the potential of automated systems in histopathology: these regions could than be scrutinized at high power using knowledge-guided image segmentation for disease grading. Systems of this kind have the potential to provide objectivity, unbiased sampling, and valuable diagnostic decision support.
• Mariuzzi, G., Mombello, A., Mariuzzi, L., Hamilton, P. W., Weber, J. E., Thompson, D., & Bartels, P. H. (1997). Quantitative study of ductal breast cancer - Patient targeted prognosis: An exploration of case base reasoning. Pathology Research and Practice, 193(8), 535-542.
  More info

  PMID: 9406246;Abstract: Current analytic methodologies allow the extraction, even from small tumor masses, of extensive information on the biologic characteristics of malignant lesions, such as tumor aggressivity, metastatic potential, drug resistance, and host interactions. Clinical practice now offers a wide range of therapeutic strategies. Information technological advances offer the opportunity to refer to very large data bases of patient anamnestic data, response to treatment and clinical outcome. There is a need to formulate therapy and prognosis for each individual case. Case based reasoning is a knowledge based methodology where the outcome for complex situations can be predicted by referring to a large data base of cases of known outcomes. The preliminary data obtained from this study suggest that case based reasoning may offer a promising approach to individual targeted prognosis.
• Montironi, R., Diamanti, L., Pomante, R., Thompson, D., & Bartels, P. H. (1997). Subtle changes in benign tissue adjacent to prostate neoplasia detected with a Bayesian belief network. Journal of Pathology, 182(4), 442-449.
  More info

  PMID: 9306966;Abstract: The aim of this paper was to test the usefulness of a Bayesian belief network (BBN) as a decision support system in the uncertainty assessment of benign prostatic tissue, either associated or not with inflammation or adjacent to prostatic adenocarcinoma (PAC) or prostatic intraepithelial neoplasia (PIN). A shallow network was used with eight first-level descendant nodes for the diagnostic clues, each independently linked by a conditional probability matrix to a root node containing the diagnostic alternatives. One diagnostic evidence node was based on the tissue architecture and the others were based on cell features. The efficacy of the network was tested on a series of 45 simple prostatectomy specimens, subdivided as follows: benign prostatic tissue not associated with other diseases (15 cases), associated with acute and/or chronic inflammation (15 cases), and adjacent to accidentally discovered PAC or PIN (15 cases). The highest belief values for the diagnostic alternative normal prostate (NP) were obtained in the 15 cases not associated with other diseases, the mean value being 0.996. The 15 cases evaluated in areas with inflammation showed the lowest belief values for NP (mean 0.774). For the 15 cases evaluated in specimens with PAC or PIN, the belief values for NP were intermediate between those from normal prostatic tissue associated with inflammation and those not associated (mean 0.925). Moreover, it was found that subtle changes were also present at a certain distance from the tumour. In conclusion, the network can be used as a decision support system to differentiate with high certainty benign prostate adjacent to PAC or PIN from benign prostatic tissue either associated or not with inflammation. The subtle morphological alterations detected with the BBN may be considered malignancy-associated changes.
• Montironi, R., Pomante, R., Colanzi, P., Thompson, D., Hamilton, P. W., & Bartels, P. H. (1997). Diagnostic distance of high grade prostatic intraepithelial neoplasia from normal prostate and adenocarcinoma. Journal of Clinical Pathology, 50(9), 775-782.
  More info

  PMID: 9389981;PMCID: PMC500177;Abstract: Objective - To develop a distance measure based methodology to support the morphological evaluation of high grade prostatic intraepithelial neoplasia (PIN), a direct precursor of prostate cancer. Methods - Eight morphological and cellular features were analysed in 20 cases of high grade PIN found in radical prostatectomy specimens from patients with adenocarcinoma. The diagnostic distance was evaluated to measure the extent to which the feature outcomes of the individual high grade PIN cases differed from the expected outcome profile of normal prostate, low and high grade PIN, and cribriform and large acinar adenocarcinoma. The belief value for high grade PIN was evaluated with a Bayesian belief network (BBN). Results - Complete separation existed between the cumulative absolute diagnostic distances of these 20 cases from the prototype feature outcomes of high grade PIN and normal prostate the values for which were ≤ 3 (range 0 to 3) and ≤ 9 (range 9 to 15), respectively. The distances from low grade PIN (range 3 to 9), cribriform adenocarcinoma (range 2 to 8), and large acinar adenocarcinoma (range 5 to 10) were intermediate and showed overlap in their distribution. When taking into consideration whether the severity of feature changes was increasing or decreasing in comparison with the category prototype outcomes, the cumulative directional diagnostic distances from high grade PIN ranged from -3 to +3. Positive distance values were seen relative to low grade PIN (range +3 to +9) and relative to normal prostate (range +9 to +15). Negative values were found relative to cribriform adenocarcinoma (range -8 to +2). The distance values from large acinar adenocarcinoma ranged from -2 to +4 and partly overlapped with those from the high grade PIN category. A bivariate scattergram derived from both diagnostic distance measures showed excellent separation between the groups' distances. BBN analysis confirmed the morphology based diagnosis. The distance evaluation resulted in 18 cases whose belief value for high grade PIN ranged from 0.60 to 0.87. In the remaining two cases the results of the BBN analysis showed a belief value of 0.50 and 0.57 for low grade PIN and of 0.49 and 0.38 for high grade PIN, respectively. Conclusions - Distance measure based methodology represents a useful diagnostic decision support tool for the accurate evaluation of high grade PIN.
• Scarpelli, M., Montironi, R., Thompson, D., & Bartels, P. H. (1997). Computer-assisted analysis of medulloblastoma: A cytologic study. Analytical and Quantitative Cytology and Histology, 19(5), 387-392.
  More info

  PMID: 9349898;Abstract: OBJECTIVE: To explore data from a set of cases of medulloblastoma to see whether quantitative image analysis might suggest evidence for the existence of lower and higher grade lesions. STUDY DESIGN: Fourteen consecutive cases of medulloblastoma were obtained. Smears were stained with toluidine blue. For each case, 50 nuclei were measured and a number of densitometric features extracted. RESULTS: The existence of two subgroups of cases, identified as lower and higher grade groups, was suggested by a plot of the total optical density versus nuclear area. Two nuclear texture features-the number of pixels with the same optical density value occurring consecutively in the nucleus and the proportion of pixels in the high optical density range- divided the cases into the same subgroups. The use of a clustering algorithm established two clusters that corresponded to that subgrouping except for one case. Discriminant analysis gave an identical classification, with the misplaced case having a borderline discriminant function score. An unsupervised learning algorithm based on an adaptive distance metric formed two clusters and assigned the borderline case to the low grade subgroup. The grouping obtained by quantitative analysis was only partly related to the grade of nuclear atypia subjectively evaluated. CONCLUSION: In our series of medulloblastomas, quantitative analysis provided a means of detecting differences in the nuclear size and texture that allowed the classification of cases into two subgroups.
• Scarpelli, M., Montironi, R., Thompson, D., & Bartels, P. H. (1997). Computer-assisted discrimination of glioblastomas. Analytical and Quantitative Cytology and Histology, 19(5), 369-375.
  More info

  PMID: 9349896;Abstract: OBJECTIVE: To measure a number of nuclear features in a series of glioblastomas and compare the data with those from a set of anaplastic and low grade astrocytomas. STUDY DESIGN: The material consisted of toluidine blue-stained smears from 13 consecutive cases of glioblastoma. Smears from 12 high grade astrocytomas and 13 low grade fibrillary astrocytomas were used for comparison. Fifty nuclei were measured in each case. Cell images were segmented by an interactive procedure. A set of features representing both morphometric and nuclear texture characteristics was computed. RESULTS: The use of a discriminant function based on two features related to the gray value distribution resulted in the separation of all low grade astrocytomas from glioblastomas. When the corresponding discriminant function was computed for high grade astrocytomas and the values were plotted against optical density, the glioblastomas formed a group at a greater distance from the low grade astrocytomas than from the high grade astrocytomas. A second discriminant function based on two features allowed complete separation of the glioblastoma cases from the high grade astrocytomas. CONCLUSION: In our material, chromatin texture analysis allowed effective separation of astrocytic tumors with different histologic grades of malignancy.
• Bartels, P. H., Thompson, D., & Montironi, R. (1996). Knowledge-based image analysis in the precursors of prostatic adenocarcinoma. European Urology, 30(2), 234-242.
  More info

  PMID: 8875205;Abstract: Objectives: It is the objective of this study to present the use of knowledge-guided procedures in quantitative image analysis and interpretation in histopathology. Methods: The knowledge-guided procedures were implemented in the form of N-gram encoding methods for the search and detection of areas of atypicality or abnormality in histopathologic sections; they were implemented as expert system for automated scene segmentation based on an associative network with frames at each node. The extraction of histometric features from the basal cell layer of prostatic lesions is presented as an example of automated image interpretation. Results: Rapid search algorithms for lesion detection were able to identify approximately 90% of areas labelled as atypical or abnormal by visual assessment, in lesions of colon, prostate and breast. Automated segmentation of very complex histopathologic imagery was possible with a success rate of approximately 80-90%, in sections of prostatic and colonic lesions. Histometry of the deterioration of the basal cell layer in prostatic lesions provided a monotonic trend curve suitable for the measurement of progression or regression. Conclusions: Knowledge-guided procedures bring external information, not offered by the imagery itself, to bear on image processing and image analytic methods. This has enabled automated analysis and interpretation of very complex imagery, such as from cribriform glands, resulting in quantitative diagnostic information.
• Bartels, P. H., Thompson, D., Montironi, R., Mariuzzi, G., & Hamilton, P. W. (1996). Automated reasoning system in histopathologic diagnosis and prognosis of prostate cancer and its precursors. European Urology, 30(2), 222-233.
  More info

  PMID: 8875204;Abstract: Objective: This article presents the rationale and options offered to diagnostic and prognostic decision support systems for prostate pathology by automated reasoning capabilities. Methods: The symbolic information used in diagnostic decision-making is systematically ordered, compared, numerically assessed in its probability, and combined such that a conclusion can be drawn. The framework for the processing of such symbolic information may be an expert system, an inference network or a case-based reasoning system. Automated reasoning is implemented by the use of a rule base and information flow control modules. Results: Automated reasoning allows decision support systems to follow highly adaptive decision sequences, capable of handling contradictory evidence, exceptions in diagnostic clue expression, and nonmonotonic decision-making. Conclusions: Automated reasoning capability in diagnostic and prognostic decision support systems allows highly flexible decision development, very close to human decision procedures.
• Hamilton, P. W., Anderson, N. H., Diamond, J., Bartels, P. H., Gregg, J. B., Thompson, D., & Millar, R. J. (1996). An interactive decision support system for breast fine needle aspiration cytology. Analytical and Quantitative Cytology and Histology, 18(3), 185-190.
  More info

  PMID: 8790830;Abstract: OBJECTIVE: To develop a computerized system to assist in the diagnosis of malignancy in breast fine needle aspiration cytology. STUDY DESIGN: A Bayesian belief network was designed to control uncertainty and allow a diagnostic decision to be reached based on the sequential collection of cytologic information. Ten cytologic features were defined as clues that contribute to the diagnostic discrimination of benign and malignant aspirates. The impact of each feature on the diagnostic decision was quantified by a conditional probability matrix. RESULTS: For the assessment of a new case, the computer guides the user through the diagnosis, prompting him or her for information on each of the diagnostic features in turn. For each feature, the user is presented with a series of digitally stored color microscopic images that have been selected to represent good examples of the different feature grades-e.g., pleomorphism: none, mild, moderate and severe. Each image is mapped to an overlapping curve, and by positioning a line on the spectrum where the user feels the case lies, a membership function vector is calculated and entered as evidence into the network. This results in an update in the belief in the diagnostic alternatives. After all the clues have been assessed, a final diagnostic probability is reported. In addition, a cumulative belief curve can be drawn that maps the change in the diagnostic probabilities after each piece of evidence has been submitted, providing unique insight into the diagnostic process. CONCLUSION: Systems like this represent an important step forward in the use of descriptive classifiers. They impose consistency in terminology, improve reproducibility in the grading of cellular abnormalities and remove subjectivity in interpreting the significance of pvisual clues to diagnosis. As such, they represent a necessary tool in pathologic decision making.
• Mariuzzi, G. M., Mariuzzi, L., Mombello, A., Santinelli, A., Valli, M., Rahal, D., Thompson, D., & Bartels, P. H. (1996). Quantitative study of ductal breast cancer progression. A progression index (P.I.) for premalignant lesions and in situ carcinoma. Pathology Research and Practice, 192(5), 428-436.
  More info

  PMID: 8832747;Abstract: The diagnostic subjective assessment of ductal premalignant proliferative lesions and in situ carcinoma of the breast produces unsatisfactory results. Since the phenotypical cell changes in tumour progression toward infiltrating cancer constitute a continuum, a grading on a continuous scale of values produces a more reliable and reproducible characterization. The diagnostic assessment for any individual patient may be expressed by a progression index (P.I.): its numerical values are based on the cellular changes measured in the individual cases. In this study, the progression index is based an two morphometric features, nuclear size and nucleolar area. In addition, the method presented may produce a ratio, stating the relative likelihood that each case represents one of the conventional diagnostic categories. Such a likelihood ratio may be obtained from the bivariate distribution of nuclear size and nucleolar area for the conventional diagnostic categories.
• Mariuzzi, G., Mombello, A., Mariuzzi, L., Thompson, D., & Bartels, P. H. (1996). Objective quantitative grading: A study of breast ductal hyperplasias and ductal carcinomas in situ. Annals of the New York Academy of Sciences, 784, 395-402.
  More info

  PMID: 8651587;
• Montironi, R., Bartels, P. H., Hamilton, P. W., & Thompson, D. (1996). Atypical adenomatous hyperplasia (adenosis) of the prostate: Development of a Bayesian belief network for its distinction from well-differentiated adenocarcinoma. Human Pathology, 27(4), 396-407.
  More info

  PMID: 8617484;Abstract: The diagnosis of atypical adenomatous hyperplasia (AAH) of the prostate and its distinction from well-differentiated prostatic adenocarcinoma with small acinar pattern (PACsmac; Gleason primary grades 1 or 2) are affected by uncertainties that arise from the fact that the knowledge of AAH histopathology is expressed in descriptive linguistic terms, words, and concepts. A Bayesian belief network (BBN) was used to reduce the problem of uncertainty in diagnostic clue assessment, while still considering the dependencies between elements in the reasoning sequence. A shallow network was designed and developed with an open-tree topology, consisting of a root node containing two diagnostic alternatives (eg, AAH v PACsmac) and 12 first- level descendant nodes for the diagnostic features. Eight of these nodes were based on cell features, three on the type of gland lumen contents and one on the gland shape. The results obtained with prototypes of relative likelihood ratios showed that belief for the diagnostic alternatives is high and that the network can differentiate AAH from PACsmac with certainty. The features that best contributed to the highest belief were those concerning the nucleolar size, frequency, and location. In particular, after the analysis of five nucleolar features (prominent nucleoli, inconspicuous nucleoli, nucleoli with diameter greater than 2.5 μm, nucleolar margination, and nuclei with multiple nucleoli), the belief for AAH was 1.0, being already close to 1.0 when three were evaluated (the value range is 0.0 to 1.0; the closer to 1.0, the greater the belief). The contribution of the three features concerning the gland lumen contents (mucinous material, corpora amylacea, and crystalloids) was such that the final belief did not exceed 0.8. Results with the group of remaining features (eg, basal cell recognition, gland shape variation, cytoplasm appearance, and nuclear size variation) were slightly better. These features allowed a substantial accumulation of belief that was already greater than 0.9 when three were polled. However, the maximum belief value was never obtained. In conclusion, a BBN for AAH diagnosis offers a descriptive classifier that is readily implemented, and allows the use of linguistic, fuzzy variables, and the accumulation of evidence presented by diagnostic clues.
• Montironi, R., Bartels, P. H., Thompson, D., Diamanti, L., & Prete, E. (1996). Androgen-deprived prostate adenocarcinoma: Evaluation of treatment-related changes versus no distinctive treatment effect with a Bayesian belief network. A methodological approach. European Urology, 30(3), 307-315.
  More info

  PMID: 8931962;Abstract: Objective: To develop and test a Bayesian belief network (BBN) for the identification of prostatic adenocarcinomas (PACs) with combination endocrine therapy (CET) changes from PACs with poor to no treatment CET effect and from untreated PACs. Methods: A network was designed with a decision node containing three diagnostic alternatives (PAC with CET effect, PAC with poor to no treatment effect, and untreated PAC) and seven first-level evidence nodes for the diagnostic features: nuclear enlargement; frequency of prominent nucleoli; cell cytoplasm vacuolization; shrunken acini; individual infiltrating tumor cells; WHO prostate cancer pattern recognition, and amount of interstitial tissue stroma. Three prototype cases, one for each diagnostic alternative, were used to develop the BBN. The BBN performance was then evaluated in 40 prostatectomies for PAC, consisting of 20 CET treated and 20 untreated cases. Results: The results obtained with the three prototypes showed that the network can identify the diagnostic alternatives with certainty when seven features are polled. When the performance was evaluated in the 40 PACs, the belief values were 1.0 or close to it in most of the cases (the value range is 0.0-1.0; the closer to 1.0, the greater the belief). Moreover, the BBN allowed an identification with high certainty of PACs with treatment-related changes from those either with poor to no treatment effect or untreated. Conclusions: A BBN for the evaluation of androgen-deprived PAC offers a descriptive classifier which is readily implemented and allows the use of descriptive, linguistic terms.
• Montironi, R., Diamanti, L., Thompson, D., Bartels, H. G., & Bartels, P. H. (1996). Analysis of the capillary architecture in the precursors of prostate cancer: Recent findings and new concepts. European Urology, 30(2), 191-200.
  More info

  PMID: 8875200;Abstract: Objective: To report on recent findings and new concepts in the remodeling of the capillary architecture in the precursors of prostate cancer. Methods: Immunohistochemical methods have been adopted in prostate cancer and in its precursors (prostatic intra-epithelial neoplasia) to investigate capillary pattern changes - which were mainly analyzed as capillary frequency - and the degree of endothelial cell proliferation. Several features related to the capillary architecture have been considered. Manual, semiautomatic, and automatic (machine vision) types of evaluation have been used to quantify the features. Results: The data available indicate that: (1) Going from normal prostate through prostatic intra-epithelial neoplasia up to invasive adenocarcinoma, an increasing proportion of capillaries becomes shorter, with open lumen and undulated external contour and with greater proliferation of the endothelial cells and greater expression of type IV collagenase. The highest proportion of touching capillaries is seen in normal prostate, while the lowest is found in invasive adenocarcinoma, being intermediate in prostatic intra-epithelial neoplasia. (2) When total androgen ablation is induced, there is no proliferation of the endothelium, whereas the capillaries are reduced in frequency and represented by small vessels lined by flat endothelial cells and with an open lumen. (3) Automation in the evaluation of the capillary architecture is feasible with a machine vision system. Conclusions: The progression in prostate carcinogenesis is associated with changes in the capillary architecture. There are some preliminary data indicating that total androgen ablation can inhibit the angiogenesis in precursors of prostate cancer.
• Montironi, R., Pomante, R., Diamanti, L., Hamilton, P. W., Thompson, D., & Bartels, P. H. (1996). Evaluation of prostatic intraepithelial neoplasia after treatment with a 5-α-reductase inhibitor (finasteride): A methodologic approach. Analytical and Quantitative Cytology and Histology, 18(6), 461-470.
  More info

  PMID: 8978870;Abstract: OBJECTIVE: To develop a methodology applicable to the morphologic study of the efficacy of finasteride on prostatic intraepithelial neoplasia (PIN), a putative precursor of prostate cancer. STUDY DESIGN: Three PIN foci were reviewed in two simple prostatectomy specimens from patients with clinical diagnoses of benign prostatic hyperplasia and treated with finasteride for six months. The feasibility of PIN diagnosis and grading based on 'diagnostic distance' was investigated. It is a measure of the 'extent' to which the observed features are different from those of the untreated prototypes representing the following diagnostic categories: normal prostate, low and high grade PIN and prostatic adenocarcinoma with a cribriform or large acinar pattern. Uncertainty in the PIN diagnosis and grading was dealt with by means of a Bayesian belief network (BBN). RESULTS: The distance measure values of the three PIN foci from the prototype of untreated, nonneoplastic prostate were 9, 7 and 8, respectively, in relative, arbitrary units. Their distance from the two prostate cancer patterns (large acinar and cribriform) was as high as 8-10. The distances of these foci from either low or high grade PIN were as low as 5, 3 and 2, and 3, 5 and 4, respectively. BBN produced the highest belief values for PIN, thus confirming the morphology-based and diagnostic distance-supported diagnosis; however, the belief values were low/or both grades. CONCLUSION: The results provided by BBN analyses and diagnostic distance measures support the conclusion that this methodology is applicable to assessing the efficacy of finasteride treatment of PIN.
• Montironi, R., Whimster, W. F., Collan, Y., Hamilton, P. W., Thompson, D., & Bartels, P. H. (1996). How to develop and use a Bayesian belief network. Journal of Clinical Pathology, 49(3), 194-201.
  More info

  PMID: 8675727;PMCID: PMC500395;
• Thompson, D., Bartels, P. H., Bartels, H. G., & Montironi, R. (1996). Knowledge-guided histometry of the basal cell layer in prostatic intraepithelial neoplasia. Analytical and Quantitative Cytology and Histology, 18(2), 177-184.
  More info

  PMID: 8744508;Abstract: OBJECTIVE: To define a procedure for the fully automated evaluation of histopathologic sections of prostatic intraepithelial neoplasia. STUDY DESIGN: The design is based on expert system-guided scene segmentation, the construction of a knowledge file and the defining of histometric measures for assessment of the basal cell layer. RESULTS: A knowledge file specifying a total of 263 entities was constructed. Of 73 multimegapixel-sized images of cribriform glands, 71 were correctly and completely processed. A histometric measure allowing precise indexing of the degree of lesion progression, based on the deterioration of the basal cell layer, was derived. CONCLUSION: The combination of image analytic and immunohistochemical procedures and fully automated processing provides a quantitative measure for evaluating prostatic intraepithelial neoplasia lesion progression and the effects of intervention.
• Whimster, W. F., Hamilton, P. W., Anderson, N. A., Humphreys, S., Boyle, M., Sundaresan, M., Rainey, A., Giles, A., Hopster, D., & Bartels, P. H. (1996). Reproducibility of Bayesian belief network assessment of breast fine needle aspirates. Analytical and Quantitative Cytology and Histology, 18(4), 267-274.
  More info

  PMID: 8862667;Abstract: OBJECTIVE: To assess the consistency of diagnosis of fine needle aspiration biopsies of breast lesions by three experienced and five less experienced pathologists using conventional means and applying a Bayesian belief network (BBN) to 10 diagnostic features to support diagnostic decision making. STUDY DESIGN: Forty fine needle aspiration biopsies, previously assessed by one of the experienced pathologists both conventionally and using a BBN, were assessed by two further experienced pathologists and five less experienced pathologists. RESULTS: Using the BBN, the experienced pathologists arrived at diagnoses in agreement with an established consensus at a slightly lower rate than by conventional means. The less experienced pathologists arrived at the correct diagnoses no more frequently with the help of the BBN than conventionally. CONCLUSION: As used in this study, the BBN did not help less experienced pathologists to interpret their observations but did enable less experienced pathologists to identify how their observations differed and affected their diagnoses. The prototype system used in this study has since been upgraded by providing computer graphic displays of the features to be observed so that a more uniform mental image can be held by the participating pathologists. This will be tested with the same study design.
• Wied, G. L., Bartels, P. H., Bibbo, M., Gupta, P. K., Gurley, A. M., Hilgarth, M., Jiménez-Ayala, M., Kato, H., Knight, B. K., McGoogan, E., Medley, G., Meisels, A., Nishiya, I., Nozawa, S., Ramzy, I., Reith, A., Rilke, F., Rivera-Pomar, J., Rosenthal, D. L., , Schenck, U., et al. (1996). Computer-assisted quality assurance.. Acta cytologica, 40(1), 1-3.
  More info

  PMID: 8604561;
• Bartels, P. H., Thompson, D., & Weber, J. E. (1995). Diagnostic and prognostic decision support systems.. Pathologica, 87(3), 221-236.
  More info

  PMID: 8570283;Abstract: Diagnostic decision support systems provide a quantitative evaluation of diagnostic evidence and the capability to combine diagnostic evidence in such a manner that a numeric measure of certainty in a final diagnostic recommendation results. Generally, expert systems serve to establish a diagnostic decision, inference networks allow a detailed analysis of the diagnostic value of diagnostic clues, case-based reasoning systems are designed to provide a prognostic assessment targeted to an individual patient. In all of these systems, symbolic information, i.e., traditional diagnostic, linguistic terms and concepts are processed and quantitatively evaluated.
• Bartels, P. H., Thompson, D., Bartels, H. G., Montironi, R., Scarpelli, M., & Hamilton, P. W. (1995). Machine vision-based histometry of premalignant and malignant prostatic lesions. Pathology Research and Practice, 191(9), 935-944.
  More info

  PMID: 8606876;Abstract: The implementation of knowledge-guided control of the processing and segmentation of histopathologic images of prostatic lesions has made automated analysis and interpretation possible. To establish correspondence between histopathologic concepts, terms and diagnostic criteria, and computed histometric entities, 'interpretive transforms' are introduced. Scene segmentation is controlled by art expert system following a model-based reasoning process. The expert system is structured as an associative network with frames at each node, which controls a knowledge file and a large library of image processing algorithms.
• Bartels, P. H., Thompson, D., Montironi, R., Hamilton, P. W., & Scarpelli, M. (1995). Diagnostic decision support for prostate lesions. Pathology Research and Practice, 191(9), 945-957.
  More info

  PMID: 8606877;Abstract: The diagnostic evaluation of premalignant and malignant lesions of the prostate may benefit from the application of an inference network. Used as a diagnostic decision support system, an inference network provides standardized assessment of diagnostic clues which is supported by computer graphics and comparison imagery, uncertainty management by possibility and probabilistic schemes and the systematic combination of different pieces of diagnostic evidence. This assessment results in a numeric measure of belief in the final diagnosis.
• Hamilton, P. W., Bartels, P. H., Montironi, R., Anderson, N., & Thompson, D. (1995). Improved diagnostic decision-making in pathology: Do inference networks hold the key?. Journal of Pathology, 175(1), 1-5.
  More info

  PMID: 7891221;
• Hamilton, P. W., Bartels, P. H., Wilson, R. H., & Sloan, J. M. (1995). Nuclear texture measurements in normal colorectal glands. Analytical and Quantitative Cytology and Histology, 17(6), 397-405.
  More info

