Global acetylation and methylation changes predict papillary urothelial neoplasia of low malignant potential recurrence: a quantitative analysis.

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.

Karyometry in the early detection and chemoprevention of intraepithelial lesions.

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.

Karyometry detects subvisual differences in chromatin organisation state between non-recurrent and recurrent papillary urothelial neoplasms of low malignant potential.

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<0.005 for more than 20 features. Based on significance, six texture features were selected for discriminant analysis. Stepwise linear discriminant analysis reduced Wilk's lambda to 0.87, indicating a highly significant difference between the two multivariate data sets, but only modest ability to discriminate (70% correct case classification). A box sequential classifier was used based on data derived from discriminant analysis. The classifier took three classification steps and classified 19 of the 20 cases correctly (95% correct case classification). To determine whether significant case grouping could also be obtained based on an objective criterion, the merged data sets of non-recurrent and recurrent cases were submitted to the unsupervised learning algorithm P-index. Two clusters were formed with significant differences. The subsequent application of a Cooley/Lohnes classifier resulted in an overall correct case classification rate of 85%. CONCLUSIONS: Karyometry and multivariate analyses detect subvisual differences in chromatin organisation state between non-recurrent and recurrent PUNLMPs, thus allowing identification of lesions that do or do not recur.

Computerized diagnostic decision support system for the classification of preinvasive cervical squamous lesions.

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 kappa statistics. Intraobserver reproducibility using conventional unaided diagnosis was reasonably good (kappa range, 0.688 to 0.861), but interobserver agreement was poor (kappa 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.

Expert system support using a Bayesian belief network for the classification of endometrial hyperplasia.

Accurate morphological classification of endometrial hyperplasia is crucial as treatments vary widely between the different categories of hyperplasia and are dependent, in part, on the histological diagnosis. However, previous studies have shown considerable inter-observer variation in the classification of endometrial hyperplasias. The aim of this study was to develop a decision support system (DSS) for the classification of endometrial hyperplasias. The system used a Bayesian belief network to distinguish proliferative endometrium, simple hyperplasia, complex hyperplasia, atypical hyperplasia and grade 1 endometrioid adenocarcinoma. These diagnostic outcomes were held in the decision node. Four morphological features were selected as diagnostic clues used routinely in the discrimination of endometrial hyperplasias. These represented the evidence nodes and were linked to the decision node by conditional probability matrices. The system was designed with a computer user interface (CytoInform) where reference images for a given clue were displayed to assist the pathologist in entering evidence into the network. Reproducibility of diagnostic classification was tested on 50 cases chosen by a gynaecological pathologist. These comprised ten cases each of proliferative endometrium, simple hyperplasia, complex hyperplasia, atypical hyperplasia and grade 1 endometrioid adenocarcinoma. The DSS was tested by two consultant pathologists, two junior pathologists and two medical students. Intra- and inter-observer agreement was calculated following conventional histological examination of the slides on two occasions by the consultants and junior pathologists without the use of the DSS. All six participants then assessed the slides using the expert system on two occasions, enabling inter- and intra-observer agreement to be calculated. Using unaided conventional diagnosis, weighted kappa values for intra-observer agreement ranged from 0.645 to 0.901. Using the DSS, the results for the four pathologists ranged from 0.650 to 0.845. Both consultant pathologists had slightly worse weighted kappa values using the DSS, while both junior pathologists achieved slightly better values using the system. The grading of morphological features and the cumulative probability curve provided a quantitative record of the decision route for each case. This allowed a more precise comparison of individuals and identified why discordant diagnoses were made. Taking the original diagnoses of the consultant gynaecological pathologist as the 'gold standard', there was excellent or moderate to good inter-observer agreement between the 'gold standard' and the results obtained by the four pathologists using the expert system, with weighted kappa values of 0.586-0.872. The two medical students using the expert system achieved weighted kappa values of 0.771 (excellent) and 0.560 (moderate to good) compared to the 'gold standard'. This study illustrates the potential of expert systems in the classification of endometrial hyperplasias.

Transcontinental communication and quantitative digital histopathology via the Internet; with special reference to prostate neoplasia.

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/microm) and high power (6 pixels/microm) 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.

Quantitative study of breast cancer progression: different pathways for various in situ cancers.

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.

Quantitative study of ductal breast cancer progression: nuclear signatures for evaluation of progression grade.

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.

Measurement of chemopreventive efficacy in skin biopsies.

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-dlornithine (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.

Karyometry of secretory cell nuclei in high-grade PIN lesions.

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. CONCLUSIONS: 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.

Multimegapixel images in histopathology.

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 (Park-ridge, 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 x 470, 512 x 470 or 1,496 x 1,120 pixels, respectively. The slide was moved while mounted on a Maerzheuser scanning stage with 0.1-micron 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 degree, very large arrays need substantial tile overlap. For object areas extending over only 5-10 mm, the effects can be controlled.

Nuclear chromatin patterns in normal, hyperplastic and atypical endometrium.

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.

Progression curves for endometrial lesions.

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.

Karyometry in Barrett's esophagus.

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.

Nuclear chromatin characteristics of breast solid pattern ductal carcinoma in situ.

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.

Early stromal invasion in cervical cancer: immunocytochemical changes occurring in infiltrating neoplastic cells.

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.

Quantitative study of ductal breast cancer progression: signatures of nuclei in proliferating breast lesions and in situ cancer.

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.

An automated machine vision system for the histological grading of cervical intraepithelial neoplasia (CIN).

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.

Data representation and reduction for chromatin texture in nuclei from premalignant prostatic, esophageal, and colonic lesions.

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.