  PMID: 8750354;Abstract: OBJECTIVE: To examine differences in epithelial cell nuclear texture along the length of normal, well-oriented colorectal glands. STUDY DESIGN: Histologically normal mucosa was obtained from four patients who had a family history of cancer and no hyperproliferation as assessed by bromodeoxyuridine immunohistochemistry. Serial sections were stained using the Feulgen technique and viewed to identify glands that were sectioned along their entire length. Two glands were sampled from each patient. Each gland was divided into three compartments based on epithelial cell counts along the length of the gland. Within each compartment, 10 nuclei were randomly selected and their nuclear texture features measured using computerized image analysis. RESULTS: Analysis of variance demonstrated that significant differences exist in the nuclear texture of cells in the different compartments of the gland. Discriminant analysis selected three variables that were important and independent in the quantitative discrimination of the gland compartments. These features were related to the elliptical shape, optical density distribution and cooccurrence structure of the chromatin texture within the epithelial nuclei. CONCLUSION: It is clear that these differences are associated with the spatial variation in proliferation and differentiation status found within colorectal mucosal glands. Nuclear measurements from colorectal epithelium with near-normal architecture should therefore be stratified in relation to position within the gland. Nuclear texture and size measurements within defined compartments may provide a sensitive means of detecting early malignant transformation.
• Hamilton, P. W., Montironi, R., Abmayr, W., Bibbo, M., Anderson, N., Thompson, D., & Bartels, P. H. (1995). Clinical applications of Bayesian belief networks in pathology.. Pathologica, 87(3), 237-245.
  More info

  PMID: 8570284;Abstract: Bayesian belief networks (BBNs) are a novel tool for representing knowledge about diagnostic decision making and for obtaining a numerical measure of certainty in the final diagnosis. Belief networks have been applied to the pathological assessment of breast, prostate and skin lesions and have been shown to provide consistency in the grading of microscopic features and improve diagnosis. These applications are reviewed in the current paper. The application of BBNs has been further facilitated through the use of standardised imagery which is stored digitally and used to enter evidence into a BBN. It is predicted that the further development of BBNs with improved logical capabilities represent the key to improved decision making in pathology.
• Hamilton, P. W., Thompson, D., Sloan, J. M., & Bartels, P. H. (1995). Knowledge-guided segmentation and morphometric analysis of colorectal dysplasia. Analytical and Quantitative Cytology and Histology, 17(3), 172-182.
  More info

  PMID: 7546051;Abstract: The histopathologic grading of colorectal adenomatous dysplasia is a subjective process. In this study, automated image analysis using knowledge- guided software was used to quantitatively assess colorectal glandular characteristics. Cases of histologically normal mucosa and of low, moderate and high grade dysplasia were examined using the technique. A total of 19 morphometric and densitometric features were measured on each gland. These included gland shape, epithelial area, nuclear stratification and nuclear optical density. Discriminant analysis of the data revealed those morphometric features which provided the best discrimination between the various histologic groups. The area of epithelium occupied by nuclei was the strongest discriminating variable, and this, in combination with a discriminant function derived from the remaining variables, was used to plot cases in bivariate sample space. The plots revealed that the data for normal glands were consistently well separated from dysplastic gland data. Data sets belonging to the various grades of dysplasia showed varying degrees of separation, depending which two histologic groups the discriminant function was based on. This study showed that automated image analysis of complex histologic scenes is possible using knowledge-guided segmentation and that it can provide useful data for the objective classification of colorectal dysplasia.
• Kronqvist, P., Montironi, R., Collan, Y. U., Bartels, P. H., & Thompson, D. (1995). Management of uncertainty in breast cancer grading with Bayesian belief networks. Analytical and Quantitative Cytology and Histology, 17(5), 300-308.
  More info

  PMID: 8534332;Abstract: OBJECTIVE: To examine the potential of different constructs of Bayesian belief networks (BBN) to manage uncertainty in breast cancer grading. STUDY DESIGN: We developed four networks, two based on Bloom-Richardson's and two on Helpap's grading systems. The function of the networks was based either on an expert's experience or frequency counts derived from subjective grading of a large number of samples. The four BBNs were tested on 20 specimens, and the resulting final beliefs were compared with tile subjective gradings. RESULTS: The BBNs showed agreement with the subjective gradings in 60-85% of cases. Different constructs of BBNs, however, differed in their performance. The mean beliefs in frequency-based networks were slightly higher than in experience based networks. In addition, as compared with the Bloom-Richardson based networks, the Helpap-based BBNs resulted in higher maximum beliefs but produced a larger fraction of discrepancies with the subjectively graded cases. Depending on the type of network, 65-90% of the BBN grades were associated with high beliefs. CONCLUSION: The results suggest that for reliable results, grading systems with more than three or four variables may be necessary. When based on relevant information, BBNs seem to have the potential to become a valuable method of assisting the pathologist in breast cancer grading.
• Montironi, R., Bartels, P. H., Thompson, D., Bartels, H. G., & Scarpelli, M. (1995). Prostatic intraepithelial neoplasia. Quantitation of the basal cell layer with machine vision system. Pathology Research and Practice, 191(9), 917-923.
  More info

  PMID: 8606874;Abstract: The aim of this paper is to report on a fully automated procedure to quantify the epithelial cell components in prostatic intraepithelial neoplasia with particular emphasis on the basal cell layer (BCL). Scene segmentation was guided by a knowledge-based system of digitized images recorded from histological section immunostained against high molecular weight keratin and counterstained with hematoxylin and eosin. Scene segmentation involved two major stages. First, the system located and identified the duct and segmented the scene, resulting in 'intermediate segmention products.' This stage was followed by a reconstruction process in which the segmentation products (i.e., the lumen and the darkly and lightly stained epithelial cell components) were assembled to form the microscopic structure to achieve working unity. Following this, histometric measurements were made of the reconstructed scene. Computed were the percentage of the duct contour with BCL (90%), and the number and length of gaps in the BCL (19, ranging from 10 to 90 microns). Automated analysis of the BCL is accurate and provides information not readily identified by human examination.
• Montironi, R., Bartels, P. H., Thompson, D., Scarpelli, M., & Hamilton, P. W. (1995). Prostatic intraepithelial neoplasia (PIN). Performance of Bayesian belief network for diagnosis and grading. Journal of Pathology, 177(2), 153-162.
  More info

  PMID: 7490682;Abstract: Prostatic intraepithelial neoplasia (PIN) diagnosis and grading are affected by uncertainties which arise from the fact that almost all knowledge of PIN histopathology is expressed in concepts, descriptive linguistic terms, and words. A Bayesian belief network (BBN) was therefore used to reduce the problem of uncertainty in diagnostic clue assessment, while still considering the dependences between elements in the reasoning sequence. A shallow network was used with an open-tree topology, with eight first-level descendant nodes for the diagnostic clues (evidence nodes), each independently linked by a conditional probability matrix to a root node containing the diagnostic alternatives (decision node). One of the evidence nodes was based on the tissue architecture and the others were based on cell features. The system was designed to be interactive, in that the histopathologist entered evidence into the network in the form of likelihood ratios for outcomes at each evidence node. The efficiency of the network was tested on a series of 110 prostate specimens, subdivided as follows: 22 cases of non-neoplastic prostate or benign prostatic tissue (NP), 22 PINs of low grade (PINlow), 22 PINs of high grade (PINhigh), 22 prostatic adenocarcinomas with cribriform pattern (PACcri), and 22 prostatic adenocarcinomas with large acinar pattern (PAClgac). The results obtained in the benign and malignant categories showed that the belief for the diagnostic alternatives is very high, the values being in general more than 0.8 and often close to 1.0. When considering the PIN lesions, the network classified and graded most of the cases with high certainty. However, there were some cases which showed values less than 0.8 (13 cases out of 44), thus indicating that there are situations in which the feature changes are intermediate between contiguous categories or grades. Discrepancy between morphological grading and the BBN results was observed in four out of 44 PIN cases: one PINlow was classified as PINhigh and three PINhigh were classified as PINlow. In conclusion, the network can grade PIN lesions and differentiate them from other prostate lesions with certainty. In particular, it offers a descriptive classifier which is readily implemented and which allows the use of linguistic, fuzzy variables.
• Stotzka, R., Manner, R., Bartels, P. H., & Thompson, D. (1995). A hybrid neural and statistical classifier system for histopathologic grading of prostatic lesions. Analytical and Quantitative Cytology and Histology, 17(3), 204-218.
  More info

  PMID: 7546055;Abstract: Neural network and statistical classification methods were applied to derive an objective grading for moderately and poorly differentiated lesions of the prostate, based on characteristics of the nuclear placement patterns. A partly trained multilayer neural network was used as a feature extractor. A hybrid classifier system using a quadratic Bayesian classifier applied to these features allowed grade assignment consensus with visual diagnosis in 96% of fields from a training set of 500 fields and in 77% of 130 fields of a test set.
• Thompson, D., Bartels, P. H., Bartels, H. G., & Montironi, R. (1995). Image segmentation of cribriform gland tissue. Analytical and Quantitative Cytology and Histology, 17(5), 314-322.
  More info

  PMID: 8534334;Abstract: OBJECTIVE: To develop procedures for the segmentation of cribriform prostatic glands. STUDY DESIGN: A knowledge-guided procedure following a model-based reasoning process was developed in the context of a set of interacting expert systems for machine vision in histometry. RESULTS: With 78 entities in the knowledge file, fully automated, correct segmentaton of approximately 70-80% of cribriform glands was attained-i.e., outlining of histologic components agreed with visual assessment. Measurement of gland size, shape, lumen area, number of lumina per gland, epithelias layer thickness, degree of cribriformity and determination of completeness of lining of a gland by a basal cell layer could be taken from the correctly segmented images. CONCLUSION: The automated procedure allows a histometric characterization of premalignant and malignant prostatic lesions. Extension of system capabilities to the utilization of spectral information is exprected to allow an increase in the correct segmentation rate.
• Bibbo, M., Xiao, J., Christen, R., Fitzpatrick, B., Galera-Davidson, H., Bartels, P. H., & Minimo, C. (1994). Use of computer graphic filters for the nuclear grading of hematoxylin and eosin-stained specimens from prostatic lesions. Analytical and Quantitative Cytology and Histology, 16(3), 183-188.
  More info

  PMID: 7522452;Abstract: An inexpensive workstation is being developed to assist pathologists in diagnosing routine hematoxylin and eosin-stained slides. A linear discriminant model was applied to karyometric features of prostate lesions, and a grade according to Mostofi was determined from the discriminant values. Twenty cases, five of hyperplasia and five each of carcinoma Mostofi grades I, II and III, for a total of 600 nuclei, were selected to train the model. Computer graphic filters were constructed from the discriminant values. Each segmented nucleus has a colored frame (the graphic filter) displayed around it. The color, determined from discriminant values and correlated with the grades, ranges from green for hyperplasia, yellow for low grade, orange for medium grade and red for high grade. An additional 20 cases, 5 of hyperplasia and 5 each of the Mostofi grades, for a total of 538 nuclei, were selected to test the graphic filter. Ninety-six percent of the hyperplasia nuclei were framed in green, 84% of low grade nuclei were framed in yellow, 90% of medium grade nuclei were framed in orange, and 89% of high grade nuclei were framed in red. These results indicate the potential of the color graphic filter to show the pathologist immediate and accurate visual information about the grade of a nucleus. This method may help with the difficult diagnosis of borderline lesions and may help in making the diagnosis from scanty biopsy material.
• Hamilton, P. W., Anderson, N., Bartels, P. H., & Thompson, D. (1994). Expert system support using Bayesian belief networks in the diagnosis of fine needle aspiration biopsy specimens of the breast. Journal of Clinical Pathology, 47(4), 329-336.
  More info

  PMID: 8027370;PMCID: PMC501936;Abstract: Aim - To develop an expert system model for the diagnosis of fine needle aspiration cytology (FNAC) of the breast. Methods - Knowledge and uncertainty were represented in the form of a Bayesian belief network which permitted the combination of diagnostic evidence in a cumulative manner and provided a final probability for the possible diagnostic outcomes. The network comprised 10 cytological features (evidence nodes), each independently linked to the diagnosis (decision node) by a conditional probability matrix. The system was designed to be interactive in that the cytopathologist entered evidence into the network in the form of likelihood ratios for the outcomes at each evidence node. Results - The efficiency of the network was tested on a series of 40 breast FNAC specimens. The highest diagnostic probability provided by the network agreed with the cytopathologists' diagnosis in 100% of cases for the assessment of discrete, benign, and malignant aspirates. Atypical probably benign cases were given probabilities in favour of a benign diagnosis. Suspicious cases tended to have similar probabilities for both diagnostic outcomes and so, correctly, could not be assigned as benign or malignant. A closer examination of cumulative belief graphs for the diagnostic sequence of each case provided insight into the diagnostic process, and quantitative data which improved the identification of suspicious cases. Conclusion - The further development of such a system will have three important roles in breast cytodiagnosis: (1) to aid the cytologist in making a more consistent and objective diagnosis; (2) to provide a teaching tool on breast cytological diagnosis for the non-expert; and (3) it is the first stage in the development of a system capable of automated diagnosis through the use of expert system machine vision.
• Mariuzzi, G. M., Mariuzzi, L., Mombello, A., Santinelli, A., Valli, M., Rahal, D., Thompson, D., & Bartels, P. H. (1994). Quantitative study of ductal breast cancer progression. Morphometric evaluation of phenotypical changes occurring in benign and preinvasive epithelial lesions. Pathology Research and Practice, 190(11), 1056-1065.
  More info

  PMID: 7746739;Abstract: Fifty-nine cases of Breast Epithelial Proliferative Lesions (BEPL) and Ductal Carcinoma in Situ (DCIS), were studied by image analysis, to evaluate the nuclear changes occurring in the conventional diagnostic categories of ductal hyperplasia, atypical ductal hyperplasia and DCIS with quantitative methods. Diagnosis reproducibility is the main practical problem of these breast lesions. In fact, with subjective methods, the reproducibility appears to be very low and precarious especially for clinical demands. The objective, quantitative evaluation of cell phenotypical changes should be the method for both practical diagnostic problems and study of ductal cancer progression. The distribution pattern of the data in the feature, obtained with quantitative analysis, strongly suggests a continuum of changes, indicating an evolutionary process of Breast Ductal Carcinoma (BDC) progression in its preinvasive stage. Each observed case may be characterised by its own cellular, objective alterations and a progressive trend toward BDC can be stated. Since the actual changes of the proliferative phenotypes can be measured, and the values are reproducible, karyometric measurement may allow an objective grading of individual cases.
• Minimo, C., Galera-Davidson, H., Xiao, J., Christen, R., Fitzpatrick, B., Bartels, P. H., & Bibbo, M. (1994). Importance of different nuclear morphologic patterns in grading prostatic adenocarcinoma: An expanded model for computer graphic filters. Analytical and Quantitative Cytology and Histology, 16(5), 307-314.
  More info

  PMID: 7840836;Abstract: The aim of this work was to continue the development of an interactive workstation for the nuclear grading of prostatic lesions by including a large range of nuclear patterns. A previous model was based on four groups: hyperplasia, Mostofi grade 1, Mostofi grade 2 and Mostofi grade 3. Each group included the most common nuclear patterns of the lesions. One set used to test the model included cases showing patterns different from the typical ones of the model. Poor results were obtained for low and medium grades. A review of all the cases in our database led to the conclusion that different nuclear patterns can belong to the same 'nuclear grade.' Thus, in this work the model was expanded to include six groups: hyperplasia, two subgroups for Mostofi grade 1, two subgroups for Mostofi grade 2 and Mostofi grade 3. A set of 900 nuclei, 150 in each group, was selected to test the model. An additional 300 nuclei, 50 in each group, were used for a test set. The overall success rate for classifying the nuclei in the test set using the new model was 93% as compared to a rate of 71% obtained for the similar test set, described above, using the previous model. Moreover, correlating karyometric features with nuclear morphology indicated a role for nucleoli in nuclear grading. The good results obtained with large and heterogeneous sets of cases indicate that the procedures used to develop this model may be adapted for the development of models for the nuclear grading of other tumors.
• Montironi, R., Bartels, P. H., Thompson, D., Scarpelli, M., & Hamilton, P. W. (1994). Prostatic intraepithelial neoplasia: Development of a Bayesian belief network for diagnosis and grading. Analytical and Quantitative Cytology and Histology, 16(2), 101-112.
  More info

  PMID: 8043157;Abstract: The diagnosis and grading of prostatic intraepithelial neoplasia (PIN) are affected by uncertainties that arise from the fact that almost all our knowledge of PIN histopathology is not expressed in numeric form but rather in descriptive linguistic terms, words and concepts. A Bayesian belief network (BBN) was used to reduce the problem of uncertainty in diagnostic clue assessment while considering the dependencies between elements in the reasoning sequence. A shallow network was developed with an open-tree topology, with a root node containing the diagnostic alternatives and seven first-level descendant nodes for the diagnostic clues. One of these nodes was based on tissue architecture and the others on cell features. The results obtained with prototypes of relative likelihood ratios showed that beliefs for the diagnostic alternatives are very high. The network can grade and differentiate PIN lesions from other prostate lesions with certainty. A number of diagnostic clues greater than seven did not significantly improve network performance, whereas a reduced number of clues resulted in decreased beliefs. A BBN for PIN diagnosis and grading offers a descriptive classifier that is readily implemented and allows the use of linguistic, fuzzy variables. A BBN allows the accumulation of evidence presented by diagnostic clues, each offering only weak evidence.
• Scarpelli, M., Bartels, P. H., Montironi, R., Galluzzi, C. M., & Thompson, D. (1994). Morphometrically assisted grading of astrocytomas. Analytical and Quantitative Cytology and Histology, 16(5), 351-356.
  More info

  PMID: 7840841;Abstract: Alcohol-fixed, toluidine-blue-stained smears from 24 astrocytomas (12 low grade and 12 high grade or anaplastic) were included in the study. In each case 50 nuclei from representative areas of the tumor were selected for analysis; quantitative features pertaining to both the entire nucleus and the nucleoli were computed. Nuclear features were nuclear area and total optical density. Nucleolar features were number of nucleoli per nucleus, nucleolar area, variance of the nucleolar area, and mean and variance of the distance of nucleoli from the nuclear membrane. The results showed distinct changes in a number of nuclear and nucleolar features from low to high grade astrocytomas. Features expressing the most pronounced nuclear changes were area, grey level nonuniformity and run percentage. Changes were also found in the following nucleolar features: number of nucleoli, nucleolar area, variance of nucleolar area, nucleolar location and variance of nucleolar location. Linear discriminant analysis was carried out to determine a direction in feature space along which astrocytomas of low and high grade might be ranked. The nuclear area, number of low gray value pixels and a run length feature provided a useful linear combination. The study showed that one can derive a set of objective criteria from morphometric measurements that allows an ordering of astrocytoma cases along an axis and that might be used for continuous grading.
• Xiao, J., Christen, R., Minimo, C., Bartels, P. H., & Bibbo, M. (1994). An algorithm for automatic tracking of nuclear boundaries. Analytical and Quantitative Cytology and Histology, 16(4), 240-246.
  More info

  PMID: 7945699;Abstract: An image segmentation algorithm, based on boundary tracking, was introduced to achieve automatic segmentation of nuclei. This will improve reliability and reproducibility for the computer-assisted grading of routinely stained material, especially from biopsies, which often offer only scanty clinical material. Nuclear grading systems using karyometric features were developed earlier. However, hematoxylin and eosin-stained tissues have proven difficult for automatic segmentation, which is a crucial part of an objective grading system. In this paper we describe an automatic tracking method that traces nuclear boundaries on the basis of edge information and local boundary features. There were two phases to the procedure. First, approximate boundaries were extracted by automatic thresholding; then, boundaries were refined through interactive tracking. The results are encouraging.
• Bartels, P. H., Thompson, D., & Weber, J. E. (1993). Diagnostic decision support by inference networks. In Vivo, 7(4), 379-386.
  More info

  PMID: 8218983;Abstract: Inference networks permit combining of diagnostic evidence in such a fashion that the mutual dependence structure of different pieces of evidence is considered, and that a probabilistic measure of the uncertainty of the final diagnostic decision is provided. Operated in an automatic reasoning mode, an inference network allows a decoupling of the false negative rate from the false positive rate in diagnostic procedures involving rare event detection such as the prescreening for cervical cancer.
• Bibbo, M., Bartels, P. H., Pfeifer, T., Thompson, D., Minimo, C., & Davidson, H. G. (1993). Belief network for grading prostate lesions. Analytical and Quantitative Cytology and Histology, 15(2), 124-135.
  More info

  PMID: 8318127;Abstract: A Bayesian belief network for grading prostatic lesions into eight primary Gleason grades was developed and tested. The network employs 13 diagnostic clues, 8 based on tissue architectural features and 5 based on nuclear features. For every diagnostic clue, three to five different outcomes are specified by membership functions. The network works in a robust fashion and attained agreement with consensus visual grading in 241 of 256 microscopic fields.
• Christen, R., Xiao, J., Minimo, C., Gibbons, G., Fitzpatrick, B. T., Galera-Davidson, H., Bartels, P. H., & Bibbo, M. (1993). Chromatin texture features in hematoxylin and eosin-stained prostate tissue. Analytical and Quantitative Cytology and Histology, 15(6), 383-388.
  More info

  PMID: 7507673;Abstract: A pilot study was undertaken to determine the expression of certain nuclear features in prostatic lesions. Twenty cases, 5 of hyperplasia and 5 each of carcinoma, Mostofi grades I-III, were selected as a training set, and an additional 20 cases were used as a test set, including 5 cases of hyperplasia and 5 cases each in Mostofi grades I-III. Images of hematoxylin and eosin-stained, 4-μm paraffin sections were obtained with a JVC BY-110 three-color camera and digitized by an IBM personal computer with a Matrox MVP-AT/NP imaging board. Thirty nuclei for each case from the training set, for a total of 600 nuclei, and 10 nuclei for each case from the test set, for a total of 200 nuclei, were analyzed by quantitative cytometric software on a SUN 3/60 workstation. A linear discriminant model was used for statistical analysis. One hundred percent of the hyperplasia group, 98% of the low grade group, 92% of the medium grade group and 82% of the high grade group were classified correctly in the test set with an overall success rate of 93%. Statistically significant chromatin texture features included heterogeneity, condensation, margination, run length nonuniformity, long run emphasis, gray level nonuniformity and inertia. Area, roundness and staining intensity (total extinction) were also significant. The results with standard hematoxylin and eosin-stained tissue sections were similar to those previously obtained with Feulgen-stained material. These results indicate that routine hematoxylin and eosin-stained material offers consistent diagnostic clues.
• Thompson, D., Bartels, P. H., Bartels, H. G., Hamilton, P. W., & Sloan, J. M. (1993). Knowledge-guided segmentation of colorectal histopathologic imagery. Analytical and Quantitative Cytology and Histology, 15(4), 236-246.
  More info

  PMID: 7691061;Abstract: Scene segmentation in images of sections of colorectal glands is guided by a knowledge-based system. The construction of the knowledge file, processing sequences and computation of a crowding index for adenomatous glands are described. The histologic sections were stained by hematoxylin and eosin, and the image segmentation was difficult. The knowledge-based system has to determine, on the basis of context and logic, which detected object groups belong to which histologic components. Image segmentation was followed by a reconstruction process. Acceptable segmentation was attained in approximately 85% of the glands processed.
• Bartels, P. H. (1992). Computer-generated diagnosis and image analysis: An overview. Cancer, 69(6 SUPPL.), 1636-1638.
  More info

  PMID: 1540904;Abstract: Image analysis provides quantitative data on morphology, cytochemical and histochemical reactions, the location of specific events or reaction sites, and statistical descriptions of the spatial distribution of such events relative to histologic structure. Image analytic methodology in correlation with histopathologic knowledge bases is an essential component in the development of an objective basis for histopathologic diagnostic decision making.
• Bartels, P. H., Thompson, D., & Weber, J. E. (1992). Construction of the knowledge file for an image understanding system. Pathology Research and Practice, 188(4-5), 396-404.
  More info

  PMID: 1409064;Abstract: To enable an image understanding system to provide an automated interpretation of diagnostic imagery it must have access to all of the concepts, procedures and methods used by human experts. The paper describes information elicitation from experts of different domains and the construction of a knowledge file. Uncertainty management is based on Bayesian belief network methods.
• Bartels, P. H., Thompson, D., & Weber, J. E. (1992). Expert systems in histopathology: IV. The management of uncertainty. Analytical and Quantitative Cytology and Histology, 14(1), 1-13.
  More info

  PMID: 1558613;Abstract: Expert systems deal with data that are categorical and conceptual, that represent elements of fuzzy sets and that often, by themselves, do not allow an unequivocal decision. The management of uncertainty in expert systems thus becomes a crucial issue. It involves defining measures of uncertainty and procedures for combining accumulating evidence in a manner that properly considers the dependence structure of diagnostic clues. Probability theory offers valuable procedures for uncertainty assessment; however, their practical application in the domain of quantitative histopathology and histopathologic diagnosis can be problematic.
• Bartels, P. H., Thompson, D., & Weber, J. E. (1992). Expert systems in histopathology: V. DS theory, certainty factors and possibility theory. Analytical and Quantitative Cytology and Histology, 14(3), 165-174.
  More info

  PMID: 1418266;Abstract: Uncertainty management for the evaluation of evidence based on linguistic and conceptual data is taking advantage of developments in the Dempster- Shafer (DS) theory of evidence, possibility theory and fuzzy logic. The DS theory offers the capability to assess the uncertainty of different subsets of assertions in a domain and the way in which uncertainty is affected by accumulating evidence. The DS theory goes beyond probability theory in its ability to represent ignorance about certain aspects of a situation. However, the theory is very sensitive to the numerical assessments provided by users and can lead to intuitively unexpected and even undesirable results. Certainty factors are widely used in various expert systems. Their definition and updating may follow either a probabilistic model or fuzzy set theoretic concept.
• Bartels, P. H., Thompson, D., Bibbo, M., & Weber, J. E. (1992). Bayesian belief networks in quantitative histopathology. Analytical and Quantitative Cytology and Histology, 14(6), 459-473.
  More info

  PMID: 1292445;Abstract: Bayesian belief networks have a dynamic range and numeric response characteristics that make them uniquely suitable for descriptive classification schemes. Features showing considerable overlap of tolerance regions may be used, in a cumulative manner, to derive unequivocal classification decisions. The numeric response characteristics of Bayesian belief networks are analyzed, and their application as control modules in automated scene segmentation in histopathology is demonstrated.
• Bartoo, G. T., Lee, J. S., Bartels, P. H., Kiviat, N. B., & Nelson, A. C. (1992). Automated prescreening of conventionally prepared cervical smears: A feasibility study. Laboratory Investigation, 66(1), 116-122.
  More info

  PMID: 1731146;Abstract: The feasibility of an automated screening device that will analyze conventionally prepared cervical smears was examined. Consecutive normal and abnormal cervical smears were selected retrospectively and analyzed on a customized microscope imaging workstation system. Specialized image processing, feature extraction, and object classification algorithms were integrated on the imaging workstation to provide an Analysis Score for each slide analyzed. A threshold was set for the Analysis Score, and all slides with a score less than the threshold were classified as normal. Examination of the distribution of the Analysis Score for normal and abnormal cervical smears provided an estimate that 50% of all slides will be sorted as normal, and consequently need no cytologist review. With the Analysis Score threshold set for a 50% sort rate, zero false-negative slides were found in this slide set. This study has shown that a large percentage of slides can be correctly classified as normal while maintaining a high sensitivity to abnormal slides. Given the current workload problems in cytology laboratories, reducing the screening workload by one half will be beneficial. This study has shown that it is now feasible with current technology to provide a clinically useful device to automate screening of conventional cervical smears.
• Duffy, F. H., Jones, K., Bartels, P., McAnulty, G., & Albert, M. (1992). Unrestricted principal components analysis of brain electrical activity: Issues of data dimensionality, artifact, and utility. Brain Topography, 4(4), 291-307.
  More info

  PMID: 1510873;Abstract: Principal components analysis (PCA) was performed on the 1536 spectral and 2944 evoked potential (EP) variables generated by neurophysiologic paradigms including flash VER, click AER, and eyes open and closed spectral EEG from 202 healthy subjects aged 30 to 80. In each case data dimensionality of 1500 to 3000 was substantially reduced using PCA by magnitudes of 20 to over 200. Just 20 PCA factors accounted for 70% to 85% of the variance. Visual inspection of the topographic distribution of factor loading scores revealed complex loadings across multiple data dimensions (time-space and frequency- space). Forty-two non-artifactual factors were successful in classifying age, gender, and a separate group of 60 demented patients by linear discriminant analysis. Discrimination of age and gender primarily involved EP derived factors, whereas dementia primarily involved EEG derived factors. Thirty- eight artifactual factors were identified which, alone, could not discriminate age but were relatively successful in discriminating gender and dementia. The need to parsimoniously develop real neurophysiologic measures and to objectively exclude artifact are discussed. Unrestricted PCA is suggested as a step in this direction.
• Lee, J. S., Bannister, W. I., Kuan, L. C., Bartels, P. H., & Nelson, A. C. (1992). A processing strategy for automated Papanicolaou smear screening. Analytical and Quantitative Cytology and Histology, 14(5), 415-425.
  More info

  PMID: 1338567;Abstract: A multilayer processing strategy was developed for the automatic screening of conventionally prepared Papanicolaou smears. The processing stages include image segmentation, feature extraction, object classification and slide classification. Mathematical morphology functions were implemented in hardware with custom-built gate array processors for image segmentation. There were 68 features used for classifier training. In object classification we combined the evidential supports of a binary decision tree classifier and a multilayer perceptron classifier to achieve an integrated decision. In this feasibility study, 449 conventionally prepared cervical Papanicolaou smears were tested in a prototype research system between January and May 1991. The 95% confidence interval for the slide false-negative rate was 1-9%, and the 95% confidence interval for the slide sort rate was 45-55%. The estimated sort rate for clearly normal slides is within the range required for a cost- efficient screening system, and the estimated false-negative rate for premalignant and malignant smears is an improvement over published false- negative rates for human performance. Several performance improvement efforts are still under way. We expect that they will result in a vastly reduced slide false-negative rate.
• Bibbo, M., Kim, D. H., Pfeifer, T., Dytch, H. E., Galera-Davidson, H., & Bartels, P. H. (1991). Histometric features for the grading of prostatic carcinoma. Analytical and Quantitative Cytology and Histology, 13(1), 61-68.
  More info

  PMID: 2025375;Abstract: Histometric features for the objective grading of prostatic adenocarcinoma in histologic specimens were analyzed in five cases each of well, moderately and poorly differentiated lesions. Tissue sections from the selected cases were stained by the Feulgen method and digitized by a video-based microphotometer. Twenty total fields were recorded for each grade: ten at high resolution (an image sampling of 0.5 μm per pixel) and ten at low resolution (0.8 μm per pixel), with two fields per case recorded at each resolution. The images were segmented by an automated expert system-guided scene segmentation procedure. The performance of that procedure was measured by comparing the automated counts of nuclei in the segmented fields to the visual counts made by a pathologist in the same fields. For well, moderately and poorly differentiated cases, respectively, the nuclear counts made by the expert system at high resolution were 2.7%, 4.2% and 4.7% higher than the visual counts (as estimated from a total of 6,628 nuclei), but 1.2%, 2.5% and 1.1% lower at low resolution (10,329 nuclei). High-resolution features and tissue textural features were computed for each case. The high-resolution features showed good separation between the three groups of cases. The tissue textural features showed consistent separation between well and moderately differentiated cases. The relaxation of the spatial resolution (to 0.8 μm/pixel spacing) did not affect the selection of features, but led to less separation between the data from different grades. In conclusion, the automated system performed satisfactorily in distinguishing sections of prostatic tumors of varying degrees of differentiation.
• Michelassi, F., Montag, A., Bartels, P. H., Dytch, H., Bania, C., Lerma, E., & Bibbo, M. (1991). Nuclear changes in the perineoplastic colonic mucosa: Is the mucosa near a colon cancer normal?. Research in Surgery, 3(2), 127-132.
• Minimo, C., Bartels, P. H., Kim, D. H., Pfeifer, T., Dytch, H. E., Galera-Davidson, H., & Bibbo, M. (1991). Value of the cribriformity factor in grading prostatic carcinoma. Analytical and Quantitative Cytology and Histology, 13(6), 411-417.
  More info

  PMID: 1666955;Abstract: Architectural and histometric features for the objective grading of prostate adenocarcinoma in histologic specimens were analyzed in five cases each of Gleason primary grades 2, 3A, 3B, 3C, 4A and 4B, selected as ''typical'' for the histopathologic images. Tissue sections from the selected cases were stained by the Feulgen method. Fifteen fields for each grade, for a total of 4,430 glands, were digitized by a video-based microphotometer at low resolution (pixel spacing of 2 μm). Outer and inner outlines of the glandular epithelium were traced manually using a mouse. For each field the number of glands, the gland area, the lumen area, the area of the glandular epithelium and the cribriformity factor were computed. The gland area and its variance proved to be useful indicators for lower-grade lesions, whereas the variance of cribriformity resulted in an excellent grading indicator in the Gleason 3-4 range when cribriform glands were present.
• Montag, A. G., Bartels, P. H., Dytch, H. E., Lerma-Puertas, E., Michelassi, F., & Bibbo, M. (1991). Karyometric features in nuclei near colonic adenocarcinoma: Statistical analysis. Analytical and Quantitative Cytology and Histology, 13(3), 159-167.
  More info

  PMID: 1716896;Abstract: The normal mucosa adjacent to colonic adenocarcinoma (marginal or transitional mucosa) has been shown to have subtle alterations of architecture, surface glycoproteins and proliferative activity. To evaluate possible changes in nuclear configurations in this marginal mucosa, a large set of cytometric features was evaluated using a computer-assisted video analysis system. Preliminary statistical analysis of the measurements identified six nuclear features useful for discriminating marginal mucosa nuclei from normal (control) mucosa nuclei: total optical density (OD), nuclear area, chromatin texture (from gray value cooccurrence matrix), chromatin coarseness, average OD of nuclear staining and peripheral tendency of the chromatin in the nucleus. An analysis of variance revealed that both patient-to-patient and gland-to-gland variation would limit the usefulness of any one feature as a screening tool. As a group, however, these six features should be investigated further as markers of preneoplastic changes in histologically normal-appearing mucosa.
• Bartels, P. H., Thompson, D., & Weber, J. E. (1990). Conceptual learning in an expert system for histopathologic diagnosis. Proceedings of SPIE - The International Society for Optical Engineering, 1206, 7-11.
  More info

  Abstract: Autonomous learning modules in a diagnostic expert system serve to reveal distributional properties of the diagnostic clue values. This leads to more exhaustive utilization of the collected information. It also results in matching the design granularity for the diagnostic discrimination to the true structure of the diagnostic data. Additional conceptual entities augment the knowledge base.
• Bibbo, M., Bartels, P. H., Salguero, M., Dytch, H. E., Lerma-Peurtas, E., & Galera-Davidson, H. (1990). Karyometric marker features in fine needle aspirates of microinvasive follicular carcinoma of the thyroid. Analytical and Quantitative Cytology and Histology, 12(1), 42-47.
  More info

  PMID: 1689163;Abstract: Karyometric measurements were performed on fine needle aspirates of clearly identifiable tumor areas and adjacent normal-appearing areas in the surgical specimens from ten patients with microinvasive follicular carcinoma of the thyroid. Similar measurements were performed of aspirates from nine patients free of thyroid disease (controls). A total of 95 karyometric features were evaluated for each nucleus. As compared with the control nuclei, the normal- appearing nuclei showed a 6% increase in total nuclear optical density (OD) while the tumor nuclei showed a 14% increase. Analysis of variance indicated a significant difference between the normal-appearing nuclei and the control nuclei, with most of the difference due to the differences of tissue origin. Discriminant analysis selected nine features that produced a statistically highly significant separation of tumor nuclei from control nuclei. A similar analysis selected five features that produced a statistically highly significant discrimination of normal-appearing nuclei from control nuclei; the validity of those karyometric features as markers of malignancy in normal-appearing nuclei from tissues adjacent to microinvasive follicular carcinomas of the thyroid was demonstrated by analysis in further training and test sets of nuclei.
• Bibbo, M., Kim, D. H., Loreto, C. d., Dytch, H. E., Galera-Davidson, H., Thompson, D., Richards, D. L., Bartels, H. G., & Bartels, P. H. (1990). Tissue architectural features for the grading of prostatic carcinoma. Analytical and Quantitative Cytology and Histology, 12(4), 229-236.
  More info

  PMID: 2206192;Abstract: In research for the development of a computer-aided workstation for the objective grading of prostatic carcinoma, tissue architectural (histometric) features were analyzed in ten cases each of well-differentiated, moderately differentiated and poorly differentiated carcinoma (as subjectively graded by the consensus of a panel of experts). Sections were cut at 4 μm, stained by the Feulgen reaction and digitized by two different video-based photometric systems. Some images were interactively segmented, considering the histometric clues to be studied; others were automatically segmented by an expert system-guided technique. The latter procedure produced good results, with over 90% of the nuclei judged to be correctly segmented in 64% of the fields studied and over 80% in another 24% of the fields. While the number of nuclei per field provided some separation of well-differentiated from other lesions, the number of nuclei per gland distinguished between well-differentiated and moderately differentiated lesions. Simplicial decomposition of the images also provided a measure of the degree of differentiation, as did the ''texture'' of the nuclear placement, based on two run-length statistics. Combination of the run-length features distinguished the three categories of lesions with statistical significance. The results of this study provided insights into the problems (such as the effect of field boundaries) faced in the design of an computer-aided grading system. They also showed the value of expert system-guided scene segmentation and of such histometric features as the field cellularity and the number of nuclei per gland for the discrimination between lesions of different grades of differentiation.
• Bibbo, M., Michelassi, F., Bartels, P. H., Dytch, H., Bania, C., Lerma, E., & Montag, A. G. (1990). Karyometric marker features in normal-appearing glands adjacent to human colonic adenocarcinoma. Cancer Research, 50(1), 147-151.
  More info

  PMID: 1688372;Abstract: The expression of nuclear marker features in normal-appearing tissue adjacent to colonic adenocarcinoma was investigated. Formalin-fixed, paraffin-embedded tissue sections of colon from 9 patients with adenocarcinoma and from 9 normal controls were cut 4 μm thick, Feulgen stained, and measured by a cell image analysis system using a Matrox MVP-AT/NP imaging board. Thirty nuclei in the tumor region, 30 nuclei 2 mm into the histologically normal-appearing distal margin, and the same number at 5, 10, 20, and 50 mm into the margin were measured for each patient. An additional 30 nuclei were recorded from 9 patients each free from colonic disease. Nuclear features were selected to discriminate between tumor nuclei and nuclei from normal control subjects and between nuclei measured in the histologically normal-appearing margin next to the tumor and control nuclei. Eight micromorphometric measures were found to be statistically significantly different in nuclei measured in the margin site, including features describing staining density (total absorbance, average absorbance 20% below mean, average absorbance 20% above mean) chromatin texture (cooccurrence matrix, run length, and peripheral tendency) and nuclear area. The category differences are statistically highly significant.
• Duffy, F. H., Jones, K., Bartels, P., Albert, M., McAnulty, G. B., & Als, H. (1990). Quantified neurophysiology with mapping: Statistical inference, exploratory and confirmatory data analysis. Brain Topography, 3(1), 3-12.
  More info

  PMID: 2094310;Abstract: Topographic mapping of brain electrical activity has become a commonly used method in the clinical as well as research laboratory. To enhance analytic power and accuracy, mapping applications often involve statistical paradigms for the detection of abnormality or difference. Because mapping studies involve many measurements and variables, the appearance of a large data dimensionality may be created. If abnormality is sought by statistical mapping procedures and if the many variables are uncorrelated, certain positive findings could be attributable to chance. To protect against this undesirable possibility we advocate the replication of initial findings on independent data sets. Statistical difference attributable to chance will not replicate, whereas real difference will reproduce. Clinical studies must, therefore, provide for repeat measurements and research studies must involve analysis of second populations. Furthermore, Principal Components Analysis can be employed to demonstrate that variables derived from mapping studies are highly intercorrelated and data dimensionality substantially less than the total number of variables initially created. This reduces the likelihood of capitalization on chance. The need to constrain alpha levels is not necessary when dimensionality is low and/or a second data set is available. When only one data set is available in research applications, techniques such as the Bonferroni correction, the ''leave-one-out'' method, and Descriptive Data Analysis (DDA) are available. These techniques are discussed, clinical and research examples are given, and differences between Exploratory (EDA) and Confirmatory Data Analysis (EDA) are reviewed.
• Galera-Davidson, H., Bartels, P. H., Fernandez-Rodriguez, A., Dytch, H. E., Lerma-Puertas, E., & Bibbo, M. (1990). Karyometric marker features in fine needle aspirates of invasive follicular carcinoma of the thyroid. Analytical and Quantitative Cytology and Histology, 12(1), 35-41.
  More info

  PMID: 1689162;Abstract: Karyometric measurements were performed on fine needle aspirates of clearly identifiable tumor areas and adjacent normal-appearing areas in the surgical specimens from ten patients with invasive follicular carcinoma of the thyroid. Similar measurements were performed on aspirates from nine patients free of thyroid disease (controls). A total of 95 karyometric features were evaluated for each nucleus. Analysis of variance of optical density values indicated (1) a similarity between tumor and normal-appearing nuclei from carcinoma cases, (2) a significant difference between those nuclei and control nuclei and (3) that most of the differences were due to the differences of tissue origin. Stepwise linear discriminant analysis selected ten features that produced a statistically highly significant separation of tumor nuclei from control nuclei. A similar analysis selected six features that produced a statistically highly significant discrimination of normal-appearing nuclei from control nuclei; the validity of those karyometric features as markers of malignancy in normal-appearing nuclei from tissues adjacent to invasive follicular carcinomas of the thyroid was demonstrated by analysis in further training and control sets of nuclei. This analysis in thyroid aspirates identified more marker features than did a previous similar analysis using tissue sections.
• Galera-Davidson, H., Gonzalez-Campora, R., Mora-Marin, J., Matilla-Vicente, A., Dytch, H. E., Bartels, P. H., Lerma-Puertas, E., Andrada-Becerra, E., & Bibbo, M. (1990). Cytophotometric DNA measurements in medullary thyroid carcinoma. Cancer, 65(10), 2255-2260.
  More info

  PMID: 2346910;Abstract: The prognostic significance of objectively measured karyometric variables (ploidy pattern, nuclear roundness, area, elongation, chromatin texture, and nearest nucleus distance) was investigated in relation to clinical (stage and type of disease) and morphologic (histologic patterns) variables in 27 patients with the diagnosis of medullary thyroid carcinoma (MTC). The DNA and karyometric measurements of Feulgen-stained nuclei were made with a video cytometry system. The five-year and ten-year adjusted survival rates were 74.4 ± 10.1% and 59.5 ± 15.6%, respectively. Cox's survival analysis for mortality showed that only stage, age, sex, and 5N exceeding rate had predictive value (overall P = 0.0012) in decreasing order. Patients with the best prognosis were young females with clinical Stage I disease and low 5N exceeding rate tumors. When karyometric and histometric variables were considered by themselves survival correlates with the standard deviation (SD) of the nearest nuclear distance and nuclear elongation; that is, patients with crowded, high cellularity tumors and elongated cells had the worst prognosis. In univariate analyses only clinical stage correlated with adjusted survival rate. Multivariate survival analysis for morbidity showed that patients in Stages ≥ II and high SD of ploidy values were free of symptoms for short intervals. When morphometric data were considered alone, patients with high variance in the chromatin texture and highly variable nuclear areas had shorter asymptomatic intervals.
• Graham, A. R., Paplanus, S. H., & Bartels, P. H. (1990). A diagnostic expert system for colonic lesions. American Journal of Clinical Pathology, 94(4), S15-S18.
  More info

  PMID: 2220680;Abstract: The diagnostic expert system for colonic lesions (DESCL) was designed to discriminate colonic adenoma and adenocarcinoma from normal colonic tissue. Although it was originally developed for use in conjunction with a machine vision analytic system, the DESCL had evolved into a teaching tool and a model of conceptual machine learning. The expert system is table driven and consists of a shell and a knowledge base. The latter comprises a series of architectural and cytologic observations and a quantitative estimate of diagnostic importance relating these observations to diagnostic outcome. In a validation study of 100 colonic lesions, the expert sysem achieved a success rate of 98%. It has the flexibility to allow individual pathologists to 'customize' the knowledge base to suit their diagnostic criteria.
• Lerma-Puertas, E., Galera-Davidson, H., Bartels, P. H., Kim, D. H., Dytch, H. E., & Bibbo, M. (1990). Karyometric marker features in fine needle aspirates of follicular adenoma of the thyroid. Analytical and Quantitative Cytology and Histology, 12(4), 223-228.
  More info

  PMID: 2206191;Abstract: Karyometric measurements were performed on fine needle aspirates of clearly identifiable adenomatous areas and adjacent normal-appearing areas in the surgical specimens from ten patients with follicular adenomas of the thyroid. Similar measurements were made on aspirates from nine patients free of thyroid disease. A total of 95 karyometric features were evaluated for each nucleus. Analysis of variance of optical density values did not show a significant difference between the three types of nuclei. Discriminant analysis selected seven karyometric features that produced a statistically highly significant separation of adenoma nuclei from control nuclei. A similar analysis selected six features that produced a statistically highly significant discrimination of normal-appearing nuclei from control nuclei. The validity of these markers for distinguishing control nuclei from adenoma nuclei and normal-appearing nuclei adjacent to adenomas was demonstrated by analysis in further training and test sets. These findings parallel those previously demonstrated for invasive and microinvasive follicular carcinomas of the thyroid.
• Shack, R., Bartels, P. H., Hillman, D., Sabb, M., & Shoemaker, R. (1990). Confocal microscope with enhanced depth discrimination. Proceedings of SPIE - The International Society for Optical Engineering, 1206, 12-18.
  More info

  Abstract: Confocal microscopy, in comparison to conventional microscopy, offers a narrower point-spread function, the rendition of phase structures in contrast, and high axial resolution. Of these, the capabilities of optical sectioning and the three-dimensional imaging of cells and tissues have attracted the most interest. The literature offers a number of studies exploring the factors affecting depth of field in confocal microscopy. In this article, the three-dimensional representation of fluorescence imagery is examined and the design of a laser scanning confocal fluorescence microscope with enhanced depth discrimination is described.
• Shoemaker, R. L., Thompson, D., Griswold, W. G., & Bartels, P. H. (1990). Performance and task scheduling studies of a multiprocessor in histopathologic image analysis. Proceedings of SPIE - The International Society for Optical Engineering, 1206, 31-39.
  More info

  Abstract: The performance of a multiprocessor computer system in histopathology applications was studied. Specifically, timing studies have been done on an operational multiprocessor, the Heidelberg Polyp, running expert system-guided scene segmentation and image analysis software. A number of actual and potential system bottlenecks and methods of attack for removing them are discussed.
• Thompson, D., Bartels, H. G., Haddad, J. W., & Bartels, P. H. (1990). Scene segmentation in a machine vision system for histopathology. Proceedings of SPIE - The International Society for Optical Engineering, 1206, 40-47.
  More info

  Abstract: Algorithms and procedures employed to attain reliable and exhaustive segmentation in histopathologic imagery of colon and prostate sections are detailed. The algorithms are controlled and selectively called by a scene segmentation expert system as part of a machine vision system for the diagnostic interpretation of histopathologic sections. At this time, effective segmentation of scenes of glandular tissues is produced, with the system being conservative in the identification of glands; for the segmentation of overlapping glandular nuclei an overall success rate of approximately 90% has been achieved.
• Weber, J. E., & Bartels, P. H. (1990). Statistical identification of subpopulations for flow cytometric data. Proceedings of SPIE - The International Society for Optical Engineering, 1206, 19-30.
  More info

  Abstract: Identification of cell subpopulations is of interest in the context of both flow cytometry and image analysis. This paper considers statistical identification of subpopulations of cells; analyses using both actual data and simulated data are discussed, the results indicate that identification of subpopulations is feasible, even for large multivariate data sets, when expert system guided preprocessing of the data is completed prior to cluster analysis. The clustering algorithm PINDEX is suggested as being particularly well-suited for identifying cell subpopulations. This algorithm uses both Euclidean distance and Mahalanobis distance in the process of determining clusters. Thus PINDEX takes advantage of the speed of calculation provided by Euclidean distance, which does not take account of covariance structure, and the adaptive aspect of Mahlanobis distance, which takes account of changes in the covariance structure of the emerging clusters. Speed and accuracy are further improved by expert system guided preprocessing of data prior to clustering by PINDEX.
• Wied, G. L., Dytch, H., Bibbo, M., Bartels, P. H., & Thompson, D. (1990). Artificial intelligence - guided analysis of cytologic data. Analytical and Quantitative Cytology and Histology, 12(6), 417-428.
  More info

  PMID: 2078264;Abstract: A design for the integration of artificial intelligence (AI) technology with large databases of clinical and objective cytologic data, such as are on file at the University of Chicago, is presented. Among the key features of this approach are the use of a knowledge representation structure based upon an associative network, the use of a Bayesian belief network as a method of managing uncertainty in the system, and the use of neural networks and unsupervised learning algorithms as a means of discovering patterns within this database. Such an automated approach is necessary, given the complexity and interdependence of these data, to gain an understanding of their dependence structure and to assist in their exploration and analysis.
• Bartels, P. H. (1989). The diagnostic pattern in histopathology. American Journal of Clinical Pathology, 91(4 SUPPL. 1), S7-S13.
  More info

  PMID: 2929515;Abstract: Defining in precise detail the diagnostic clues used by histopathologists to arrive at a diagnosis is a difficult process; however, better analytic understanding of this process is a prerequisite for the design of diagnostic expert systems. Such systems offer the potential for consistency and setting standards in the diagnostic evaluation of difficult situations. Diagnostic clues as offered by the image represent two-dimensional information. Their description is entered into the knowledge base of an aspect system, however, in the form of a description, which is one-dimensional. Preserving all of the two-dimensional dependence structure in this conversion is a fundamental problem. Use of the description to generate simulated imagery tests whether or not the two-dimensional dependence structure of a given tissue architecture has been adequately represented in the analytic description. This analytic description would then be entered into the knowledge base of an expert system.
• Bartels, P. H., & Hiessl, H. (1989). Expert systems in histopathology. II. Knowledge representation and rule-based systems. Analytical and Quantitative Cytology and Histology, 11(3), 147-153.
  More info

  PMID: 2663006;Abstract: Two aspects of expert systems for use in diagnostic histopathology and cytopathology are examined: knowledge representation and the structure and operation of rule-based systems. Knowledge may be represented, e.g., in semantic networks, frames, multiple contexts and model-based structures; the choice of structure should be matched to the type of information to create an efficient and logically adequate expert system. In a rule-based system, knowledge is represented as 'rules', often in the form of 'IF (condition)-THEN (conclusion)' rules. The anatomy of such rules and their operation is explored via the use of examples. Uncertainty in rules is briefly addressed, and their processing by the symbolic reasoning of the 'inference engine' of the expert system is described, including both 'forward-chaining' ('data-driven') operations and 'backward-chaining' ('goal-driven') operations.
• Bartels, P. H., & Thompson, D. (1989). Expert systems in histopathology. III. Representation of knowledge as 'structured objects'. Analytical and Quantitative Cytology and Histology, 11(6), 367-374.
  More info

  PMID: 2690840;Abstract: In designing expert systems for histopathology, the representation of knowledge as 'structured objects' may offer advantages over the use of a set of rules. A structured-object representation using straightforward, descriptive, declarative statements can allow consideration of the relationships between the facts and constraining conditions of a domain in a richer, more complex and complete manner. The use of associative (semantic) networks and frames to organize histopathologic knowledge as structured objects is discussed and demonstrated with examples that consider the complex tissue structures found in the kidney and thyroid. As an example, the design and function of a structured object-based expert system for follicular thyroid aspirates is discussed, including a consideration of the types of information that such a system would require. Such a system utilizes model-based reasoning rather than the rule-based reasoning of rule-based expert systems.
• Bartels, P. H., & Weber, J. E. (1989). Expert systems in histopathology. I. Introduction and overview. Analytical and Quantitative Cytology and Histology, 11(1), 1-7.
  More info

  PMID: 2655647;Abstract: An introduction to, and overview of, expert systems is presented, along with some preliminary comments on their application in diagnostic and analytical histopathology and cytopathology. The terminology common to expert systems is defined, and the nature of expert systems is discussed. In particular, the differences between expert systems and other types of computer programs (e.g., algorithms) or means of solving problems are explored. The rationale for their use and the types of tasks for which they are appropriate are also discussed.
• Bartels, P. H., Bibbo, M., Graham, A., Paplanus, S., Shoemaker, R. L., & Thompson, D. (1989). Image understanding system for histopathology. Analytical Cellular Pathology, 1(4), 195-214.
  More info

  PMID: 2487045;Abstract: An image understanding machine vision system for histological diagnoses is based on three interacting expert systems: a diagnostic expert system utilizing terms familiar to pathologists, an interpretive expert system relating human diagnostic concepts to computable histometric features and a scene segmentation expert system which extracts the diagnostic information from the imagery. The control software for the image understanding system resides on a multiprocessor computer. This article details measures to maintain system efficiency and to accommodate the requirements of interprocess communication and processing task scheduling. © 1989.
• Bartels, P. H., Thompson, D., & Weber, J. E. (1989). Unsupervised conceptual learning in a diagnostic expert system. Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings, 11 pt 6, 1780-1781.
  More info

  Abstract: A diagnostic expert system for assessing colonic sections is augmented by the addition of an unsupervised learning module. A simple distance metric between diagnostic clue sequences, as they occur for each assessed case, is defined. The statistical significance of the modes detected in the conceptual data sets is established. Even the relatively simple unsupervised learning module implemented in this system has led to a number of insights. For example, if learning capability is considered for a diagnostic expert system, it is advisable to use a fine grading and multiple diagnostic clue values. This will allow better resolution by a distance measure. Also, it is possible to establish distances between concepts if an ordering can be attained and, based on such an ordering, the statistical significance of a grouping of cases, described only in conceptual terms, can be established.
• Bartels, P. H., Thompson, D., Bartels, H. G., & Shoemaker, R. (1989). Machine vision system for diagnostic histopathology. Pathology Research and Practice, 185(5), 635-646.
  More info

  PMID: 2626374;Abstract: In a machine vision system for the diagnostic assessment of histopathologic sections, human diagnostic knowledge and human ability to recognize components in a complex image need to be emulated. This is attempted by three integrated expert systems, supported by a multiprocessor computer with data-driven, dynamically reconfigurable architecture.
• Bibbo, M., Dytch, H. E., Alenghat, E., Bartels, P. H., & Wied, G. L. (1989). DNA ploidy profiles as prognostic indicators in CIN lesions. American Journal of Clinical Pathology, 92(3), 261-265.
  More info

  PMID: 2773848;Abstract: The prognostic significance of DNA ploidy measurements in cervical tissues was examined. Microphotometric measurements of 302 Feulgen-stained tissue sections (91 normal squamous epithelia, 14 condylomata, 29 cervical intraepithelial neoplasia (CIN) I, 78 CIN II, and 90 CIN III) were performed with a personal computer (PC)-based video microphotometry system. Analysis of these data shows that the DNA profile provides significant prognostic information: CINs with a polyploid DNA profile are more likely to return to normal than are those exhibiting an aneuploid pattern. Of 211 abnormal cases, 38% had polyploid DNA profiles and 62% were aneuploid. Eighty-six percent of the cases that regressed were polyploid and 14% were aneuploid. Of the 130 aneuploid DNA cases, 95% remained static or progressed and only 5% regressed. Of these nonregressing aneuploid lesions, 90 remained static and 34 progressed, whereas within the nonregressing polyploid group 37 remained static and only 6 progressed. This result holds across diagnostic categories. Several other ploidy-related descriptors also showed prognostic significance (including mean ploidy, the 5N exceeding rate and 2N deviation index, and discriminant functions derived from order statistic analysis of the cumulative DNA histograms), but not to the degree or with the consistency of expression as the DNA profile categorization. These results indicate that important information about the prognosis of CIN lesions may be obtained by the DNA profile on Feulgen-stained tissue specimens. The data were acquired by a cytophotometry system of relatively modest cost consisting of readily available hardware components.
• Bibbo, M., Montag, A. G., Lerma-Puertas, E., Dytch, H. E., Leelakusolvong, S., & Bartels, P. H. (1989). Karyometric marker features in tissue adjacent to invasive cervical carcinomas. Analytical and Quantitative Cytology and Histology, 11(4), 281-285.
  More info

  PMID: 2765075;Abstract: A companion study showed the existence of statistically significant changes in the value of karyometric 'marker features' in the nuclei of histologically normal-appearing ectocervical epithelium adjacent to carcinomas in situ. In this second part of the study, the results obtained in patients with invasive cervical carcinomas were analyzed. Formalin-fixed, paraffin-embedded tissue sections were Feulgen stained and analyzed with a microTICAS video microphotometer. The results, demonstrated for the first time in histologic material, indicate that the marker features are clearly expressed in a majority of nuclei observed in the normal-appearing tissue adjacent to invasive lesions. This effect was statistically significant. The best marker features selected by the discriminant analysis were nuclear roundness, nuclear perimeter length, total optical density and a run length texture measure. These findings may reflect a subtle transformation of the apparently normal cervical epithelium adjacent to an invasive cervical carcinoma.
• Jordan, S. W., Brayer, J. M., Bartels, P. H., & Anderson, R. E. (1989). Quantitative morphology of late renal radiation injury. International Journal of Radiation Oncology, Biology, Physics, 16(1), 101-106.
  More info

  PMID: 2912931;Abstract: Late radiation injury manifests itself in all morphologic compartments of the kidney, but loss of cell mass is most significant in the proximal convoluted tubules. Development of an end stage or nonfunctional kidney requires 12 or more months after single fraction X ray exposures of about 12 Gy (2, 5, 13) and is associated with marked morphologic alterations of renal tubules. Radiation induced changes were studied at 6 months after irradiation, a time interval when histological alterations appear minor, but in previous studies were shown to correlate with the later end stage alterations (5, 8, 9). Renal alterations were graded objectively based on renal weight, variation in size of tubule cell nuclei, and glomerular nuclear volume fraction. Irradiation was associated with loss of renal weight, increased variability of tubule nuclear size, and previously unappreciated changes in glomerular nuclear volume fraction. A classification index derived from a weighted combination of renal weight ratio and tubule cell nuclear variability correlates with radiation dose and with previously established subjective histologic grading of renal damage, and allows objective comparisons of various fractionation schedules. © 1989.
• Montag, A. G., Bartels, P. H., Lerma-Puertas, E., Dytch, H. E., Leelakusolvong, S., & Bibbo, M. (1989). Karyometric marker features in tissue adjacent to in situ cervical carcinomas. Analytical and Quantitative Cytology and Histology, 11(4), 275-280.
  More info

  PMID: 2765074;Abstract: Subtle changes in nuclear chromatin structure have been documented in the histologically normal mucosa adjacent to neoplastic lesions. To evaluate the expression of such 'marker features' in tissue adjacent to squamous carcinomas in situ (CIS) of the uterine cervix, normal-appearing ectocervical tissues from five cases of CIS were compared with ectocervical tissues from control patients with squamous metaplasia. Formalin-fixed, paraffin-embedded tissue sections were Feulgen stained and analyzed with the micro TICAS video microphotometer. Discriminant analysis revealed seven features that helped to distinguish nuclei from ectocervical tissue adjacent to CIS from those of control tissue. These features reflected changes in nuclear shape and chromatin distribution that were not detected by routine histopathologic analysis. The findings may reflect a subtle premalignant change in the apparently normal mucosa adjacent to a cervical neoplasm; they may also reflect the influence of the neoplasm on the adjacent mucosa
• Weber, J. E., & Bartels, P. H. (1989). Colonic lesion expert system. Evaluation of sensitivity. Analytical and Quantitative Cytology and Histology, 11(4), 249-254.
  More info

  PMID: 2669783;Abstract: In a previous study, an expert system using visually assessed diagnostic clues for diagnosing colonic tissue as normal, adenoma or adenocarcinoma arrived at diagnoses agreeing with the evaluation by pathologists ('correct diagnoses') for all 49 cases of normal colon, for 49 of 50 cases of adenoma and for 48 of 49 cases of adenocarcinoma. The present study examined the robustness and sensitivity of the expert system to changes in the knowledge base, to changes in criteria specified by the user and to missing information. Alternative rules for combining certainty factors are discussed.
• Weber, J. E., & Bartels, P. H. (1989). Guided data reduction for flow cytometry. Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings, 11 pt 6, 1783-1784.
  More info

  Abstract: A statistical procedure for identifying subpopulations in large samples is discussed in the context of the analysis of flow cytometric data. The PINDEX algorithm is proposed for analyzing such multimodal data. PINDEX is a clustering algorithm which considers the covariance structure of clusters. When such an algorithm is run in a brute force manner, the number of runs required for convergence may involve computation which, for large samples, is prohibitive. A solution to this problem using starting centroids chosen by an expert system is proposed. Sample runs were made using data sets for lymphoid cells; these data sets are of modest dimensionality and contain four subpopulations. The results demonstrate that substantive economy of processing can be achieved by having the expert system provide start-up cluster centroids close to the final coordinates.
• Wied, G. L., Bartels, P. H., Bibbo, M., & Dytch, H. E. (1989). Image analysis in quantitative cytopathology and histopathology. Human Pathology, 20(6), 549-571.
  More info

  PMID: 2656497;
• Bartels, P. H. (1988). Validation of expert-system image understanding in histopathology. IEEE/Engineering in Medicine and Biology Society Annual Conference, 10 pt 3, 1385-1386.
  More info

  Abstract: The automated diagnostic assessment of tissue sections relies on knowledge-base-supported evaluation of tissue architecture and nuclear placement patterns. Given the multifaceted dependence structure of the relative placement of tissue components, it is not straightforward to ascertain whether the description used for the classification properly considers that dependence. Simulated imagery allows testing of the completeness of the descriptive information used in diagnostic assessment.
• Bartels, P. H., Paplanus, S., Graham, A., & Bibbo, M. (1988). Image understanding in expert systems in histopathology. IEEE/Engineering in Medicine and Biology Society Annual Conference, 10 pt 3, 1373-1374.
  More info

  Abstract: The use of model-based reasoning, combined with locally adaptive selection of segmentation procedures, has already been found productive in expert-system-guided scene segmentation of histopathologic imagery. It applies human understanding of segmentation problems, with suitable remedial procedures, and knowledge of the structure of the tissues to the segmentation. Expert-system-guided scene segmentation thus implements certain aspects of image understanding to attain robustness. For diagnostic expert systems, though, image understanding in a much broader sense is required. A pathologist's verbal description of histopathologic patterns must be related to specific information extraction and analytic processes, which are to be executed by the automated system.
• Bartels, P. H., Weber, J. E., & Duckstein, L. (1988). Machine learning in quantitative histopathology. Analytical and Quantitative Cytology and Histology, 10(4), 299-306.
  More info

  PMID: 3166674;Abstract: The role of expert systems functioning as process controllers in learning image understanding systems is discussed. Numeric learning systems already have found a number of applications in cytologic and histopathologic diagnosis. Depending on the required capabilities, systems of increasing complexity are needed. Expert systems to guide scene segmentation in histopathologic imagery require model-based reasoning. Diagnostic image interpretation with learning capability demands a full model of the human expert's competence, including a considerable variety of knowledge representation schemes and inference strategies, coordinated by a meta-process controller.
• Duckstein, L., Bartels, P. H., & Weber, J. E. (1988). Organization of a knowledge base by Q-analysis. Applied Mathematics and Computation, 26(4), 289-301.
  More info

  Abstract: Q-analysis is used in organizing a histopathological knowledge base which is a component of the diagnostic expert system at the University of Arizona. This expert system has three subsystems or modules: the first module guides the dynamic reconfiguration of the computer system; the second module guides scene decomposition and the extraction of diagnostic information; the third module uses a rule-based system to obtain a diagnostic assessment. A specific example of the usefulness of Q-analysis is given in the context of the third module; the data represent expert opinion concerning diagnosis of colonic cancer and are summarized in a matrix representing four diagnostic categories and nineteen diagnostic clues. The example shows that Q-analysis may be helpful to diagnosticians in defining and explaining the process they use in arriving at a diagnosis and in using this information as a basis for structuring the knowledge base. © 1988.
• Graham, A. R., Paplanus, S. H., & Bartels, P. H. (1988). Micromorphometry of colonic lesions. Laboratory Investigation, 59(3), 397-402.
  More info

  PMID: 3411940;Abstract: In the formulation of a quantitative component for the knowledge base of a diagnostic expert system, we developed a mathematical model to describe nuclear placement patterns in colon. Glands from normal colon, tubular adenomas, and adenocarcinomas were studied, and a ratio developed of percentage internally displaced nuclei to percentage displacement area. This displacement area represented an inner epithelial zone bounded centrally by the lumen and peripherally by the apical limit of basal nuclei. The mean percentage internally displaced nuclei to percentage displacement area ratio was 0.15 ± 0.13 for normal glands, 0.75 ± 0.09 for adenomas, and 1.04 ± 0.13 for carcinomas. When the percentage internally displaced nuclei to percentage displacement area ratio was plotted against total number of nuclei, there was statistically significant separation of the diagnostic categories at α ≤0.05. Relative nuclear area and its variance were also studied for normal colonic glands, adenomas, and adenocarcinomas. Mean relative nuclear area for cells from normal glands was 284.2, from adenomas 342.4, and from adenocarcinomas 416.4, these being statistically significantly different at α ≤0.05. Variance of relative nuclear area, compared for the three diagnostic categories, was statistically significantly increased only in adenocarcinoma. Quantitation of nuclear placement patterns appears to provide a useful diagnostic clue for the knowledge base of a diagnostic expert system in the discrimination among normal colonic glands, adenomas, and adenocarcinomas.
• Jordan, S. W., Brayer, J. M., Bartels, P. H., & Anderson, R. E. (1988). Video-based image collection for quantitative histopathology. Analytical and Quantitative Cytology and Histology, 10(1), 37-46.
  More info

  PMID: 2451527;Abstract: Obtaining histologic images for computer-based morphometric analysis is associated with a number of standardization problems, which must be solved if reproducible data collection is expected. These problems include tissue processing, sectioning and staining, standardizing and calibrating the video camera and determining the appropriate sampling rate (pixels/μm). Suggested solutions for these problems are presented for a specific image analysis system, but are applicable to other systems with similar capabilities. Biologic variability is not eliminated by computer-assisted analysis, so it is important to minimize data-collection artifacts by parallel processing of experimental and control material, as in other investigative work.
• Paplanus, S., Graham, A., Thompson, D., & Bartels, P. H. (1988). Expert system-guided scene segmentation. IEEE/Engineering in Medicine and Biology Society Annual Conference, 10 pt 3, 1380-1381.
  More info

  Abstract: The complexity of histopathologic imagery presents substantial difficulties for scene segmentation preceding diagnostic information extraction. An expert-system-guided segmentation strategy founded on model-based reasoning allows a locally adaptive selection of segmentation procedures and performance control. This introduces increased stability and reliability to the process.
• Weber, J. E., & Bartels, P. H. (1988). Performance evaluation of an expert system using rescaled certainty factors. IEEE/Engineering in Medicine and Biology Society Annual Conference, 10 pt 3, 1371-1372.
  More info

  Abstract: Certainty factors are frequently used as an alternative to probabilities for representing uncertainty in expert systems. Rules proposed for combining certainty factors are much more easily applied than are the Bayesian rules for revising probabilities, since Bayesian rules require sequential assessment of conditional likelihood functions. However, an inconsistency arises in the results obtained when certainty factor rules are applied in sequential steps to update certainty factors having differing signs. A possible simple solution to this problem, involving rescaling the certainty factors, is discussed.
• Weber, J. E., Bartels, P. H., Griswold, W., Kuhn, W., Paplanus, S. H., & Graham, A. R. (1988). Colonic lesion expert system. Performance evaluation. Analytical and Quantitative Cytology and Histology, 10(2), 150-159.
  More info

  PMID: 3291892;Abstract: A computer-based expert system for diagnosing colonic sections as normal, adenoma or adenocarcinoma is described, along with an evaluation of its performance. On the basis of its knowledge base, consisting of the values of diagnostic clues and their associated certainty factors for the possible diagnoses, the system will suggest the diagnosis for new cases presented to it. Using the data provided for 16 diagnostic clues, the system arrived at correct diagnoses for all cases of normal colon, for 49 of 50 cases of adenoma and for 48 of 49 cases of adenocarcinoma. Sample outputs from the expert system are presented and discussed, and the effects of possible alterations in the data base are considered.
• Bartels, P. H., Weber, J. E., Paplanus, S. H., & Graham, A. R. (1987). Detection of diagnostic clues in statistical histometry. Analytical and Quantitative Cytology and Histology, 9(4), 355-368.
  More info

  PMID: 3311069;
• Bibbo, M., Bartels, P. H., Dytch, H. E., Puls, J. H., & Wied, G. L. (1987). Rapid cytophotometry and its application to diagnostic pathology. Applied Pathology, 5(1), 33-46.
  More info

  PMID: 3620206;
• Bibbo, M., Galera-Davidson, H., Dytch, H. E., Gonzalez, J., Lopez-Garrido, J., Bartels, P. H., & Wied, G. L. (1987). Karyometry and histometry of renal-cell carcinoma. Analytical and Quantitative Cytology and Histology, 9(2), 182-187.
  More info

  PMID: 3606777;Abstract: In an attempt to more objectively predict the outcome of renal cancers, karyometric and histometric studies were performed using an interactive computer-based system for the quantitative analysis of tissue sections. Analysis showed a significant relationship between patient survival and metastases and the histometric parameters of nuclear elongation, nuclear crowding and mitotic density, as well as tumor grade. Patients who died tended to have a high mitotic density, elongated and crowded nuclei and high-grade tumors. Ploidy showed no significant correlation with prognosis while nuclear elongation and crowding did. Differences in histologic grade were significantly associated with several histometric variables, including nuclear area, shape, crowding, elongation and mitotic density.
• Duckstein, L., Weber, J. E., & Bartels, P. H. (1987). EXPERT SYSTEM COMPLEXITY IN QUANTITATIVE HISTOPATHOLOGY.. IEEE/Engineering in Medicine and Biology Society Annual Conference, 1555-1556.
  More info

  Abstract: As expert systems find application in the quantitative diagnostic assessment of clinical materials in the histopathology, a number of systems executing increasingly complex control functions are beginning to emerge. Closed-domain production systems are being followed by systems capable of learning, as represented by automatic updating of decision rules, automated augmentation of the rule set, and automated exploration of data structure. Finally, systems that sequence decisions based not on attributes or object, but rather on procedures and sets of procedures, are under development. The authors discuss rule-based production systems and learning systems in this context.
• Dytch, H. E., Bibbo, M., Bartels, P. H., Puls, J. H., & Wied, G. L. (1987). An interactive microcomputer-based system for the quantitative analysis of stratified tissue sections. Analytical and Quantitative Cytology and Histology, 9(1), 69-78.
  More info

  PMID: 3580086;Abstract: This paper describes a microcomputer-based system that allows diagnostically relevant properties of stratified tissue sections to be objectively measured. The results of detailed nuclear image analyses are examined in the broader context of the position of nuclei within the tissue section and relative to histologic structures and each other. Quantitative measures are obtained for important morphometric and densitometric properties of individual nuclei and mitotic figures and especially for their distribution and orientation within the tissue section relative to the stratum germinativum and each other. Recorded karyometric and histometric parameters include measures of nuclear DNA content (based on optical density measurements), size, roundness, texture, shape, distance to the basal layer, angle with the stratum germinativum, epithelial height and proximate nuclear distance. Statistics generated describe normalized mitotic density as a function of depth in the epithelium, and a composite mitotic index is produced based upon weighting of these densities relative to their distance from the stratum germinativum. These properties and derived statistics may be examined as a function of epithelial depth and nuclear type and may be plotted as a function of other diagnostic features in addition to the observed stratum. The system is one part of microTICAS-STRATEX, an expert diagnostic system for the clinical evaluation of stratified tissue sections, now under development as an outgrowth of the microTICAS system. Results of measurements made with this system will be compared with site-specific and diagnosis-specific reference profiles and used in conjunction with clinical data derived from a data base at the University of Chicago of over 1.5 million patients to generate diagnostic and prognostic evaluations.
• Maenner, R., Shoemaker, R. L., & Bartels, P. H. (1987). HEIDELBERG POLYP SYSTEM.. IEEE Micro, 7(1), 5-13.
  More info

  Abstract: A description is given of the Polyp (from 'polyprocessor'), a reconfigurable multiprocessor being used as a real-time data processor in high-energy physics experiments. Design consideration are discussed, and the system architecture is described. Each Polyp module consists of a combination of five units - CPU, cache, host interface, memory, and bus switch unit - all of which are connected to an internal module bus. The bus system, the Polybus, is a two-stage interconnection network with a multiple global data bus structure, well-suited to general-purpose multiprocessors of the MIMD type. The Polybus is the common communication path used for all data transfers among modules. Hardware for scheduling and dispatching tasks is also described.
• Shack, R. V., Bartels, P. H., Buchroeder, R. A., Shoemaker, R. L., Hillman, D. W., & Vukobratovich, D. (1987). Design for a fast fluorescence laser scanning microscope. Analytical and Quantitative Cytology and Histology, 9(6), 509-520.
  More info

  PMID: 3435628;Abstract: The design of a fast fluorescence laser scanning microscope is described and illustrated, with discussion of the design consideration of the principal components, including the optical elements. The system, now under construction at the Optical Sciences Center of the University of Arizona, is expected to provide very-high-speed scanning, at a high spatial sampling density, of large object areas while retaining a flexibility of applications. The projected scanning rate approximates the rate achieved by flow cytometry; the projected rates of information generation should be orders of magnitude higher.
• Weber, J. E., Bartels, P. H., & Wied, G. L. (1987). FUZZY REASONING, POSSIBILITY THEORY, AND PROBABILITY THEORY IN EXPERT SYSTEMS FOR HISTOPATHOLOGY.. IEEE/Engineering in Medicine and Biology Society Annual Conference, 1560-1562.
  More info

  Abstract: Decisions in diagnostic cytology, whether made by human diagnosticians or expert systems, involve uncertainty. The study examines ways in which the design of a diagnostic expert system may consider uncertainty resulting from inherent randomness of a process (e. g. , the expression of a diagnostic clue) and uncertainty attributable to the vagueness in the assessment of a diagnostic clue by the diagnostician. The appropriate use of probability theory, and of the theory of possibility for modeling either kind of uncertainty, are discussed.
• Weber, J. E., Bartels, P. H., Bartels, H. G., & Bibbo, M. (1987). Discrimination of DNA ploidy patterns by order statistics. Analytical and Quantitative Cytology and Histology, 9(1), 60-68.
  More info

  PMID: 3580085;
• Wied, G. L., Weber, J. E., & Bartels, P. H. (1987). EXPERT SYSTEMS AS CLASSIFIERS IN DIAGNOSTIC CYTOPATHOLOGY.. IEEE/Engineering in Medicine and Biology Society Annual Conference, 1915-1917.
  More info

  Abstract: The use of rule-based expert systems in quantitative cytopathology that may readily implement both statistical and descriptive classification procedures is discussed. In the case of statistical classification algorithms, hierarchic decision sequences based on features with real-valued attributes are involved. Descriptive classification techniques allow highly specific classification into a large number of categories; they are readily implemented simply by using forward-chaining expert systems with strict constraints on rule selection. Automated updating may require modification of membership functions, and thus intervention by the designer.
• Wied, G. L., Weber, J. E., Dytch, H., Bibbo, M., & Bartels, P. H. (1987). TICAS-STRATEX, AN EXPERT DIAGNOSTIC SYSTEM FOR STRATIFIED CERVICAL EPITHELIUM.. IEEE/Engineering in Medicine and Biology Society Annual Conference, 1557-1559.
  More info

  Abstract: TICAS-STRATEX is an expert system under development at the University of Chicago for the diagnostic assessment of precursor lesions of cervical epithelium. Tissue architectural patterns are assessed and probabilities are derived, indicating one or another diagnostic condition. The likelihood function for the expert system in its utilization of diagnostic clues is discussed.
• Bartels, P. H., Duckstein, L., Weber, J. E., Paplanus, S., & Graham, A. (1986). SYSTEMS ENGINEERING APPROACH TO ORGANIZING A KNOWLEDGE BASE.. IEEE/Engineering in Medicine and Biology Society Annual Conference, 761-765.
  More info

  Abstract: A systems engineering tool called Q-analysis is used for assessing the structure relating diagnostic clues and diagnoses in histopathology. This analysis provides a basis for structuring a knowledge base for a diagnostic expert system. An example is presented which relates to diagnostic clues in tissue analyses for colon cancer.
• Bibbo, M., Bartels, P. H., Galera-Davidson, H., Dytch, H. E., & Wied, G. L. (1986). Markers for malignancy in the nuclear texture of histologically normal tissue from patients with thyroid tumors. Analytical and Quantitative Cytology and Histology, 8(2), 168-176.
  More info

  PMID: 2425826;Abstract: The chromatin of nuclei from histologically normal-appearing glands in tissue sections from patients with follicular adenoma, papillary carcinoma and follicular carcinoma of the thyroid exhibited changes that were statistically significant when compared to the nuclear chromatin in tissue sections of thyroid from normal individuals. Certain chromatin texture features assumed values approximating those found in tumor cell nuclei. Staining was affected only slightly, and morphometric features, such as nuclear area, roundness and ellipticity, were statistically not different from those in normal controls. In patients with Hurthle-cell tumors, such marker features could not be found. Results from measurements on 30 patients (approximately 1,000 nuclei) are reported.
• Bibbo, M., Dytch, H. E., Puls, J. H., Bartels, P. H., & Wied, G. L. (1986). Clinical applications for an inexpensive, microcomputer-based DNA-cytometry system. Acta Cytologica, 30(4), 372-378.
  More info

  PMID: 3461647;Abstract: Some of the clinical capabilities of an inexpensive, microcomputer-based DNA-cytometry system are described. A variety of applications is illustrated: (1) ploidy assessment, which provides significant information for the management of patients with tumors from almost any organ site; (2) quality control of cytologic and histologic diagnoses; (3) determination of tumor heterogeneity; (4) monitoring of cell profile changes during therapy; and (5) quantitation of immunocytochemical staining.
• Duffy, F. H., Bartels, P. H., & Neff, R. (1986). A response to Oken and Chiappa. Annals of Neurology, 19(5), 494-497.
• Dytch, H. E., Bibbo, M., Bartels, P. H., & Wied, G. L. (1986). Computer graphics in cytologic and pathologic microscopy. Tools for the clinician and researcher. Analytical and Quantitative Cytology and Histology, 8(2), 81-88.
  More info

  PMID: 2425827;Abstract: Computer graphics capabilities are becoming practical economically and ergonometrically for the first time on a widespread basis. The factors that have brought this about and the importance of such graphic interfaces to computer technology in the clinical and research laboratory are discussed. The uses of computer graphics in cytologic and pathologic microscopy are reviewed in terms of objectives, types of display and software and hardware display methods, and examples of various types of displays are presented.
• Dytch, H. E., Bibbo, M., Puls, J. H., Bartels, P. H., & Wied, G. L. (1986). Software design for an inexpensive, practical, microcomputer-based DNA cytometry system. Analytical and Quantitative Cytology and Histology, 8(1), 8-18.
  More info

  PMID: 3754144;
• Galera-Davidson, H., Bibbo, M., Bartels, P. H., Dytch, H. E., Gonzalez-Campora, R., Sanchez, F., & Wied, G. L. (1986). Differential diagnosis between follicular adenoma and follicular carcinoma of the thyroid by marker features. Analytical and Quantitative Cytology and Histology, 8(3), 195-200.
  More info

  PMID: 2430591;Abstract: Morphometric features descriptive of nuclear morphology, nuclear chromatin texture and nuclear staining density were studied in tissues from encapsulated follicular carcinoma and follicular adenoma of the thyroid, whose differentiation is one of the most difficult tasks in thyroid pathology. Measurements were made on nuclei in both tumorous and normal-appearing areas in the same section from six patients with each diagnosis, as well as on nuclei in sections from six normal individuals. The average nuclear optical density (OD) was significantly lower in the follicular carcinomas than in the follicular adenomas while the nuclear area was significantly increased in the carcinomas. The variance of the nuclear OD values was higher in both tumor and normal-appearing areas in the sections from the follicular adenomas than in the follicular carcinomas; this variance was lowest in tumor areas from the follicular carcinomas. There were also significant differences in another chromatin feature (the average difference in OD values between adjacent pixels) between nuclei from follicular carcinomas and those from follicular adenomas, whether comparing measurements between the tumor areas or between the histologically normal-appearing areas of the sections. These results suggest that the high-resolution study of marker features, such as nuclear chromatin and staining density, may help in the differentiation between follicular adenoma and encapsulated follicular carcinoma of the thyroid.
• Galera-Davidson, H., Bibbo, M., Bartels, P. H., Dytch, H. E., Puls, J. H., & Wied, G. L. (1986). Correlation between automated DNA ploidy measurements of Hurthle-cell tumors and their histopathologic and clinical features. Analytical and Quantitative Cytology and Histology, 8(2), 158-167.
  More info

  PMID: 3730089;Abstract: The treatment of Hurthle-cell tumors of the thyroid is controversial because of their rarity and the inconsistent histopathologic criteria for their diagnosis. In order to obtain more objective criteria for the management of Hurthle-cell tumors, the nuclear DNA content of cells from 20 cases was measured with the MicroTICAS system and the correlation between the DNA distribution patterns and the clinical and histopathologic findings was evaluated. Three main DNA patterns were found: euploid, polyploid and aneuploid. The euploid or polyploid Hurthle-cell tumors came from patients who did not develop distant metastases or recurrence whereas the aneuploid variants came from patients who died of their disease and/or developed distant metastases and recurrence. Various correlation analyses were performed between DNA ploidy and age, sex, size of tumor, growth pattern, pleomorphism, invasion and metastases. Our data suggest that an aneuploid DNA pattern or one with a large percentage of aneuploid nuclei with DNA content exceeding 5N may predict eventual metastases or recurrence from Hurthle-cell tumor.
• Weber, J. E., Duckstein, L., & Bartels, P. H. (1986). BAYESIAN USE OF CONTEXT TO TREAT RARE OCCURRENCES.. IEEE/Engineering in Medicine and Biology Society Annual Conference, 766-769.
  More info

  Abstract: Incorporation of rare occurrences of diagnostic clues in a histopathological knowledge base is studied as a function of context, which includes extent and direction of deviation from normality. Examples of simple sets of decision rules are given to illustrate the use of context information concerning both distances and directions of observed outliers. The differences between standard analyses of rate occurrences and analyses considering context information are noted. Also, a procedure appropriate for analyzing purely sequentially available data is contrasted with a procedure for analyzing simultaneously available data, such as the data available in computerized image analysis.
• Wied, G. L., Bartels, P. H., Bibbo, M., Dytch, H., & Weber, J. E. (1986). EXPERT SYSTEM DESIGN UNDER UNCERTAINTY OF HUMAN DIAGNOSTICIANS.. IEEE/Engineering in Medicine and Biology Society Annual Conference, 757-760.
  More info

  Abstract: Early detection of cervical cancer requires diagnostic grading of precursor lesions, frequently leading to rather diverse assessments of the same clinical materials by different experts. An expert system is proposed to provide expert resolution by using objective measures of atypia and offers direct visual comparison with a large image database.
• Bartels, P. H., & Rosenthal, D. L. (1985). Symposium by correspondence on automated microscopy. Analytical and Quantitative Cytology and Histology, 7(1), 2-3.
• Bartels, P. H., Weber, J. E., & Bibbo, M. (1985). Ploidy pattern analysis. Statistical considerations. Analytical and Quantitative Cytology and Histology, 7(2), 126-130.
  More info

  PMID: 4026944;Abstract: Availability of large data sets of ploidy measurements makes it possible to study ploidy patterns for the diagnostic and prognostic clues they can provide. Appropriate statistical analyses can improve the accuracy and precision of these studies. Such statistical analyses include considerations of sample size requirements for the detection of different types of deviations from normal, analyses of sources of variability in ploidy patterns and assessment of the probabilities of both types of possible errors in patient classification. The advantages of statistical assessment in the classification of ploidy patterns associated with diagnostic categories are discussed in the context of these considerations.
• Bibbo, M., Bartels, P. H., Dytch, H. E., & Wied, G. L. (1985). Ploidy measurements by high-resolution cytometry. Analytical and Quantitative Cytology and Histology, 7(2), 81-88.
  More info

  PMID: 4026948;Abstract: Technology has made DNA ploidy assessment a clinically and economically feasible procedure, 30 years after its clinical potential was established. This paper reviews the history of ploidy measurements, DNA cytophotometry, DNA interpretation and the prognostic value of nuclear DNA analysis in various anatomic sites. Emphasis is placed on static high-resolution cytometry, not flow cytometric methods, with a description given of the systems for rapid DNA cytometry developed as part of the TICAS Project at the University of Chicago.
• Bibbo, M., Bartels, P. H., Dytch, H. E., & Wied, G. L. (1985). Ploidy patterns in cervical dysplasia. Analytical and Quantitative Cytology and Histology, 7(3), 213-217.
  More info

  PMID: 4052223;Abstract: The ploidy patterns determined for groups of patients with cervical dysplasia (cervical intraepithelial neoplasia [CIN]) were subjected to statistical analysis. The patterns were based on the measurement of at least 100 Feulgen-stained nuclei from 30 patients with normal cervices, 10 cases of CIN I, 18 cases of CIN II and 33 cases of CIN III. The scale of the patterns was a log transformation of the ratio of the total extinction (optical density) of the nuclei to that of the 2N reference; this widens the intervals for higher ploidies, alleviating sampling requirements for intervals in which occurrences are rare and helping to maintain a reasonable sample size-to-dimensionality ratio. Pairwise discriminant analyses showed clear distinctions between the ploidy pattern for normal cases and those for CIN I, CIN II and CIN III. The distinctions between the different grades of CIN, based on these modest sample sizes, were less clearcut, largely due to pronounced patient-to-patient variability. Analysis of variance confirmed that the patient groups constitute statistically distinct entities. An aneuploid pattern did not seem to develop until CIN III lesions were involved. The diagnostic and prognostic significance of these preliminary findings require further study using larger data sets and correlations to patient survival.
• Paplanus, S. H., Graham, A. R., Layton, J. M., & Bartels, P. H. (1985). Statistical histometry in the diagnostic assessment of tissue sections. Analytical and Quantitative Cytology and Histology, 7(1), 32-37.
  More info

  PMID: 3890653;Abstract: A statistical histometric model to distinguish adenoma of the colon from normal colon is described. The model, based on the number and location of glandular nuclei, with a dependency scheme based upon displacement, was tested and shown to be significant and capable of stimulating the architecture of an adenoma.
• Weber, J. E., Baldessari, B. A., & Bartels, P. H. (1985). Test statistics for detecting aneuploidy and hyperdiploidy. Analytical and Quantitative Cytology and Histology, 7(2), 131-139.
  More info

  PMID: 4026945;Abstract: Possible approaches to the analytical evaluation of ploidy patterns are discussed and two specific problems are considered: detection of early onset of aneuploidy and detection of moderate hyperdiploidy. A statistical model for a euploid DNA pattern is formulated in terms of a mixture distribution. A test statistic for detecting deviations from this pattern is defined, and its performance is evaluated for simulated data representing differing degrees of severity of aneuploidy. An analysis based on a discriminant function using order statistics of the sample cumulative distribution functions is proposed for detecting hyperdiploidy. This procedure has the advantage of being relatively distribution-free; its performance is evaluated for simulated data and is compared with that of its classical counterparts. Although the results reported are only preliminary, they indicate that tailor-made statistical analyses can provide early detection of aneuploidy and hyperdiploidy with known and acceptable error rates using clinically reasonable sample sizes.
• Wied, G. L., Bibbo, M., Dytch, H. E., & Bartels, P. H. (1985). Computer grading of cervical intraepithelial neoplastic lesions. I. Cytologic indices. Analytical and Quantitative Cytology and Histology, 7(1), 52-60.
  More info

  PMID: 4003966;Abstract: In an attempt to derive a computer-assisted grading for cervical intraepithelial neoplastic (CIN) lesions, apparent discrepancies with the visual diagnostic evaluation could be explained from the data structure of the objective assessment. The 40 cases of dysplasia in this study appeared to fall into two subgroups having different curves of progression. A smaller subgroup, visually graded as milder, less-severe CIN, exhibited dysplastic cells with unexpectedly high abnormality indices. A larger subgroup showed dysplastic cells with abnormality incices in agreement with the tissue assessment.
• Bartels, H. G., Bartels, P. H., Bibbo, M., & Wied, G. L. (1984). Stabilized binary hierarchic classifier in cytopathologic diagnosis. Analytical and Quantitative Cytology, 6(4), 247-261.
  More info

  PMID: 6397084;Abstract: A binary tree classifier (BTC) algorithm for computer-assisted cell image analysis has been developed that overcomes the problem of overtraining due to inadequate sample size/dimensionality ratio at the higher-order nodes of a hierarchic decision structure. Provisions have been introduced that ensure that decision rules created at each node are based on samples representative of the subpopulation routed to the node. These provisions eliminate problems caused by truncation effects resulting from the application of decision rules at preceding decision nodes. The BTC performs better than do single-stage classifiers in situations where the categories' mean vectors are not well separated and no equality of covariance matrices exists. In applications in which noticeable deterioration of classifier performance on test-set data is common, the classification success rate of the BTC algorithm is not statistically significantly different between the training-set and test-set data.
• Bartels, P. H., Layton, J., & Shoemaker, R. L. (1984). Digital microscopy.. Monographs in clinical cytology, 9, 28-61.
  More info

  PMID: 6387450;
• Bibbo, M., Bartels, P. H., Dytch, H. E., & Wied, G. L. (1984). Computed cell image information.. Monographs in clinical cytology, 9, 62-100.
  More info

  PMID: 6387451;
• Jordan, S. W., Brayer, J. M., Bartels, P. H., Olson, G. B., & Anderson, R. E. (1984). Computer-assisted morphometric analysis of late renal radiation injury.. Monographs in clinical cytology, 9, 117-147.
  More info

  PMID: 6387448;
• Wied, G. L., Bibbo, M., Pishotta, F. T., & Bartels, P. H. (1984). Intermediate cell markers for malignancy. Consistency of expression. Analytical and Quantitative Cytology, 6(4), 243-246.
  More info

  PMID: 6084969;Abstract: Studies of samples from a larger patient population confirmed the consistency of expression of the previously reported markers for malignancy in normal-appearing intermediate cervical cells in samples from patients with abnormal cytology (moderate dysplasia and severe dysplasia/carcinoma in situ). Based on samples of only 30 cells per case, a false-negative rate of 10% to 30% was estimated. The expression of the marker features thus provides a clear indication of uterine abnormal cytology; the lack of expression, however, does not entirely rule out the possibility of uterine abnormalities. The use of larger sample sizes and better staining protocols could further enhance the usefulness of marker feature studies in the prescreening for cancer.
• Bartels, P. H. (1983). Numerical evaluation of cytologic data. XIII. Curve fitting and curvilinear regression. Analytical and Quantitative Cytology, 5(4), 229-235.
  More info

  PMID: 6367577;
• Bartels, P. H., Bibbo, M., Dytch, H. E., Pishotta, F. T., & Wied, G. L. (1983). Marker features for malignancy in ectocervical cells - Statistical evaluation. Cell Biophysics, 5(2), 71-77.
  More info

  PMID: 6197176;Abstract: Marker features for malignancy have recently been observed in ectocervical cells, even in cells that are visually normal in appearance. This study assessed the statistical significance of these marker features using a mixed-model nested-design analysis of variance (ANOVA). Features in blue intermediate cells from patients with normal cytology, moderate dysplasia, and severe dysplasia/carcinoma in situ, nonkeratinizing cells from patients with moderate dysplasia, severe dysplasia/carcinoma in situ, and invasive cancer, and dysplastic cells from areas of metaplasia from patients with moderate dysplasia, severe dysplasia/carcinoma in situ, and invasive cancer were tested. ANOVA clearly demonstrated that the marker features differentiate between cells of the same cell type originating from patients in different diagnostic categories. In every instance, the differences owing to the diagnostic category were statistically significantly greater than those caused by patient-to-patient variability. Although the discriminating marker features in the intermediate cells were almost exclusively spectral features reflecting staining differences, morphometric features were also marker features in the dysplastic cells. © 1983 Humana Press Inc.
• Bibbo, M., Alenghat, E., Bahr, G. F., Bartels, P. H., Dytch, H. E., Herbst, A. L., Keebler, C. M., Pishotta, F. T., & Wied, G. L. (1983). A quality-control procedure on cervical lesions for the comparison of cytology and histology. Journal of Reproductive Medicine for the Obstetrician and Gynecologist, 28(12), 811-822.
  More info

  PMID: 6663582;Abstract: A new method for the quality control of cytologic and histologic diagnoses of cervical lesions is based on the automated high-resolution scanning, image processing and computer analysis of cytometric data by the TICAS system. It determines and then compares optical-density-based ploidy patterns of cells in cytologic smears and the corresponding histologic sections, with the results available both as computergraphic displays and printouts. Examples of the former appear for an 'agreement case', in which the cytologic and histologic patterns corresponded, and a noncorrespondence (nonrepresentative) case, in which the tissue sample had been nonrepresentative of the lesion sampled by cytology. Computergraphic examples concern one case of condyloma and one of tissue repair, in both of which both the cytologic and histologic diagnoses had been overcalled. A further example shows the method's use in monitoring response to therapy. The DNA ploidy patterns on which this method is based can give diagnostic and prognostic clues when morphology alone may be equivocal or insufficient. The utility of ploidy pattern determinations of material from other body sites is also well established. With the use of microprocessors, the system described could be made inexpensive and operationally simple for the routine quality control of many cytopathologic studies as well as for the clinical follow-up of patients.
• Bibbo, M., Bartels, P. H., Dytch, H. E., Puls, J. H., Pishotta, F. T., & Wied, G. L. (1983). High-resolution color video cytophotometry. Cell Biophysics, 5(1), 61-69.
  More info

  PMID: 6190568;Abstract: Comparison was made between cytophotometric measurements obtained using two data acquisition systems, one a microphotometer and the other a rapid video camera system, to ascertain whether the degradation of data with the faster video acquisition system still results in recorded images of sufficient quality to permit computer discrimination between cells of very similar appearance. Normal-appearing intermediate cells from cases with normal cytology and those from patients with dysplasia or malignant disease, as well as the subvisual markers within these cells that have rendered them capable of cytophotometric discrimination, were used for the study. Comparison of the data recorded by the two systems indicates that the diagnostic information is preserved in the change-over to a full-field, video-rate scanning system, with differences in the data caused primarily by differences in the spectral response of the two systems. This was reflected in the substantial differences observed in the color-related features and the lesser differences seen in the textural features, while the morphometric features (outline and shape) were virtually unaffected. The differences were primarily expressed on a cell-to-cell basis; in sets of about 300 cells, which would be used in patient-to-patient comparisons, the feature values showed remarkable consistency between the two systems. © 1983 The Humana Press Inc.
• Dytch, H. E., Bartels, P. H., Bibbo, M., Pishotta, F. T., & Wied, G. L. (1983). The rejection of noncellular artifacts in papanicolaou-stained slide specimens by an automated high-resolution system. Identification of important cytometric features. Analytical and Quantitative Cytology, 5(4), 241-249.
  More info

  PMID: 6670792;
• Maenner, R., Ueberreiter, B., Bille, J., Bartels, P. H., & Shoemaker, R. L. (1983). MULTIPROCESSOR SYSTEM IN MEDICAL IMAGING.. Proceedings of SPIE - The International Society for Optical Engineering, 435, 28-36.
  More info

  Abstract: In this paper a description is given of the application of a microprocessor-based multiprocessor system (the Heidelberg POLYP polyprocessor) to chromosome aberration scoring in biologic dosimetry and to the screening of monolayer cell preparations such as those prepared for cancer prescreening. The POLYP system is characterized by the grouping of processors under software control, by multiple data buses, and by a syncbus that dynamically adjusts priority scheduling among processors, thus eliminating, for all practical purposes, the usual bus bandwidth limitations.
• Wied, G. L., Bartels, P. H., & Bibbo, M. (1983). The range of concepts of visual and automated cytodiagnosis - how closely do they touch?. Microscopica acta. Supplement, 6, 179-186.
  More info

  PMID: 6578403;
• Wied, G. L., Bartels, P. H., Dytch, H. E., & Bibbo, M. (1983). Rapid DNA evaluation in clinical diagnosis. Acta Cytologica, 27(1), 33-37.
  More info

  PMID: 6573828;Abstract: The design of an automated system for rapid DNA ploidy measurements in cytologic and histologic material is described, and examples of the resulting computergraphic displays are presented. This system offers the possibility of making DNA determinations for diagnostic and prognostic purposes a practical part of routine clinical cytology laboratory activity.
• Abmayr, W., Bartels, P. H., Giaretti, W., & Dormer, P. (1982). Identification on cell-cycle compartments by textural and chromatin features. Annals of Clinical and Laboratory Science, 12(4), NO. 1.
• Anderson, R. E., Klingler, E. L., Evan, A. P., Bartels, P. H., & Jordan, S. W. (1982). The pathogenesis of radiation injury in complex tissues. An overview of problems and probes. Analytical and Quantitative Cytology, 4(3), 165-173.
  More info

  PMID: 6816115;Abstract: The evaluation of radiation injury in complex organs has tended to lag behind comparable areas of investigative interest. This observation is somewhat surprising in view of the increased use of radiotherapy as a primary or adjuvant mode of therapy for malignant disease. In part, this problem appears to relate to the difficulty of identifying and quantifying the morphologic consequences of radiation injury in complex organs, in which the various component tissues exhibit a broad spectrum of radiosensitivities. Two approaches have been employed to address this problem: (1) utilization of sophisticated probes to evaluate the functional and morphologic sequelae of radiation injury and (2) segmentation of complex tissues into their component parts, which are then evaluated individually. Both approaches are illustrated in the papers presented in this issue. The purpose of this overview is to call attention to some of the attendant difficulties of the former approach, as seen in an ongoing investigative program concerned with radiation injury of the kidney.
• Bartels, P. H. (1982). Numerical evaluation of cytologic data. XI. Nested designs in multivariate analysis of variance. Analytical and Quantitative Cytology, 4(2), 81-89.
  More info

  PMID: 7114623;Abstract: Nested designs in multivariate analysis of variance allow the investigator to assess at what stages in an experimental procedure variability enters and to what extent. When employed in conjunction with a fixed-effects level in the design, a mixed model results, which provides the appropriate significance tests for the fixed-level effect; e.g., differences in cells between control patients and treated patients must be tested against the patient-to-patient variability, not against the cell-to-cell variability. A worked example for a three-level, mixed-model nested design is given, including the significance tests.
• Bartels, P. H. (1982). Numerical evaluation of cytologic data. XII. Curve fitting. Analytical and Quantitative Cytology, 4(4), 241-250.
  More info

  PMID: 6762123;Abstract: Graphic presentation of research data often requires the drawing of a curve. Such curves are frequently drawn by hand so that the observed data points are connected and a trend in the values of an observed variable as a function of the value of another variable becomes evident. The method of least squares allows one to draw such curves so that they fit the data in an optimized way and on the basis of an objective criterion. In this paper, the calculating scheme for fitting curves using the orthonormal polynomial method is demonstrated.
• Bartels, P. H., Bibbo, M., Dytch, H. E., Pishotta, F. T., Yamauchi, K., & Wied, G. L. (1982). Diagnostic marker displays for intermediate cells from the uterine cervix. Acta Cytologica, 26(1), 29-34.
  More info

  PMID: 6950623;Abstract: Studies have shown the existence of computer-recognizable subvisual markers in intermediate cells from different cervical conditions that show good potential for discriminating between such 'normal' cells from patients with normal, severe dysplasia/carcinoma in situ or invasive cancer cytologies. This paper presents examples of computer graphic displays, based on these markers, that could provide recognizable diagnostic clues to the cytologist, with a discussion of the factors involved in the interpretation of such displays.
• Bibbo, M., Kluskens, L., Azizi, F., Bartels, P. H., Wied, G. L., Yamauchi, K., & Herbst, A. L. (1982). Accuracy of three sampling technics for the diagnosis of endometrial cancer and hyperplasias. Journal of Reproductive Medicine for the Obstetrician and Gynecologist, 27(10), 622-626.
  More info

  PMID: 7175831;Abstract: Intrauterine aspiration (Vakutage), endocervical aspiration cytology and vaginal, ectocervical and endocervical (VCE) smears were used as outpatient procedures for diagnosis of endometrial cancer and hyperplasias in 840 patients. The results of the three technics were correlated with D&C or hysterectomy specimens. Ninety-three percent of the malignant lesions (82 adenocarcinomas, 5 mixed mesodermal tumors, 5 mixed adenosquamous carcinomas and 5 metastatic adenocarcinomas) were diagnosed by the Vakutage sample but only 67% by the VCE smears and 68% by endocervical aspiration cytology. The diagnostic accuracy of Vakutage in 50 cases of cystic hyperplasia was 88% (as compared to 14% for VCE smears and endocervical aspiration cytology) and 89% in 90 cases of adenomatous hyperplasias (as compared to 20% for the other two technics). Endometrial polyps were diagnosed exclusively from the Vakutage tissue sample in 83% of the 42 cases.
• Dytch, H. E., Bartels, P. H., Bibbo, M., Pishotta, F. T., & Wied, G. L. (1982). Computergraphics in cytodiagnosis. Analytical and Quantitative Cytology, 4(4), 263-268.
  More info

  PMID: 7165181;Abstract: Computer analysis of digitized cell images has brought major advances to the practice of cytology and to cytologic research, especially in its ability to detect subvisual morphometric feature and to quantitate the results of its investigations. Given the visual orientation of cytotechnologists and cytopathologists, a graphic display of the results of such computer analyses are presented of the color computergraphic displays developed within the TICAS system, in which color-coded masks, or frames, are generated for cell images by the choice of appropriate parameters. These masks are used to indicate the normality or degree of abnormality of the displayed cell images. Examples are also given of such masked cell images ordered according to the chosen parameters. These displays have a great potential not only in cytodiagnosis but also in the training of cytodiagnosticians, as does the generation of synthetic cell images from the data utilized in these displays; examples of synthetic cell images are also shown.
• Jordan, S. W., Olson, G. B., Bartels, P. H., Yuhas, J. M., & Anderson, R. E. (1982). Computer-assisted morphometric analysis of renal radiation response. Analytical and Quantitative Cytology, 4(3), 174-180.
  More info

  PMID: 7149483;Abstract: A single X-ray dose of 1,200 to 1,600 rads to the mouse kidney is associated with definite morphologic alteration but minimal functional impairment at six months; this progresses to profound structural and functional impairment by one year after irradiation. Subjective morphologic assessment of renal damage at six months correlates well with total radiation dose, fractionation schedule and energy characteristics of the radiation beam but does not provide adequate quantitative numerical data for sophisticated statistical tests of significance or for comparisons of effect variability at given dose levels. This investigation assessed the applicability of computer-assisted morphometric analysis (CAMA) for quantitation of effects and in making statistical comparisons of significance between kidneys subjectively classified as to degree of histologic alterations. Images of renal cortex tubula nuclei from the various histologic grades were digitized, recorded and analyzed with the CAMA system. Results indicate that the reliability of specific grade assignment by CAMA for individual nuclei was inadequate but that separation of irradiated and unirradiated renal tissue (bivariate group means) was quite distinct and of high reliability. Differences were present among the four irradiated histologic grades, but they were not marked, especially among the three highest grades. More accurate quantitation of nuclear size variations was achieved, and chromatin textural differences were detected that were not apparent to the eye. Computer-assisted morphometric analysis appears to have a valuable application in the quantification and analysis of chronic radiation effects.
• Olson, G. B., Bartels, P. H., & Anderson, R. E. (1982). Computer-assisted morphometric analysis of radiation injury in murine lymphocytes. Analytical and Quantitative Cytology, 4(3), 181-187.
  More info

  PMID: 6183999;Abstract: Groups of mice were killed 1, 3, 7 and 15 days after whole body exposure to 0, 5, 50 or 500 rads of irradiation. Lymphocytes were harvested from the peripheral blood and spleen for total and differential leukocyte counts, for determination of the absolute numbers of thymic-dependent (T) and bone marrow-dependent (B) cells and the in vitro responsiveness to phytohemagglutinin (PHA), concanavalin A (Con A) and lipopolysaccharide (LPS) mitogens and for computer-assisted morphometric analysis (CAMA) of Feulgen-stained nuclei. Alterations in functional activity and morphometric composition showed the same trends with respect to radiation injury, with maximum effects observed three days after exposure. The spleen contained a greater proportion of altered cells than did the peripheral blood. Comparison of the results with similar data obtained from lymphocytes irradiated in vitro showed that the same micromorphometric descriptors serve to distinguish irradiated from control cells in both cases.
• Shack, R., Bell, B., Hillman, D., Kingston, R., Landesman, A., Shoemaker, R., Vukobratovich, D., & Bartels, P. H. (1982). ULTRAFAST LASER SCANNER MICROSCOPE - FIRST PERFORMANCE TESTS.. Proceedings of SPIE - The International Society for Optical Engineering, 372, 49-57.
• Shoemaker, R. L., Bartels, P. H., Hillman, D. W., Jonas, J., Kessler, D., Shack, R. V., & Vukobratovich, D. (1982). ULTRAFAST LASER SCANNER MICROSCOPE FOR DIGITAL IMAGE ANALYSIS.. IEEE Transactions on Biomedical Engineering, BME-29(2), 82-91.
  More info

  Abstract: The design of an ultrafast laser scanner microscope has been completed and an experimental model has been constructed. The instrument is described and the considerations that led to the choice of scanning method and optical and electronic system design are discussed. With its data rate of 64 MHz, the instrument is the fastest such device ever designed, allowing the user to scan a microscope slide area of 2 multiplied by 2 cm with 1. 6 billion image points within 60 s. This brings the full information content of high-resolution imagery to bear on the problem of automated cytology at data rates equal to or even greater than those used in flow cytometry, and will make the automated prescreening of cervical and bladder samples feasible once the required special processors are available.
• Wied, G. L., Bartels, P. H., Dytch, H. E., Pishotta, F. T., & Bibbo, M. (1982). Rapid high-resolution cytometry. Analytical and Quantitative Cytology, 4(4), 257-262.
  More info

  PMID: 6187253;Abstract: Examples are presented of the color computergraphic capabilities of the TICAS system for the display of its high-resolution cytometric determinations. The interactive software of the system permits the simultaneous display of a selected cell and its digitized image along with the nuclear and cellular boundaries and the values of selected features, such as the nuclear and cellular areas and optical densities (ODs) and the nuclear-cytoplasmic ratio. By altering the range of OD values displayed in the digitized image, one may examine the fine structure of the cell, such as the spatial distribution of its chromatin. The relationship between cells in feature spaces of varying parameters may be depicted and manipulated in appropriate displays. The mapping capabilities of the system permit subsequent alternate restaining and remeasuring of previously identified cells; this allows, e.g., the rapid Feulgen DNA determination of ploidy and the degree of aneuploidy in cells and populations, the results of which are displayed in color-coded DNA histograms. This opens up the possibility of making this valuable diagnostic and prognostic assessment a practical clinical test. The heuristic value of this system for rapid high-resolution cytometry is also discussed.
• Wied, G. L., Bartels, P. H., Dytch, H. E., Pishotta, F. T., Yamauchi, K., & Bibbo, M. (1982). Diagnostic marker features in dysplastic cells from the uterine cervix. Acta Cytologica, 26(4), 475-483.
  More info

  PMID: 6957100;Abstract: A study of dysplastic cells, visually classified as nonkeratinizing (DYSN), keratinizing (DYKS) or severely dysplastic from an area of metaplasia (DYSM), demonstrated statistically significant differences between the same type of cell depending on the diagnostic category of the patient in which it originated: cases with moderate dysplasia (MOD), severe dysplasia/carcinoma in situ (CSD) or invasive carcinoma (INV). The results suggest that the cell types should include an indication of their origin, that, e.g., DYSN is an insufficient designation and that such cells should be denoted as DYSN-MOD, DYSN-CSD or DYSN-INV; this has major implications for the collection of cell image data bases for quantiative cytology as well as for the effort to characterize patients' cytologic samples by 'cytodiagnostic profiles.' The results also indicate that the samples as a whole contain many clues to patient diagnosis; examples are given of diagnostic profiles and the related computerographic displays for different patient categories.
• Bacus, J. W., & Bartels, P. H. (1981). Introduction to the special issue on pattern recognition of cell images - Part two. Pattern Recognition, 13(4), 277-.
• Bacus, J. W., & Bartels, P. H. (1981). Introduction to the special issue on the pattern recognition of cell images-part one. Pattern Recognition, 13(1), 1-.
• Bartels, P. H. (1981). Numerical evaluation of cytologic data. IX. Search for data structure by principal components transformation. Analytical and Quantitative Cytology, 3(3), 167-177.
  More info

  PMID: 7027855;Abstract: Principal components transformation may be used to explore the structure of a p-dimensional data set. It is difficult to detect inhomogeneities in a data set of multivariate variables by mere visual inspection of the numerical data. Plotting each variable's distribution is often either impractical, due to the number of variables involved, or might fail to reveal the presence of subpopulations due to high correlations. A practical example is given in which principal components transformation revealed the presence of subpopulations in a four-dimensional data set.
• Bartels, P. H. (1981). Numerical evaluation of cytologic data. VII. Multivariate significance tests. Analytical and Quantitative Cytology, 3(1), 1-8.
  More info

  PMID: 7015939;Abstract: Multivariate analysis is commonly used to 'prove' the existence of significant, if frequently small, differences between samples. Methods with numerical examples are presented for three test statistics used in multivariate analysis to assess the chance of an error of the first kind, alpha, that the differences observed are merely the results of chance. Hotelling's T2 test is a measure of the significance of the difference between groups. Wilk's lambda is used to assess the significance of the separation between groups. Box's M statistic is used to test the hypothesis that two variance-covariance matrices are the same. The question is raised of the power of test statistics: while they may possess high sensitivity in terms of assessing the significance of differences observed, alone they lack the specificity to determine absolutely the cause of the differences.
• Bartels, P. H. (1981). Numerical evaluation of cytologic data. VIII. Computation of the principal components. Analytical and Quantitative Cytology, 3(2), 83-90.
  More info

  PMID: 7020519;Abstract: The principal components transformation offers an effective method for dimensionality reduction and for the assessment of the mutual dependence of observed variables in a data set. An iterative procedure, the so-called power method, for finding a multivariate distribution's eigenvectors and eigenvalues is demonstrated. The projection of feature vectors onto the principal components is shown.
• Bartels, P. H. (1981). Numerical evaluation of cytologic data. X. Introduction to multivariate analysis of variance. Analytical and Quantitative Cytology, 3(4), 251-260.
  More info

  PMID: 7036810;Abstract: Quantitative measurement on cytologic material usually involves several features for each cell. The data are thus multivariate and are represented as feature vectors. Analysis of variance on univariate data is well established for the detection of small differences between sets of data. For multivariate data, the calculating schemes are much less well known. They are presented in this paper in direct analogy to univariate procedures. The calculating schedule in a factorial design of bivariate data is demonstrated.
• Bartels, P. H., & Wied, G. L. (1981). Automated image analysis in clinical pathology. American Journal of Clinical Pathology, 75(3 Suppl.), 489-493.
  More info

  PMID: 7008582;Abstract: The applications of automated cell image analyses are primarily in cytopathology, hematology, and cellular immunology. This paper presents data evaluation technics and statistical methods for automated cell analysis.
• Bartels, P. H., Abmayr, W., Bibbo, M., Burger, G., Soost, H. J., Taylor, J., & Wied, G. L. (1981). Computer recognition of ectocervical cells. Image features. Analytical and Quantitative Cytology, 3(2), 157-164.
  More info

  PMID: 7020518;Abstract: Machine recognition of ectocervical cells has been achieved with good classification success by a number of research groups. Even though a substantial number of cell image features have been introduced, only a moderate number are required for classification. For them, definitions and measures of their discriminatory potential are presented.
• Bartels, P. H., Buchroeder, R. A., Hillman, D. W., Jonas, J. A., Kessler, D., Shoemaker, R. M., Shack, R. V., Towner, D., & Vukobratovich, D. (1981). Ultrafast laser scanner microscope. Design and construction. Analytical and Quantitative Cytology, 3(1), 55-66.
  More info

  PMID: 7235391;Abstract: The design of an ultrafast laser scanner microscope has been completed, and an experimental model has been constructed. Details of the novel objective lens design, the automatic focus system, the high-speed polygon scanner and the fast clock system are given. Results from initial tolerance testing as well as the first recorded images are presented.
• Bartels, P. H., Olson, G. B., Bartels, H. G., Brooks, D. E., & Seaman, G. V. (1981). The automated analytical electrophoresis microscope. Cell Biophysics, 3(4), 371-386.
  More info

  PMID: 6175421;Abstract: The components of an automated computer-controlled analytical electrophoresis microscope (AEMS) are described. Computer tracking of migrating cells projected under phase contrast onto a vidicon permits the rapid taking of multiple velocity measurements per cell so that reliable determinations of electrophoretic mobilities thereby result. The computer-controlled cell search and tracking algorithms allow high speed operation so that statistically valid profiles and data bases can be collected rapidly. Initial electrophoretic mobility evaluations have been carried out on populations of lymphocytes, erythrocytes, and platelets. © 1981 The Humana Press Inc.
• Bibbo, M., Bartels, P. H., Sychra, J. J., & Wied, G. L. (1981). Chromatin appearance in intermediate cells from patients with uterine cancer. Acta Cytologica, 25(1), 23-28.
  More info

  PMID: 6162305;Abstract: This paper presents the differences found by the computer analysis of digitized images of visually apparently normal intermediate cells from patients with normal cytology and patients with uterine cancer. There exist differences in the absolute amount of nuclear staining, the average density of staining, cell shape and size and also differences in spectral contrasts of the nuclear chromatin. Visual cytologic detection of these changes became more conspicuous when the photomicrographs were arranged in order of their discriminant function scores.
• Duffy, F. H., Bartels, P. H., & Burchfiel, J. L. (1981). Significance probability mapping: An aid in the topographic analysis of brain electrical activity. Electroencephalography and Clinical Neurophysiology, 51(5), 455-462.
  More info

  PMID: 6165544;Abstract: We illustrate the application of significance probability mapping (SPM) to the analysis of topographic maps of spectral analyzed EEG and visual evoked potential (VEP) activity from patients with brain tumors, boys with dyslexia, and control subjects. When the VEP topographic plots of tumor patients were displayed as number of standard deviations from a reference mean, more subjects were correctly identified than by inspection of the underlying raw data. When topographic plots of EEG alpha activity obtained while listening to speech or music were compared by t statistic to plots of resting alpha activity, regions of cortex presumably activated by speech or music were delineated. Different regions were defined in dyslexic boys and controls. We propose that SPM will prove valuable in the regional localization of normal and abnormal functions in other clinical situations. © 1981.
• Olson, G. B., & Bartels, P. H. (1981). Computer discrimination of splenocytes and peripheral blood lymphocytes from mice infected with friend murine leukemia virus. Pattern Recognition, 13(1), 57-64.
  More info

  Abstract: Mice infected with Friend virus were evaluated at various times after infection to correlate the biologic descriptors of a leukemic process with cell image features obtained by computerized scanning microphotometry of cells removed from the mice. Results show that cells removed from mice 32 days after infection, when biologic descriptors were positive, have features which differentiate them from normal cells. These features were used to detect viral altered cells at various times after infection. Computerized microphotometric analysis detects viral altered cells first in the spleen, then in the peripheral blood. A significant level of viral altered cells can be detected by day 4 in the spleen and by day 8 in the peripheral blood. Clinical signs of the disease are not apparent at this time. This study demonstrates the value of computerized microphotometric analysis for the early detection of leukemic processes before other signs of the disease are evident. © 1981.
• Sherman, A., Koss, L. G., Adams, S., Schreiber, K., Moussouris, H. F., Freed, S. Z., Bartels, P. H., & Wied, G. L. (1981). Bladder cancer diagnosis by image analysis of cells in voided urine using a small computer. Analytical and Quantitative Cytology, 3(3), 239-249.
  More info

  PMID: 7294542;Abstract: The results of diagnostic evaluation of digitized cell images from the sediment of voided urine of 25 patients with various forms of bladder cancer using a small laboratory computer are presented. The study was based on assessment of visually selected consecutive mononucleated urothelial cells with reasonably well preserved cytoplasmic and nuclear features. Cell feature extraction and analysis were performed on 4,340 cell images from the 25 patients (an average of 174 cells per patient, with a range of 93 to 215) and on 446 control cell images from the sediment of 3 normal volunteers. The cell images were divided into a training set and an object set. Staining differences among samples were adjusted by computer, using the average optical density of nuclei of benign urothelial cells as a point of reference. The computer algorithms were used in a two-level hierarchic classification. The first level of analysis served to eliminate small epithelial cells, presumably of renal tubular origin, and the second level classified the remaining urothelial cells into four classes: NEG, ATY I, ATY II and POS. In all 25 patients, using object sets only, the sum of ATY II and POS cells was 29% or higher (range, 29% to 88%), with an average of 62.5%. In patients with flat carcinoma in situ, either as a sole lesion or as a peripheral lesion, the sum of ATY II and POS cells averaged 66.7%. The results were tested against theoretically constructed benign but atypical samples made up of various proportions of benign cells (NEG and ATY I) and their misclassification errors by computer. Under the least favorable circumstances, when the control sample was composed of 100% ATY I cells, the profiles of 23 patients still would have been considered abnormal and 2, ambiguous. With diminishing proportion of ATY I cells in the contrived sample, all 25 patients would be identified as abnormal. The significance of the respective diagnostic roles of the POS and ATY II cells in the samples of the 25 patients was analyzed using binomial distributions and tested against theoretical negative samples containing, respectively, 25% and 50% ATY I cells. In all but two patients, a sufficiently high level of 'alarms' would be generated by the POS cells alone to establish a presumptive diagnosis of cancer. The remaining two profiles would be considered 'suspicious' based on the proportion of ATY II cells. The data presented in this paper support the concept of automated evaluation of the cytologic samples of voided urine for the diagnosis of bladder cancer, using a hierarchic classification to establish patients' profiles based on a relatively small sample of cells. Further extensive testing of the concept is mandatory. The clinical significance of the approach used in this study is discussed.
• Wied, G. L., Bartels, P. H., Bibbo, M., & Keebler, C. M. (1981). Frequency and reliability of diagnostic cytology of the female genital tract. Acta Cytologica, 25(5), 543-549.
  More info

  PMID: 6945019;Abstract: In light of recent recommendations on the frequency of cytologic examinations for the detection of uterine carcinoma, data are presented on patient response, detection rates, sampling quality, cytotechnologists' proficiency and laboratory quality assurance.
• Wied, G. L., Bartels, P. H., Bibbo, M., Sychra, J. J., & Taylor, J. (1981). Computer discrimination of ectocervical cells. Assessment of the value of spectral information. Analytical and Quantitative Cytology, 3(3), 225-234.
  More info

  PMID: 7027857;Abstract: The recording of spectral images in two or three spectral bands brings a significant improvement in the rate of correct recognition of ectocervical cellular material. The results from discriminant analyses relying solely on spectral features and analyses where color features were excluded are compared against analyses where all features were made available.
• Wied, G. L., Bibbo, M., & Bartels, P. H. (1981). Computer analysis of microscopic images: application in cytopathology.. Pathology Annual, 16 Pt 1, 367-409.
  More info

  PMID: 7329734;
• Bartels, P. H. (1980). Numerical evaluation of cytologic data. IV. Discrimination and classification. Analytical and Quantitative Cytology, 2(1), 19-24.
  More info

  PMID: 6990845;Abstract: When observed data have to be assigned to one or another category, classification rules are needed. Linear discriminat functions provide easily computed rules; weighting the discriminant function according to the variances in the data sets helps reduce classification errors. Classification on the basis of a probability density involves nonlinear decision boundaries. Simple numerical examples for bivariate feature vectors are worked out to demonstrate these approaches to classification.
• Bartels, P. H. (1980). Numerical evaluation of cytologic data. V. Bivariate distributions and the Bayesian decision boundary. Analytical and Quantitative Cytology, 2(2), 77-83.
  More info

  PMID: 7447184;Abstract: The evaluation of cytologic data often involves the classification of observations into alternative categories (data sets). Plotting the elliptical contours of bivariate distributions provides immediate insight into the structure of data sets and their mutual relations. In this paper, the computation of tolerance ellipses and confidence ellipses for bivariate distributions is demonstrated, and the finding of a Bayesian decision boundary between two bivariate distributions is illustrated.
• Bartels, P. H. (1980). Numerical evaluation of cytologic data. VI. Multivariate distributions and matrix notation. Analytical and Quantitative Cytology, 2(3), 155-160.
  More info

  PMID: 6999953;Abstract: To represent multivariate data, matrix notation is required. Its equivalence to the arithmetic notation is shown with a bivariate example. With matrix notation one can then readily represent and evaluate multivariate data sets; however, most multivariate procedures require that one find the inverse of the variance-covariance matrix. This is shown using the Cholesky factorization with a numerical example.
• Bartels, P. H., Olson, G. B., Lockart, R., & Wied, G. L. (1980). Cytophotometric studies of cell populations. Cell Biophysics, 2(4), 339-351.
  More info

  PMID: 6163545;Abstract: Technologic advances in the recording of digitized imagery have made the study of large cell populations by image analytical methods feasible. Computed image information provides quantitative, and novel, information that allows an exact measurement of minute changes in the chromatin distribution of cell nuclei, and the detection of subpopulations of cells or changes in the functional state of the cells. The sensitivity of the detection exceeds that of human observers; the specificity of the measured changes must be the subject of basic cell biologic research. © 1980 Humana Press, Inc.
• Duffy, F. H., Denckla, M. B., Bartels, P. H., & Sandini, G. (1980). Dyslexia: Regional differences in brain electrical activity by topographic mapping. Annals of Neurology, 7(5), 412-420.
  More info

  PMID: 7396420;Abstract: Electroencephalographic (EEG) and evoked potential data were recorded during behavioral testing from 8 dyslexic and 10 normal boys aged 9 to 11 years. Topographic mapping of their brain electrical activity revealed four discrete regions of difference between the two groups involving both hemispheres, left more than right. Aberrant dyslexic physiology was not restricted to a single locus but was found in much of the cortical region ordinarily involved in reading and speech. Prominent group differences were observed in the bifrontal area in addition to the more expected left temporal and left posterior quadrant regions. Although activation tests produced more prominent group difference, dyslexics differed from normal subjects at rest as well. EEG alpha activity was increased for the dyslexics, suggesting relative cortical inactivity in that group.
• Duffy, F. H., Denckla, M. B., Bartels, P. H., Sandini, G., & Kiessling, L. S. (1980). Dyslexia: Automated diagnosis by computerized classification of brain electrical activity. Annals of Neurology, 7(5), 421-428.
  More info

  PMID: 7396421;Abstract: The authors describe a method for the diagnosis of dyslexia based upon a study of electroencephalographic and evoked potential data recorded from 13 normal and 11 dyslexic boys. Measurements were made from topographic maps of brain electrical activity recorded during resting and activated testing conditions. Using a statistically based technique, we developed rules for classification that successfully diagnosed 80 to 90% of subjects not used in the initial rule development. The nature of the most useful measurements suggests that aberrant neurophysiology in dyslexia involves both hemispheres and is present at rest as well as during complex testing. The method has promise for future diagnosis and research.
• Koss, L. G., & Bartels, P. H. (1980). Urinary cytology. Device capabilities and requirements. Analytical and Quantitative Cytology, 2(1), 59-65.
  More info

  PMID: 6990848;Abstract: The practical and theoretical considerations leading to the automated diagnostic assessment of cells in the urinary sediment by a computer-based, high-resolution system are discussed. Problems of sample preparation, sample-size requirements, coefficients of variation and computer-generated features are summarized. There are no major theoretical obstacles to achieving a workable system for the diagnosis of high-grade urothelial cancer by the proposed approach.
• Koss, L. G., Bartels, P. H., & Wied, G. L. (1980). Computer-based diagnostic analysis of cells in the urinary sediment. Journal of Urology, 123(6), 846-849.
  More info

  PMID: 7382000;Abstract: Evaluation of images of cells from the urinary sediment by a computer-based microscope has yielded important preliminary results. Computer identification of benign and malignant urothelial cells can now be accomplished with a small margin of error. Furthermore, 2 categories of atypical urothelial cells can be identified. This information has been applied successfully to the study of 12 patients with bladder cancer, including 3 with non-papillary carcinoma in situ. Computer-generated diagnostic profiles based on a relatively modest number of urothelial cells (60 to 150 per patient) proved to be of diagnostic value. It is anticipated that within a few years a practical computerized system of identification of cells in the urinary sediment can be made available for general laboratory use.
• Koss, L. G., Bartels, P. H., Sherman, A., Sychra, J. J., Schreiber, K., Moussouris, H. S., & Wied, G. L. (1980). Computer identification of degenerated urothelial cells. Analytical and Quantitative Cytology, 2(2), 107-111.
  More info

  PMID: 7447181;Abstract: Computer algorithms were developed to identify very poorly preserved, degenerated urothelial cells from the urinary sediment. Cell images were analyzed by two methods: the first one based on nucleus-finding algorithms (determination of nuclear boundaries) and the second based on the analysis of histograms of optical density. Discrimination between benign and malignant degenerated cells could be accomplished by both methods, although the misclassification of cells rose from ±10% by the first approach to ±15% by the second approach. Most important, however, algorithms were obtained and incorporated into the TICAS file for the identification of this large group of urothelial cells as a significant step in the autominated analysis of the cells in the urinary sediment.
• Koss, L. G., Bartels, P. H., Sherman, A., Sychra, J. J., Schreiber, K., Moussouris, H. S., & Wied, G. L. (1980). Computer identification of multinucleated urothelial cells. Analytical and Quantitative Cytology, 2(2), 112-116.
  More info

  PMID: 7447182;Abstract: Computer analysis of multinucleated urothelial cells from the urinary sediment is reported. The cell images were analyzed by two methods: the first based on nucleus-finding (determination of nuclear boundaries) and the second based on algorithms derived from the histograms of optical density. Both approaches resulted in good discrimination between benign and malignant cells, although the misclassification rate rose slightly from ±7% by the first approach to ±10% by the second approach. Algorithms identifying multinucleated cells are now a part of the TICAS file as an important step in automated analysis of cells in the urinary sediment.
• Koss, L. G., Sherman, A., Bartels, P. H., Sychra, J. J., & Wied, G. L. (1980). Hierarchic classification of multiple types of urothelial cells by computer. Analytical and Quantitative Cytology, 2(3), 166-174.
  More info

  PMID: 7425437;Abstract: A two-level hierarchic classification of 915 urothelial cell images from the urinary sediment was performed by computer using the TICAS programs. The purpose of the analysis was to determine whether adequate discrimination could be obtained among several classes of cells, such as degenerated (DG), multinucleated (MN) and mononucleated well-preserved (WP) ones. The first level in the classification hierarchy showed that the 3 classes of cells could be identified by computer and that the identification of the group of WP cells was particularly satisfactory. At the second level of the classification hierarchy, the analysis of WP cells, using training and object sets, documented once again that the identification of diagnostically significant subgroups could be achieved with a very small misclassification error, which was less than 1% for benign (NEG) and malignant (POS) cells. In view of the prior successful application of the classification of images of WP cells to establish patient profiles, the results of the hierarchic classification suport the concept of automated analysis of cells in the urinary sediment by computer.
• Nair, K. K., Bartels, P. H., Mahon, D. C., Olson, G. B., & Oloffs, P. C. (1980). Image analysis of hepatocyte nuclei from chlordane-treated rats. Analytical and Quantitative Cytology, 2(4), 285-289.
  More info

  PMID: 7469204;Abstract: The treatment of rats with hepatic dysfunction with chlordane at below the 'no-effect' dose produces an increase in the number of hepatocyte nuclei with doubled DNA content. More important, though, it can be shown that even cells with unaltered DNA contents undergo statistically significant changes in their chromatin distribution pattern. Thus, image analysis techniques are capable here of detecting and substantiating toxic effects not apparent by visual inspection.
• Olson, G. B., Donovan, R. M., & Bartels, P. H. (1980). Microphotometric differentation of human T and B cells tagged with monospecific immunoadsorbent beads. Analytical and Quantitative Cytology, 2(2), 144-152.
  More info

  PMID: 7004290;Abstract: A comparative and statistical study was done of the classification of peripheral blood lymphocytes (PBLs) as T cells, B cells and monocytes by various immunologic procedures and computerized microphotometric analyses. PBLs from 15 healthy male subjects were examined by immunofluorescence, E rosettes, and immunoadsorbent beads (IAB) for T cells and B cells, by phase contrast microscopy and as fixed slide preparations. Cells tagged with IAB for T cells and B cells were fixed, stained with Papanicolaou stain and analyzed. Evaluation of immunologic data shows 50% to 57% T cells, 9.7% to 24.1% B cells, 13.4% to 16.1% monocytes and about 20% unmarked cells. Analysis of T cells shows significant correlations between E and T cell-IAB rosettes, but neither rosetting procedure reveals a positive correlation with immunofluorescence-labelled T cells. Comparison of B cells show an insignificant correlation between B-cell IAB and immunofluorescence. Results showed that permanent slides of rosetted cells can be made without alteration in relative numbers of rosetted cells. Assessment of immunologically tagged cell samples by image analysis correctly classified 80% to 90% of cells as T and B cells. Evaluation of homogeneity of the T and B cell population shows the existence of four subsets of B cells and five subsets of T cells.
• Wied, G. L., Bartels, P. H., Bibbo, M., & Sychra, J. J. (1980). Cytomorphometric markers for uterine cancer in intermediate cells.. Analytical and Quantitative Cytology, 2(4), 257-263.
  More info

  PMID: 7469201;Abstract: Precise microphotometric assessment of intermediate cells from patients with normal cervical cytology and from patients with dysplasia or carcinoma in situ shows the existence of small but consistent differences. Marker features for the presence of premalignant and malignant disease can be extracted from the cell images of "normal"-appearing intermediate cells. The marker features and their diagnostic classification potential are described.
• Bartels, P. H. (1979). Numerical evaluation of cytologic data. I. Description of profiles. Analytical and Quantitative Cytology, 1(1), 20-28.
  More info

  PMID: 397789;Abstract: Results of cytologic studies are often presented as graphs, distribution curves or profiles. Further analysis of such data requires description of the profiles in numerical form. This article describes the use of moments about the mean, decomposition into Gaussian components and calculation of Fourier co-efficients for the characterization of cytologic data profiles. Fully worked numerical examples suitable for execution on a pocket calculator are given.
• Bartels, P. H. (1979). Numerical evaluation of cytologic data. III. Selection of features for discrimination. Analytical and Quantitative Cytology, 1(3), 153-159.
  More info

  PMID: 396839;Abstract: The proper selection of variables is important in assembling a profile to best describe a given group, whether of patients or cells, vis-a-vis other groups. The need often arises to determine which variables in comparable profiles best discriminate between the profiles. Three techniques for the evaluation and selection of variables on the basis of their potentiality for discrimination are discussed in this article. The Kruskal Wallis test is useful in determining if a certain feature (variable) has any statistical significance between groups. The ambiguity function after Genchi and Mori and the measure of detectability (d') are discussed as direct measurements of a feature's ability to discriminate between groups. Fully worked numerical examples suitable for execution on a pocket calculator are given.
• Bartels, P. H. (1979). Numerical evaluation of cytologic data: II. Comparison of profiles. Analytical and Quantitative Cytology, 1(2), 77-83.
  More info

  PMID: 396837;Abstract: Results of cytologic studies are often presented as profiles of some type. The numerical characterization of these profiles may be statistically compared using methods of multivariate analysis. This article describes the determination of the statistically significant differences between portions of profiles, the computation of the Mahalanobis distance measure between profiles and the use of the chi-square statistic as a measure of the typicality of a profile. Fully worked numerical examples suitable for execution on a pocket calculator are given.
• Bartels, P. H., Bibbo, M., & Wied, G. L. (1979). Estimation of proportion of the patients with a very low number of tumor cells from carcinoma in situ in the cervical smear. Analytical and Quantitative Cytology, 1(2), 136-142.
  More info

  PMID: 543575;Abstract: The proportion of patients with cervical carcinoma in situ exhibiting a very low number of tumor cells in cytologic samples from the female reproductive tract was assessed with a model based on a multivariate beta distribution. The results of the study indicate that it is highly unlikely that a case of carcinoma in situ will be missed when 50,000 cells are screened in a given sample. However, the study also shows the importance of sampling methodology in the design of screening systems for cervical cancer and the crucial role that the coefficient of variation between patients assumes.
• Bibbo, M., Reale, F. R., Reale, J. C., Azizi, F., Bartels, P. H., Wied, G. L., Hajj, S. N., & Herbst, A. L. (1979). Assessment of three sampling technics to detect endometrial cancer and its precursors. A preliminary report. Acta Cytologica, 23(5), 353-359.
  More info

  PMID: 294068;Abstract: The purpose of this study was to assess the accuracy of three technics in the detection of endometrial cancer and its precursors after a preparatory stage in which the clinicians involved had the opportunity to master the technics: endometrial aspiration (Vakutage), endocervical aspiration, and vaginal, ectocervical and endocervical smears (VCE). The results of the three technics were subsequently correlated with available D&C or hysterectomy specimens. The detection rate, combining cell and tissue diagnosis, for 33 adenocarcinomas and one carcinosarcoma by Vakutage was 100% while the VCE and endocervical aspirate were diagnostic in 67% of cases each. In the five cases of atypical hyperplasia the detection rate was 100% by the Vakutage and 20% by the VCE smears and endocervical aspirate. 86% of the adenomatous hyperplasias (30 cases) were detected by the Vakutage technic but only 6% by the VCE and endocervical aspirate. 91% of the cystic hyperplasias (13 cases) were detected by the Vakutage, but none were picked up by the VCE or endocervical aspirate. Endometrial polyps were detected exclusively by the tissue sample of the Vakutage in 80% of the cases (15 cases). In the preparatory stage the diagnostic accuracy of the three sampling technics was lower, as expected. The combination of either VCE smears or endocervical aspirate with the Vakutage sample did not increase the detection of endometrial lesions significantly. Therefore, the Vakutage technic seems to be more reliable than the VCE smears and endocervical aspirate in the detection of endometrial adenocarcinoma and its percursors of endometrial adenocarcinoma and its precursors and is considered to be a useful outpatient tool.
• Duffy, F. H., Burchfiel, J. L., Bartels, P. H., Gaon, M., & Sim, V. M. (1979). Long-term effects of an organophosphate upon the human electroencephalogram. Toxicology and Applied Pharmacology, 47(1), 161-176.
  More info

  PMID: 425115;Abstract: The brain electrical activity of workers exposed to the organophosphate compound (OP), sarin, was compared to that of control subjects. Exposed workers had a history of one or more documented accidental exposures to toxic levels of sarin. However, no exposed subject had exposure within 1 year of his examination. The comparison included standard clinical electroencephalograms (EEGs), computer-derived EEG spectral analysis, and standard overnight sleep EEGs. It was not possible to diagnose subjects individually by expert visual inspection of their EEGs. However, statistically significant between-group differences for both the visually inspected and computer-derived data were reported by both univariate and multivariate statistical methods. Different EEG changes revealed by visual inspection and computer-derived spectral analysis appear to reflect the differing sensitivites of these two analytic techniques. Statistically significant group differences included increased beta activity, increased delta and theta slowing, decreased alpha activity, and increased amounts of rapid eye movement sleep in the exposed population. It is suggested that the above findings represent an unexpected persistence of known short-term OP actions. It is also suggested that these results, when taken along with the reported long-term behavioral effects of OP exposure, provide parallel evidence that OP exposure can produce long-term changes in brain function. © 1979.
• Olson, G. B., Anderson, R. E., & Bartels, P. H. (1979). Computer analysis of defined populations of lymphocytes irradiated in vitro. III. Evaluation of human T and B cells of peripheral blood origin. Human Pathology, 10(2), 179-190.
  More info

  PMID: 311330;Abstract: The radiosensitivities of lymphocytes of peripheral blood origin obtained from two healthy 45 year old male donors were studied simultaneously at yearly intervals over a three year period. Hypaque-Ficoll purified cells were exposed in vitro to 0, 5, 50, and 500 rads and then evaluated serially for viability of T and B cells, responsiveness to PHA and Con A, and morphologic evidence of injury as documented by standard light microscopy and computer assisted morphometric analysis. The results showed that T cells in both subjects were less radiosensitive than B cells. Differences between the two subjects also existed in the radiosensitivity of these two subpopulations of lymphocytes, differences that remained constant over the three year period of observation. The differences correlated with similar discrepancies in mitogenic responsiveness and are thought to relate to variations in the relative proportions of subpopulations of T and B cells. In the mouse, T and B cell subpopulations differ in radiosensitivity. The data reported herein are consistent with a similar situation in man. © 1979 W. B. Saunders Company.
• Olson, G. B., Bartels, P. H., & Anderson, R. E. (1979). Characterization of murine T and B cells by computerized microphotometric analysis. Cell Biochemistry and Biophysics, 1(3), 229-242.
  More info

  PMID: 95180;Abstract: Splenocytes and column-separated T cells are differentiated into subpopulations of T and B cells on the basis of computer-assisted morphometric analysis of Feulgen-positive nuclear DNA. Differentiation is based upon the analysis of computable image information related to DNA distribution patterns. The technique at the present time does not allow immunofluorescent and morphometric measurements to be made on a given cell. However, the differentiation obtained by using descriptors proven capable of detecting pure populations of T and B cells shows excellent agreement with the differentiation obtained by immunofluorescence analysis. The descriptors and decision rules used in the discrimination among splenocytes are reproducible from one experiment to another and remain valid for the differentiation of lymphocytes from animals of different sex and strain. © 1979 The Humana Press Inc.
• Olson, G. B., Bartels, P. H., & Anderson, R. E. (1979). Subclassification of murine T cells by computerized microphotometric analysis. Cell Biochemistry and Biophysics, 1(3), 243-254.
  More info

  PMID: 95181;Abstract: Splenocytes separated by physical means and classified as T cells by immunologic tests and computerized microphotometric analysis are differentiated into subgroups by analysis of the distribution patterns of Feulgen-positive nuclear DNA. In like fashion T cells obtained as purified preparations after separation on a nylon column, and accepted as T cells by micromorphometric analysis were subjected to further computerized morphometric analysis of nuclear DNA to form subgroups of cells. In each case, the number and composition of the detected subgroups were consistent. The classification does not appear to reflect any obvious phases of the cell cycle and is not dependent upon the sex and strain of mice from which the cells were obtained. © 1979 The Humana Press Inc.
• Reale, F. R., Bartels, P. H., Bibbo, M., Chen, M., Schreiber, H., Sychra, J. J., & Wied, G. L. (1979). Differentiation of TICAS analysis of cell populations of tracheal aspirates from hamsters with squamous-cell carcinoma. American Journal of Clinical Pathology, 72(1), 52-58.
  More info

  PMID: 453111;Abstract: Individual cells from the tracheal aspirates of hamsters exposed to benzo-a-pyrene were scanned at .5μm in three colors. Features relating to size, shape, and color were extracted and calculated by computer. The single cells were then classified by these features into separate populations with varying degrees of atypia, extending up to frank cancer cells. A high degree of accuracy was attained in classification by these methods.
• Shack, R., Baker, R., Buchroeder, R., Hillman, D., Shoemaker, R., & Bartels, P. H. (1979). ULTRAFAST LASER SCANNER MICROSCOPE.. Array, 153-159.
  More info

  Abstract: Advances in monolayer deposition of cervical cells have removed one of the last serious obstacles to the design of high-resolution automated diagnostic assessment systems. In this article are described the design considerations for a system that is capable of acquiring, within 60 sec, a 0. 5 mu m digitized image of a 4 cm**2 area on a standard glass slide. The most feasible approach is found to be a system using a rotating polygon to sweep the focused spot from a laser across a 2-mm scan line while the slide is uniformly translated perpendicular to the scan direction. The use of laser sources (a helium-neon laser at 632 nm and a krypton ion laser at 568 and/or 476 nm) as compared to the incoherent light sources used in conventional microscope systems alleviates many of the optical design problems and provides the proper wavelengths needed for recognition of Papanicolaou stained cells. It is also found that focus control of the scanning spot should be achievable using a technique involving a holographic grating. Other relevant considerations such as sample heating problems, multiphoton absorption by the sample, detector signal-to-noise ratios, laser amplitude noise control, and the digitization and buffering of the data stream are also discussed.
• Shack, R., Baker, R., Buchroeder, R., Hillman, D., Shoemaker, R., & Bartels, P. H. (1979). Ultrafast laser scanner microscope. Journal of Histochemistry and Cytochemistry, 27(1), 153-159.
  More info

  PMID: 374570;
• Wied, G. L., Bartels, P. H., Bibbo, M., Chen, M., Reale, F. R., Schreiber, H., & Sychra, J. J. (1979). Discriminant analysis on cells from developing squamous cancer of the respiratory tract. Cell Biophysics, 1(1), 39-54.
  More info

  PMID: 95167;Abstract: Cytologic preparations made from the tracheobronchial tree taken by the Schreiber catheter have been scanned by three color microphotometry. The digitized cell images were processed by the analytical cytodiagnostic programs of the TICAS system. Cells were sorted into two control groups and five groups of increasing atypia ranging from normal epithelium to invasive squamous cell carcinoma. Standard statistical tests, including Wilk's Lambda, Rao's V, and the Kruskal-Wallis tests are performed on these subsets of cell image features. This study demonstrates that discriminant analyses permit differentiation between normal cells and those from marked atypia or carcinoma and that the classification achieves a high degree of agreement with visual assignment. © 1979 The Humana Press Inc.
• Bartels, P. H., Bibbo, M., Richards, D. L., Sychra, J. J., & Wied, G. L. (1978). Patient classification based on cytologic sample profiles. I. Basic measures for profile construction. Acta Cytologica, 22(4), 253-260.
  More info

  PMID: 281828;Abstract: Past efforts in the field of automated cell recognition have focused upon the separation and classification of cell types. From these efforts, large data banks have been built and work in the field is now shifting towards the practical application of this information for clinical diagnoses. This paper presents the initial results of work on a system developed to undertake the reduction of the masses of data into diagnostically useful patient cytologic sample profiles.
• Bartels, P. H., Chen, Y. P., Durie, B. G., Olson, G. B., Vaught, L., & Salmon, S. E. (1978). Discrimination between human T and B lymphocytes by computer analysis of digitized data from scanning microphotometry. II. Discrimination and automated classification. Acta Cytologica, 22(6), 530-537.
  More info

  PMID: 367028;Abstract: Human B and T lymphocytes were purified and stained according to the Feulgen procedure. Stained preparations were scanned on a microphotometer and the digitized images were processed by analytical programs on a computer. Features of the cells relating to the chromatin distribution were extracted. Computer algorithms automatically derive classification rules and a machine recognition of B cells and T cells results in better than 80% correct assignments. The image data reveal a rich feature structure within both the purified B and T cell populations. Each population appears to contain several subpopulations of cells with distinctive and different chromatin texture.
• Bartels, P. H., Koss, L. G., Sychra, J. J., & Wied, G. L. (1978). Indices of cell atypia in urinary tract cytology. Acta Cytologica, 22(5), 387-391.
  More info

  PMID: 364908;Abstract: A sample profile based on assessment of digitized cell images of normal (NEG), malignant (POS) and two classes of atypical (ATY I and ATY II) urothelial cells was established. Data distributed in a multidimensional feature space were projected into a two-dimensional display space using linear discriminant functions as composite features. Fitting of a polynomial led to the derivation of an atypicality index for each cell and to statistically clearly significant differences in the atypicality values for the four groups of urothelial cells. The application of these findings to patients' profiles will be examined in a subsequent communication.
• Bibbo, M., Bartels, P. H., Chen, M., Harris, M. J., Truttmann, B., & Wied, G. L. (1978). The numerical composition of cellular samples from the female reproductive tract. V. Cell cluster patterns in cases of invasive squamous carcinoma of uterine cervix. Acta Cytologica, 22(4), 250-252.
  More info

  PMID: 281827;Abstract: One hundred fifty thousand cells in vaginal, extocervical and endocervical smears from 25 patients with poorly to moderately differentiated squamous cancer of the uterine cervix were evaluated to determine the numerical composition of cellular clusters, the sizes of the clusters and the cellular types contained. The study is a baseline assessment for the design of automated cytology devices for which monocellular layers or individual cells are of importance. The evaluation indicates that the majority of cells appear in actual or facultative clusters, and that dispersement or breaking up of these clusters can yield improved machine-readable samples.
• Koss, L. G., Bartels, P. H., Sychra, J. J., & Wied, G. L. (1978). Computer discriminant analysis of atypical urothelial cells. Acta Cytologica, 22(5), 382-386.
  More info

  PMID: 364907;Abstract: Prior computer studies of digitized cell images by the TICAS system have shown that the category of urothelial cells classified visually as atypical may be composed of 2 subgroups, one clustering mainly with benign cells and the other with malignant cells. As a consequence, a visual review of the group of atypical cells was conducted and tested by computer discriminant analysis. The computer classification confirmed the visual reclassification and subdivison of atypical urorethelial cells into 2 subgroups, ATY I and ATY II. This is yet another example of feedback from computer diagnosis to visual assessment of cells. The significance of these observations in terms of diagnosis will be the subject of subsequent communications.
• Koss, L. G., Bartels, P. H., Sychra, J. J., & Wied, G. L. (1978). Diagnostic cytologic sample profiles in patients with bladder cancer using TICAS system. Acta Cytologica, 22(5), 392-397.
  More info

  PMID: 364909;Abstract: Computer analysis of digitized cell images was applied to consecutively encountered, well preserved urothelial cells in the urinary sediment of 12 patients with bladder cancer. It was shown that the composition of the cell sample, and notably the proportion of cells classified in the ATY II and POS (malignant) categories, were of diagnostic significance. This work indicated that a computer-generated diagnosis based on the cells in a urinary sample could probably be achieved with a relatively small number of urothelial cells. The study also suggested that computer-generated cytologic profiles of patients with non-papillary carcinoma in situ can be distinguished from other forms of urothelial cancer. The atypicality indices computed for all cells from each patient provided important information but were per se insufficient for diagnostic purposes. This preliminary study, based on a small group of patients, suggests that high resolution scanning offers a promising approach to automation of cytology of the urinary sediment.
• Sychra, J. J., Bartels, P. H., Bibbo, M., Taylor, J., & Wied, G. L. (1978). Computer recognition of binucleation with overlapping in epithelial cells. Acta Cytologica, 22(1), 22-28.
  More info

  PMID: 349990;Abstract: A technique for computer recognition of binucleation with overlapping in epithelial cells was derived. The computer program uses analytic features related to nuclear contour only. The feature set includes moments, Fourier transform features and linguistic features. The algorithm works satisfactorily in instances where the human eye would detect the presence of two nuclei at first glance.
• Taylor, J., Bartels, P. H., Bibbo, M., & Wied, G. L. (1978). Automated hierarchic decision structures for multiple category cell classification by TICAS. Acta Cytologica, 22(4), 261-267.
  More info

  PMID: 364905;Abstract: TICAS-PROBE, a computerized system, has been developed to generate hierarchic decision structures with little human intervention. It has been tested with several tasks, and the results are comparable, if not better, than those produced by human designers. Additional methods of generating the individual elements will be included in the near future.
• Taylor, J., Puls, J., Sychra, J. J., Bartels, P. H., Bibbo, M., & Wied, G. L. (1978). A system for scanning biological cells in three colors. Acta Cytologica, 22(1), 29-35.
  More info

  PMID: 349991;Abstract: A system for scanning biological cells under high magnification has been developed which utilizes a 3-color photometer. The spectral information has proved useful in improving computer recognition of certain cell types.
• Anderson, R. E., Olson, G. B., Autry, J. R., Howarth, J. L., Troup, G. M., & Bartels, P. H. (1977). Radiosensitivity of T and B lymphocytes. IV. Effect of whole body irradiation upon various lymphoid tissues and numbers of recirculating lymphocytes. Journal of Immunology, 118(4), 1191-1200.
  More info

  PMID: 300402;Abstract: Groups of 10-wk-old female CBA/J mice were exposed in whole body fashion to 0, 5, 50, and 500 rads and sacrificed in serial fashion 1, 3, 5, 7, 9, 15 and 30 days after irradiation for morphologic evaluation of thymus, spleen, lymph node and Peyer's patch, and assessment of the relative numbers of thymus-derived (T) and bone marrow-derived (B) cells in these tissues. The absolute and relative numbers of recirculating T and B cells mobilizable by thoracic duct cannulation were also determined and compared with similar determinations with respect to peripheral blood lymphocytes. B cell depletion occurred more quickly and was more pronounced in the spleen and lymph node than T cell depletion at all 3 exposure doses. Depletion of T and B cells was roughly equal in peripheral blood and thoracic duct lymph. When present, regeneration of the T cell component occurred more rapidly than did B cell restoration. The latter often was incomplete at the time of the final sacrifice (day 30). PHA-responsive and Con A-responsive cells also appeared to differ with respect to the kinetics of cell death after whole body irradiation.
• Bartels, P. H., & Wied, G. L. (1977). COMPUTER ANALYSIS AND BIOMEDICAL INTERPRETATION OF MICROSCOPIC IMAGES: CURRENT PROBLEMS AND FUTURE DIRECTIONS.. Proceedings of the IEEE, 65(2), 252-261.
  More info

  Abstract: Automated analysis and recognition of microscopic images of cells and tissue sections is undergoing active development. With regard to cells, the problems posed by samples of white and red blood cells presented in a monocellular layer are well under control, and automated laboratory devices are in the clinical testing stage. The analysis of white blood cell populations offers good promise in immunologic, radiation biologic, and experimental chemotherapeutic research. The situation encountered in prescreening for cervical cancer in Papanicolaou-stained gynecologic samples calls for significantly more stringent cell recognition and sample size requirements. Methods based on high-resolution digitized imagery appear currently to be the only approach with the required discriminatory power. As applied to tissue sections, the complexity of the scene-segmentation, analysis, and classification problems has generally restricted research work to the area of feature extraction and to the determination of descriptive stereologic parameters. The field of microscopic image analysis demands competence in a wide range of disciplines, from the practical clinical situation to optics, electrooptical devices, digital logic design, computer architecture, software engineering, multivariate statistics, and decision theory.
• Bibbo, M., Bartels, P. H., Chen, M., Harris, M. J., Truttman, B. J., & Wied, G. L. (1977). The numerical composition of cellular samples from the female reproductive tract. IV. Carcinoma in situ cases exhibiting other than normal vaginal flora. Acta Cytologica, 21(5), 705-709.
  More info

  PMID: 272114;
• Koss, L. G., Bartels, P. H., Bibbo, M., Freed, S. Z., Sychra, J. J., Taylor, J., & Wied, G. L. (1977). Computer analysis of atypical urothelial cells. I. Classification by supervised learning algorithms. Acta Cytologica, 21(2), 247-260.
  More info

  PMID: 324216;Abstract: Computer discrimination of atypical (ATY) urothelial cells from the urinary sediment by means of supervised learning algorithms disclosed that these cells form a distinct, although ill defined, family of cells which differs from normal (NEG) and malignant (POS) cell groups. The clinical significance of this observation must await longterm clinical followup. The possibility of issuing computer displays on individual patients with possible diagnostic and prognostic implications is discussed.
• Koss, L. G., Bartels, P. H., Sychra, J. J., & Wied, G. L. (1977). Computer analysis of atypical urothelial cells. II. Classification by unsupervised learning algorithms. Acta Cytologica, 21(2), 261-265.
  More info

  PMID: 324217;Abstract: Computer evaluation of digitized cell images of a set of cells from the urinary tract has shown 'atypical cells' (ATY cells) to be a distinct cytologic category. They occupy in variously defined feature spaces, regions between the benign and tumor cells of urothelial origin. There is moderate overlap with both the region into which the benign cells, and the region into which the tumor cells are mapped. The results of classification of these cell images by supervised learning, however, have not been conclusive with respect to the manner of distribution of the atypical cell images themselves. Two-dimensional displays employing composite features suggested the possibility that ATY cells may fall into 2 subsets, one mode representing ATY cells more closely resembling benign urothelial cells, and the other more closely resembling tumor cells. There are obvious prognostic implications. Unsupervised learning algorithms are of value when one searches for objective and statistically secured evidence of structure, e.g. biomodality, in a multivariate data set. It was, therefore, decided to process the data obtained from the normal, tumor and ATY cell images by an unsupervised learning algorithm in an attempt to search for the possibility of extracting prognostic information. It was found that this group of cells does form a separate and distinct, though ill-defined, group of urothelial cells. The majority of ATY cells are classed as such by the PINDEX program (McClellan, R.P., Thesis, Dept. EE, University of Arizona, 1971) although a minority falls into the POS (tumor cell images) and NEG (benign cell) groupings. By the same token, a small proportion of POS and NEG cells is classed as ATY by the unsupervised learning algorithm. It must also be pointed out that the ATY cells as shown in 3 figures provide classification on individual patient cell population and thus lend themselves to computerized displays of possible clinical diagnostic and prognostic value. This and prior communications on the subject of computer classification of urothelial cells are based on primary classification of urothelial cells by visual microscopic criteria. The results are sufficiently encouraging to attempt to classify, by computer, large populations of urothelial cells which can not be preselected by microscopic criteria.oxidoreductase may be a controlling enzyme in iron release from ferritin.
• Sychra, J. J., Bartels, P. H., Bibbo, M., & Wied, G. L. (1977). Dimensionality reducing displays in cell image analysis.. Acta Cytologica, 21(6), 747-752.
  More info

  PMID: 349987;Abstract: The paper discusses the problem of representation of multivariate cell data sets in two dimensions such that the essence of the situation as represented by the multivariate feature space is preserved. A corresponding projectional technique has been developed and illustrated on a set of five different cell types of ectocervical cells, including normal and carcinoma cells.
• Sychra, J. J., Bartels, P. H., Bibbo, M., Taylor, J., & Wied, G. L. (1977). Computer recognition of abnormal ectocervical cells. Comparison of the efficacy of contour and textural features.. Acta Cytologica, 21(6), 765-769.
  More info

  PMID: 349989;Abstract: The efficacy of contour and texture features for the computer recognition of abnormal ectocervical cells is compared. For this comparison, three cell groups were formed: benign, suspicious and malignant. Nine decision rules were derived. The discrimination attained by features derived from the contours of nucleus and cytoplasm is comparable to that achieved by textural features. A combination of both kinds of features gives best results.
• Wied, G. L., Bartels, P. H., Bibbo, M., Puls, J., Taylor, J., & Sychra, J. J. (1977). Computer recognition of ectocervical cells. Classification accuracy and spatial resolution.. Acta Cytologica, 21(6), 753-764.
  More info

  PMID: 349988;Abstract: The effect of a relaxation of the spatial resolution of images of ectocervical cells on classification accuracy by computer has been studied. The discriminatory capabilities of optical density, texture and shape features have been evaluated. Comparable computer classifications have been derived. The differences in the discriminating power of .5 and 1 mu resolution are found to be smaller than the differences corresponding to 1 and 2 mu resolution. While 2 mu resolution may be marginally acceptable in certain situations, the 4 mu resolution did not render acceptably good classification rates.
• Bartels, P. H., Bibbo, M., & Wied, G. L. (1976). Modeling of histologic images by computer. Acta Cytologica, 20(1), 62-67.
  More info

  PMID: 773071;Abstract: It should be understood that a computer assessment may be optimized for the detection of change. As in any measurement problem, it is here that the question of the 'resolution of the methodology' arises. The question is, how small or subtle a change can be detected in the placement pattern of a tissue's nuclei, or even only in the mutual dependencies of their placement. 'Resolution' would imply that the change must pass a multivariate significance test. In addition, the potential capabilities of what might properly be called 'computer microscopy' extend far beyond the extraction and recording of sets of image descriptors. The researcher has at his disposal most powerful analytical procedures, multivariate statistics, complex multivariate trend analytical methods, an enormous data base, and the adaptive, self learning capabilities of advanced software systems. This substantial analytical power enables him to search for hidden relations, to detect very subtle changes, and to establish their significance. The applications are in the area of long term toxicity effects, environmental influences, or aging. The research reported here describes the concepts underlying the development of a program which simulates images of tissue sections. Experience has shown that for tissues with marked heterogeneity, i.e., different tissue components, better discrimination would result if an assessment were made independently on each component. The problem of driving a quantitative description of an entire tissue section is closely related to the situation encountered in 'scene analysis.' Within each subscene which would correspond here to a tissue component one may then assess the distribution pattern of its components. Here this would be the nuclei. They could be described as number per unit area, placement pattern, orientation, and the variances associated with these tissue descriptors. There is no doubt that the present modeling program can at this point handle only very simple tissue patterns. The diversity of tissue patterns presents a great challenge to formulate and define in machine compatible form the essentials of the two dimensional information. The general problem, in fact, is the necessity of converting two dimensional information to a one dimensional description while preserving all of the essential contextual information in the picture.
• Bibbo, M., Bartels, P. H., Chen, M., Harris, M. J., Truttmann, B., & Wied, G. L. (1976). The numerical composition of cellular samples from the female reproductive tract. II. Cases with invasive squamous carcinoma of the uterine cervix. Acta Cytologica, 20(3), 249-254.
  More info

  PMID: 1064276;Abstract: In a previous publication the numerical composition was presented of cellular samples from patients with carcinoma in situ of the uterine cervix. This paper deals with the composition of cellular samples from patients with invasive carcinoma of the uterine cervix. Previous analytic assessments of the cellular changes observed in the presence of squamous cell cancer of the uterine cervix were concerned primarily with the study of abnormal cells present in the cellular sample. The present study also takes into consideration the relation of malignant cells to nonmalignant cells, the number of well preserved versus degenerated cells, the relative amount of cell degeneration in relation to the site of smear preparation and the relative amount of nonepithelial elements. Also analyzed are the number of isolated cells versus cells in clusters. In the 25 cases evaluated an average of 45 isolated well preserved carcinoma cells were observed in the vaginal smears, 52 in the ectocervical smears and 148 in the endocervical smears in a total of 2,000 epithelial cells in each of the V, C or E smears. Isolated well preserved non keratinizing dysplastic cells, dysplastic metaplastic cells and keratinizing dysplastic cells accompanying tumor cells were observed in larger number in the endocervical smear (6, 3 and 1 respectively) followed by the vaginal smear (1, 1 and 3) and by the cervical smear (1, 1 and 1). Therefore, the average number of abnormal well preserved cells lying singly (i.e. carcinoma cells plus dysplastic cells) was 50 in the vaginal smear, 55 in the ectocervical smear and 158 in the endocervical smear. Of 10,000 cells surveyed in preparations from 100 women with proven cancer of the uterine cervix, 50.9% were isolated, 7.7% were arranged in sheets and 41.4% were arranged in syncytia. In this study a lower percentage of isolated cells was found than in similar studies, which can be explained if one takes into consideration what was called a 'cluster' for the purpose of this study, namely, not only cells in sheets or syncytia, but also actually isolated cells which were overlapping. The figures presented here are thus a lower bound for the number of isolated, diagnostically interesting cells which could be assessed by an automated device.
• Bibbo, M., Bartels, P. H., Chen, M., Harris, M. J., Truttmann, B., & Wied, G. L. (1976). The numerical composition of cellular samples from the female reproductive tract. III. Cases with mild and moderate dysplasia of uterine cervix. Acta Cytologica, 20(6), 565-572.
  More info

  PMID: 1069453;
• Sherwood, E. M., Bartels, P. H., & Wied, G. L. (1976). Feature selection in cell image analysis: use of the ROC curve. Acta Cytologica, 20(3), 255-261.
  More info

  PMID: 775870;
• Sychra, J. J., Bartels, P. H., Taylor, J., Bibbo, M., & Wied, G. L. (1976). Cytoplasmic and nuclear shape analysis for computerized cell recognition. Acta Cytologica, 20(1), 68-78.
  More info

  PMID: 773072;
• Anderson, R. E., Olson, G. B., Howarth, J. L., Wied, G. L., & Bartels, P. H. (1975). Computer analysis of defined populations of lymphocytes irradiated in vitro. II. Analysis of thymus dependent versus bone marrow dependent cells. American Journal of Pathology, 80(1), 21-31.
  More info

  PMID: 1080360;PMCID: PMC1912831;Abstract: Three uniform populations of T and B cells exposed to varying amounts of X irradiation are examined utilizing computer assisted morphometric analysis. These populations are: thoracic duct lymphocytes (TDL) from congenitally athymic (nude) mice (B cells); TDL from CBA mice treated with anti Ig plus complement (T cells); and computer selected untreated T cells from CBA TDL. Irradiated B cells show a more even dispersion of the nuclear chromatin and a dose dependent increase in relative nuclear area beginning with the lowest dose evaluated (50 rads); no significant change in total optical density (OD) is demonstrable over the dose range evaluated (0 to 2000 rads). Anti Ig treated irradiated T cells demonstrate an initial shift toward lower OD values as a function of dose followed by a marked rise of OD values at 2000 rads, where numerous densely staining Feulgen positive aggregates are identified. The relative nuclear area of this cell population also shows a biphasic response to radiation injury with an initial increase at the lower dose levels followed by a progressive decline to approximate control levels at 2000 rads. This effect is mirrored by the alteration in total OD which, after a decrease at low dose levels, approximates control values at 2000 rads. The computer selected T cells show little change in OD values at the low dose levels but show a marked increase in the more densely staining Feulgen positive material following 2000 rads. This population reveals no apparent change in either relative nuclear area or total OD as a function of dose. Thus, untreated computer selected T cells exhibit remarkably little evidence, morphologically, or radiation injury at doses associated with pronounced alterations on the part of B cells. In addition, treatment of a mixed cell population (CBA TDL) with anti Ig plus complement to remove the B cells appears to alter the response of the residual T cells to radiation injury. These results, in conjunction with recent evidence to support the concept that T cells possess surface Ig, suggest that an Ig anti Ig interaction may alter the radiosensitivity of T cells.
• Anderson, R. E., Olson, G. B., Shonk, C., Howarth, J. L., Wied, C. L., & Bartels, P. H. (1975). Computer analysis of defined populations of lymphocytes irradiated in vitro: I. Evaluation of murine thoracic duct lymphocytes. Acta Cytologica, 19(2), 126-133.
  More info

  PMID: 47676;Abstract: Computer assisted morphometric analysis of murine lymphocytes obtained by thoracic duct cannulation demonstrates two populations of cells; the larger population (73%) appears to by thymus derived and the remaining 27% is of bone marrow origin. Following exposure to varying amounts of x radiation, morphologic alterations in both populations are evident. The smaller cell populations are evident. The smaller cell population exhibits some of these changes at lower dose levels than does the larger population. In addition, the character of the radiation induced changes appears to be different for the two populations of lymphocytes. After 500 rad, the nuclei of the larger population appears unchanged; the nuclei of the population representing 27% of the cells have become enlarged and vacuolated and are thought to be edematous. After 2000 rad, the nuclei of the larger population appear pyknotic with coarsely clumped chromatin. In the examined set of cells, the smaller population could no longer be detected after 2000 rad. Such disparate responses to radiation induced injury may correlate with known differences in immunologic function which serve to distinguish thymic dependent and bone marrow derived small lymphocytes.
• Bartels, P. H., Olson, G. B., Layton, J. M., Anderson, R. E., & Wied, G. L. (1975). Computer discrimination of T and B lymphocytes. Acta Cytologica, 19(1), 53-57.
  More info

  PMID: 46662;Abstract: The ability to recognize T and B cells on the basis of distribution patterns of the nuclear chromatin is of considerable practical and clinical importance. The analysis of digitized images not only permits an identification of cells, but it also allows one to follow gradual changes in the chromatin patterns and thus to study trends and gradual responses in these cells. This latter information can not be obtained by immunofluorescent techniques. It was therefore of interest to investigate the potential of image analytic techniques in the recognition of T and B cells from an analysis of their Feulgen stained digitized cell images.
• Bhattacharya, P. K., Bartels, P. H., Bibbo, M., Taylor, J., & Wied, G. L. (1975). Estimation procedure for the cellular composition of cervical smears. Acta Cytologica, 19(4), 366-373.
  More info

  PMID: 1057839;Abstract: This paper presents a sampling scheme which allows an estimation of the relative frequencies of occurrence of different cell types and the construction of confidence regions. It allows the setting of standards for a given group of patients, and given sampling sites. Furthermore, the procedure permits an assignment of a sample to a class with known probability of error of the first, and of the second kind. In this communication, the practical data evaluation are given with several numerical examples.
• Bibbo, M., Bartels, P. H., Chen, M., Harris, M. J., Truttman, B., & Wied, G. L. (1975). The numerical composition of cellular samples from the female reproductive tract. I. Carcinoma in situ. Acta Cytologica, 19(5), 438-447.
  More info

  PMID: 1058616;Abstract: The aim of this paper is to show in VCE cell samples [vaginal (V), cervical (C) and endocervical (E) smears] from patients with carcinoma in situ the number of isolated cells versus cells in clusters, the number of well preserved versus degenerated cells, the relative amount of cell degeneration in relationship to site, the relation of carcinoma in situ to nonmalignant cells, the relation of VCE in cell counts, the occurrence of glandular cells and the relative amount of non epithelial elements. The hardware system consisted of a special microscope with a glarex screen top, a 10μ scanning stage interfaced to a LINC 8 computer, and a DEC CRT connected to a PDP 10 computer. Several programs were developed to operate this system. In the 25 cases evaluated an average number of 29 isolated well preserved carcinoma in situ cells were observed in the V smears, 68 in the C smears and 105 in the E smears in a total of 2,000 epithelial cells per V, C or E smear. Isolated well preserved non keratinizing dysplastic cells and dysplastic metaplastic cells were also observed showing the same trend as for carcinoma in situ cells, i.e. more cells in the E smear (38 and 33 respectively) followed by the C smear (22 and 14) and by the V smear (9 and 6). These findings are consistent with studies performed by several investigators, provided that the difference in methods employed for the cell counts is taken into consideration. From the average number of well preserved isolated carcinoma in situ cells found in the C and E smears, it appears that the prospect of detecting and identifying isolated cells by image processing techniques in a prescreening system looks promising.
• Koss, L. G., Bartels, P. H., Bibbo, M., Freed, S. Z., Taylor, J., & Wied, G. L. (1975). Computer discrimination between benign and malignant urothelial cells. Acta Cytologica, 19(4), 378-391.
  More info

  PMID: 1099854;Abstract: Computer discrimination between benign and malignant cells by means of the Taxonomic Intra Cellular Analytic System routines and subroutines was accomplished with a small error of classification, not greater than 10%. One feature of major diagnostic significance, namely, nuclear texture was singled out for discussion. This study augurs well for the future of computerized cytology of the urinary tract.
• Taylor, J., Bahr, G. F., Bartels, P. H., Bibbo, M., Richards, D. L., & Wied, G. L. (1975). Development and evaluation of automatic nucleus finding routines: thresholding of cervical cytology images. Acta Cytologica, 19(3), 289-298.
  More info

  PMID: 1096517;
• Wied, G. L., Bahr, G. F., Bibbo, M., Puls, J. H., Taylor Jr., J., & Bartels, P. H. (1975). The TICAS RTCIP real time cell identification processor. Acta Cytologica, 19(3), 286-288.
  More info

  PMID: 1096516;
• Bahr, G. F., Taylor Jr, J., Bartels, P. H., & Wied, G. L. (1974). Distinguishing normal human blood lymphocytes from lymphocytes in dengue and typhoid fever. VIRCHOWS ARCH.ABT.B ZELL PATH., 16(2), 205-210.
  More info

  PMID: 4140606;Abstract: Blood smears of clinical cases of dengue and typhoid fever were obtained from the Philippines and stained by the Feulgen method. Scanning microscopic densitometry was used to digitize cell images for analysis by principles of pattern recognition. Global features of size and shape, as well as textural features, served to distinguish populations of normal lymphocytes, from lymphocytes of both dengue and typhoid fever. Distinction between dengue and typhoid failed, possibly because in both diseases a shift in the relative frequency of cells in normally occurring subpopulations had taken place as an expression of a general response of the lymphocytic system or because of a failure to detect differences between the 2 diseases at the optical resolution of 0.5μ used.
• Bahr, G. F., Taylor, J., Bartels, P. H., & Wied, G. L. (1974). Distinguishing normal human blood lymphocytes from lymphocytes in dengue and typhoid fever. Virchows Archiv B Cell Pathology, 16(1), 205-210.
  More info

  Abstract: Blood smears of clinical cases of dengue and typhoid fever were obtained from the Philippines and stained by the Feulgen method. Scanning-microscopic densitometry was used to digitize cell images for analysis by principles of pattern recognition. Global features of size and shape, as well as textural features, served to distinguish populations of normal lymphocytes, from lymphocytes of both dengue and typhoid fever. Distinction between dengue and typhoid failed, possibly because in both states of disease a shift in the relative frequency of cells in normally occurring subpopulations had taken place as an expression of a general response of the lymphocytic system or because of a failure to detect differences between the two diseases at the optical resolution of 0.5 μ used. © 1974 Springer-Verlag.
• Bartels, P. H., & Wied, G. L. (1974). Performance testing for automated prescreening devices in cervical cytology. Journal of Histochemistry and Cytochemistry, 22(7), 660-662.
  More info

  PMID: 4853581;Abstract: Performance specifications for automated prescreening devices for cervical cancer will concern rates for missed positive samples and rates for false alarms. Testing whether a given device adheres to such specifications within given confidence limits should be feasible for the missed positive rates. The test for adherence to the specified false alarm rate presents a serious sampling problem.
• Bartels, P. H., Bahr, G. F., Bibbo, M., Richards, D. L., Sonek, M. G., & Wied, G. L. (1974). Analysis of variance of the papanicolaou staining reaction. Acta Cytologica, 18(6), 522-531.
  More info

  PMID: 4141206;
• Bartels, P. H., Bahr, G. F., Jeter, W. S., Olson, G. B., Taylor Jr., J., & Wied, G. L. (1974). Evaluation of correlational information in digitized cell images. Journal of Histochemistry and Cytochemistry, 22(2), 69-79.
  More info

  PMID: 4274623;Abstract: The processing of digitized cell images by computer may yield information not only about a set of cell image properties but also about their mutual dependencies. Observation of the covariance matrix of image properties of cellular material showing a response to chemotherapeutic treatment, ionizing radiation or antigenic challenge or following a developmental trend may permit a quantitative description of small trendal charges. The covariance between certain cell image properties may show statistically significant changes before the mean values of the image properties are affected. Methods of reducing the dimensionality of the representation in an efficient manner are described.
• Bartels, P. H., Bhattacharya, P. K., Bellamy, J. C., Bahr, G. F., Bibbo, M., & Wied, G. L. (1974). Computer generated, synthetic cell images. Acta Cytologica, 18(2), 155-164.
  More info

  PMID: 4133110;Abstract: The desire to automate certain aspects of cytologic prescreening may be responsible for some of the attempts at quantitation, but most of the problems demanding objective quantitative answers exist in their own right. Recent results of objective cell image analysis suggest that there may in fact not even be an 'objective' basis for the assumption of discretely different, distinct cell types. Subjectively these are well established entities, but they may well be the result of complex psychophysical human recognition mechanisms, rather than a matter of fact. To understand any process it is helpful to bring it under experimental control, and to make it reproducible. As far as decision making on cytologic material is concerned, this means that one would have to be able to simulate, from a given stochastic model, the images of cells with known properties, and to have such images assessed by qualified cytologists. With the generation of cell images under computer control, each determining parameter defining the cell image properties can be varied. It should then be possible to explore in an iterative fashion the processes entering into diagnostic decision making. This paper presents the approach taken, and some initial results obtained with a program generating synthetic cell images by computer.
• Bartels, P. H., Bibbo, M., Taylor, J., & Wied, G. L. (1974). Cell recognition from the statistical dependence of gray values in digitized images. Acta Cytologica, 18(2), 165-169.
  More info

  PMID: 4133111;Abstract: The interdependencies between gray values recorded in digitized images of cells from normal patients, and from patients with malignant disease, are examined. Consistent differences between such dependencies are found in the material. Cell recognition on the basis of the covariance structure of gray values is possible with acceptable error rates when value strings of at least ten values are considered. It is found that a model relying on two way interdependencies is sensitive to training set inhomogeneities, and that the additional computational effort is not warranted.
• Bartels, P. H., Olson, G. B., Jeter, W. S., & Wied, G. L. (1974). Evaluation of unsupervised learning algorithms in the computer analysis of lymphocytes. Acta Cytologica, 18(5), 376-388.
  More info

  PMID: 4280223;
• Olson, G. B., Anderson, R. E., & Bartels, P. H. (1974). Differentiation of murine thoracic duct lymphocytes into T and B subpopulations by computer cell-scanning techniques. Cellular Immunology, 13(3), 347-355.
  More info

  PMID: 4141647;Abstract: Computer-assisted cytometric analysis of murine lymphocytes, and of pure populations of T and B cells obtained from thoracic duct lymphocytes (TDL) revealed the existence of descriptors capable of differentiating T and B lymphocytes. Optical density values of Feulgen-stained nuclear DNA of the cells were recorded as digitized images by a computerized scanning microscope. Computer programs extracted from each cell and for each cell population computable image information related to the DNA distribution patterns. Nuclei of B cells possess smaller areas, less total optical density, denser staining DNA granules than the nuclei of T cells and distinct differences in the arrangement of optical density values. These descriptors allowed the differentiation of TDL into two subpopulations possessing properties identical to those found in pure T and B cell populations. © 1974.
• Taylor, J., Bartels, P. H., Bibbo, M., Bahr, G. F., & Wied, G. L. (1974). Implementation of a hierarchical cell classification procedure. Acta Cytologica, 18(6), 515-521.
  More info

  PMID: 4614642;
• Bahr, G. F., Bartels, P. H., Bibbo, M., Nicolas, M. D., & Wied, G. L. (1973). Evaluation of the Papanicolaou stain for computer assisted cellular pattern recognition.. Acta Cytologica, 17(2), 106-112.
  More info

  PMID: 4120547;
• Bartels, P. H., Bibbo, M., Bahr, G. F., Taylor, J., & Wied, G. L. (1973). Cervical cytology: descriptive statistics for nuclei of normal and atypical cell types. Acta Cytologica, 17(5), 449-453.
  More info

  PMID: 4517960;Abstract: Cells from 8 categories recognized in cervical squamous cytology (superficial cells, intermediate cells, parabasal cells, metaplastic cells, metaplastic dysplastic cells, non keratinizing dysplastic cells, cells from carcinoma in situ, and cells from invasive carcinoma) were scanned, and the extinction value histograms for the nucleus alone were analyzed for trendal changes from category to category. The coefficients of the first eigenvector for the covariance matrix for each category show a continuous trend from category to category, and the ordering of categories is in agreement with that observed by cytopathologists.
• Bhattacharya, P. K., Bartels, P. H., Taylor, J., & Wied, G. L. (1973). A decision procedure for automated cytology: test statistic for detecting sample abnormality, and inadequacy. Acta Cytologica, 17(6), 538-548.
  More info

  PMID: 4585160;
• Bibbo, M., Bartels, P. H., Bahr, G. F., Taylor, J., & Wied, G. L. (1973). Computer recognition of cell nuclei from the uterine cervix.. Acta Cytologica, 17(4), 340-350.
  More info

  PMID: 4579185;
• Jahoda, E., Bartels, P. H., Bibbo, M., Bahr, G. F., Holzner, J. H., & Wied, G. L. (1973). Computer discrimination of cells in serous effusions. I. Pleural fluids.. Acta Cytologica, 17(2), 94-105.
  More info

  PMID: 4570552;
• Jahoda, E., Bartels, P. H., Bibbo, M., Bahr, G. F., Holzner, J. H., & Wied, G. L. (1973). Computer discrimination of cells in serous effusions. II. Peritoneal fluid. Acta Cytologica, 17(6), 533-537.
  More info

  PMID: 4585159;Abstract: The feasibility of discriminating between 'normal mesothelial cells' in peritoneal fluids, reactive mesothelial cells, and ovarian and esophageal tumor cells by means of digitized image analysis is investigated. Satisfactory classification results were attained in both training and object sets; the image properties most suitable for discrimination, error rates, and likelihoods for correct recognition of individual cells are given for each of the cell types.
• Olson, G. B., Wied, G. L., & Bartels, P. H. (1973). Differentiation of chicken thymic and bursal lymphocytes by cell image analysis. Acta Cytologica, 17(5), 454-461.
  More info

  PMID: 4542792;Abstract: Thymic and bursal lymphoid cell populations of embryonic and postnatal chicks were assayed by objective image analysis. Descriptors derived from the distribution of the nuclear chromatin lend themselves as indicators of trends in the development of lymphoid cells during morphogenesis. By day 7 of postnatal life it is possible to differentiate between small lymphocytes from the bursa and the thymus with approximately an 80% correct classification.
• Olson, G. B., Wied, G. L., & Bartels, P. H. (1973). Differentiation of lymphoid tissue by analysis of digitized images.. Acta Cytologica, 17(2), 89-93.
  More info

  PMID: 4540223;
• Wied, G. L., & Bartels, P. H. (1973). Modern trends in cytologic diagnosis (report) (author's transl). Verhandlungen der Deutschen Gesellschaft fur Pathologie, 57, 169-180.
  More info

  PMID: 4142169;
• Zajicek, J., Bartels, P. H., Bahr, G. F., Bibbo, M., Jakobsson, P. A., & Wied, G. L. (1973). Computer analysis of needle aspirates from breast carcinomas during radiotherapy. Acta Cytologica, 17(3), 179-187.
  More info

  PMID: 4121637;Abstract: The fine needle aspiration method was applied for sampling breast carcinoma cells prior to and during radiotherapy for morphological study. Three breast carcinoma cases exhibiting different morphological features in smear preparations were analyzed. The aspirate from carcinoma CA 61 consisted of ductal carcinoma cells with rather large nuclei, that from CA 64 consisted of cytoplasm rich cells of apocrine type and that of CA 66 of small sized ductal cells. By the TICAS programs the cell populations from these 3 carcinomas can be separated from each other. By comparison of cytophotometric data prior to irradiation with those recorded at 800 rads and 1600 rads the changes in cellular morphology after a given dose can be evaluated. The analysis of cytophotometric data shows that the differences between these 3 types of carcinoma persist after the dosis of 800 rads and 1600 rads. Collection of cells by fine needle aspiration biopsy during irradiation and processing of cytophotometric data by TICAS programs may open a new approach to the clinical study on the effect of ionizing irradiation on neoplastic cells.
• Bartels, P. H., Bahr, G. F., Bibbo, M., & Wied, G. L. (1972). Objective cell image analysis.. Journal of Histochemistry and Cytochemistry, 20(4), 239-254.
  More info

  PMID: 5031849;
• Meulen, V. t., Bartels, P. H., Bahr, G. F., Bibbo, M., Cremer, N., Lennette, E. H., & Wied, G. L. (1972). Computer assisted analysis of a carrier culture infected with Moloney leukemia virus.. Acta Cytologica, 16(5), 454-463.
  More info

  PMID: 4561245;
• Zajicek, J., Bartels, P. H., Bahr, G. F., Bibbo, M., & Wied, G. L. (1972). Computer analysis of lymphocytes from cases with lymphadenitis and lymphocytic lymphoma.. Acta Cytologica, 16(4), 284-296.
  More info

  PMID: 4556410;
• Bartels, P. H., Layton, J. M., Jarkowski, T. L., Bellamy, J. C., Bahr, G. F., & Wied, G. L. (1971). Cell recognition by multivariate gray value analysis in digitized images.. Acta Cytologica, 15(3), 284-288.
  More info

  PMID: 4933976;
• Bhattacharya, P. K., Bartels, P. H., Bahr, G. F., & Wied, G. L. (1971). A test statistic for detecting the presence of abnormal cells in a sample.. Acta Cytologica, 15(6), 533-544.
  More info

  PMID: 5289970;
• Jarkowski, T. L., Layton, J. M., Bahr, G. F., Wied, G. L., Bellamy, J. C., & Bartels, P. H. (1971). Computer recognition of cells from asymptomatic lymphocytic leukemia. I. Methodologic study.. Acta Cytologica, 15(2), 147-153.
  More info

  PMID: 4933105;
• BARTELS, P. H., & BELLAMY, J. (1970). SELF-OPTIMIZING, SELF- LEARNING SYSTEM IN PICTORIAL PATTERN RECOGNITION. Applied Optics, 9(11), 2453-2458.
  More info

  Abstract: A system of computer programs discriminates between pictorial patterns by determining a substantial number of numerically encoded pattern properties. Supervised learning is used to find both an optimum decision sequence and the thresholds for decision rules. These are applied to patterns from an object set to test the consistency of the classification procedure. Nonsupervised learning is used in the pattern detection section of the program. The system has been extensively tested in the discrimination of biomedical patterns from their digitized microscopic images, specifically in the machine recognition of tumor cells andof cells involved in immune response. Objective cytodiagnostic decision making was shown to be more consistent, and, in certain instances, capable of finer discrimination than assessment by qualified human observers.
• Bartels, P. H., Bahr, G. F., Bellamy, J. C., Bibbo, M., Richards, D. L., & Wied, G. L. (1970). A self-learning computer program for cell recognition.. Acta Cytologica, 14(8), 486-494.
  More info

  PMID: 4926009;
• Bartels, P. H., Bahr, G. F., Calhoun, D. W., & Wied, G. L. (1970). Cell recognition by neighborhood grouping techniques in TICAS.. Acta Cytologica, 14(5), 313-324.
  More info

  PMID: 4935161;
• Bibbo, M., Bartels, P. H., Bahr, G. F., Ng, A. B., Reagan, J. W., Richards, D. L., & Wied, G. L. (1970). Data bank for endometrial cells. Operation of the TICAS file project.. Acta Cytologica, 14(9), 574-582.
  More info

  PMID: 5277454;
• Wied, G. L., Bahr, G. F., Bartels, P. H., & Bibbo, M. (1970). Cytologic diagnosis by computer.. Minerva Ginecologica, 22(23), 1168-1170.
  More info

  PMID: 4930078;
• Wied, G. L., Bibbo, M., Bahr, G. F., & Bartels, P. H. (1970). Computerized microdissection of cellular images.. Acta Cytologica, 14(7), 418-433.
  More info

  PMID: 4921860;
• Wied, G. L., Bibbo, M., Bahr, G. F., & Bartels, P. H. (1970). Computerized recognition of uterine glandular cells. 3. Assessment of the efficacy of 1 micron vs 0.5 micron spot sizes.. Acta Cytologica, 14(3), 1361-.
  More info

  PMID: 5265917;
• Bartels, P. H., Bahr, G. F., & Wied, G. L. (1969). Cell recognition from line scan transition probability profiles.. Acta Cytologica, 13(4), 210-217.
  More info

  PMID: 5253403;
• Bartels, P. H., Bahr, G. F., Griep, J., Rappaport, H., & Wied, G. L. (1969). Computer analyses of lymphocytes in transformation. A methodologic study.. Acta Cytologica, 13(10), 557-568.
  More info

  PMID: 5260006;
• Wied, G. L., Bahr, G. F., Oldfield, D. G., & Bartels, P. H. (1969). Computer assisted identification of cells from uterine adenocarcinoma. II. Measurements at 590 nm.. Acta Cytologica, 13(1), 21-26.
  More info

  PMID: 5248716;
• Wied, G. L., Bartels, P. H., & Bahr, G. F. (1969). Laboratory organization in the detection and diagnosis of early cervical neoplasia.. Obstetrical and Gynecological Survey, 24(7 Pt 2), 935-966.
  More info

  PMID: 5212439;
• Wied, G. L., Bartels, P. H., Bahr, G. F., & Reagan, J. W. (1969). TICAS assessment of cells from atypical hyperplasia of the endometrium.. Acta Cytologica, 13(10), 552-556.
  More info

  PMID: 5260005;
• Bartels, P. H., Bahr, G. F., & Wied, G. L. (1968). Cell recognition by cluster analysis in pure parameter hyperspaces.. Acta Cytologica, 12(5), 371-380.
  More info

  PMID: 5245329;
• Bartels, P. H., Wied, G. L., & Bahr, G. F. (1968). Cell recognition from equiprobable extinction range contours.. Acta Cytologica, 12(3), 205-217.
  More info

  PMID: 5241461;
• Wied, G. L., Bahr, G. F., Oldfield, D. G., & Bartels, P. H. (1968). Computer-assisted identification of cells from uterine adenocarcinoma. A clinical feasibility study with TICAS. I. Measurements at wavelength 530 nm.. Acta Cytologica, 12(5), 357-370.
  More info

  PMID: 5245328;
• Wied, G. L., Bartels, P. H., Bahr, G. F., & Oldfield, D. G. (1968). Taxonomic intra-cellular analytic system (TIICAS) for cell identification.. Acta Cytologica, 12(3), 180-204.
  More info

  PMID: 5241460;
• Douglas, S. D., Spicer, S. S., & Bartels, P. H. (1966). Microspectrophotometric analysis of basic protein rich sites stained with Biebrich scarlet.. Journal of Histochemistry and Cytochemistry, 14(4), 352-360.
  More info

  PMID: 4164216;
• Bartels, P. H. (1963). Polarized light analysis. Microchemical Journal, 7(1), 98-119.