Search

To reduce the problem of heterogeneity in astrocytic tumors by means of computer-assisted microscope analysis of Feulgen-stained nuclei.Thirty-eight glial tumors for which we obtained 227 stereotactic biopsies were subjected to digital cell image analysis of Feulgen-stained nuclei. This series of 38 glial tumors included 36 supratentorial astrocytic tumors (13 astrocytomas, 7 anaplastic astrocytomas and 16 glioblastoma multiformes) and 2 grade 3 astrocytic tumors of the cerebellum.The results suggest a new methodology, enabling the biologic characteristics of the brain parenchymal area surrounding a given glial tumor to be characterized. This methodology relies on the performance of three successive steps. The first is quantitative characterization of nuclear morphology and its chromatin pattern by means of 15 morphonuclear variables. This characterization is carried out by means of the computer-assisted microscope analysis of Feulgen-stained nuclei. The second step consists of setting up morphonuclear data banks, with each process giving the precise portrait of a given cell nuclear population. This process is carried out by means of multivariate analysis, taking into account the 15 variables mentioned above. Multivariate analysis includes principal components analysis followed by the canonical transformation of the data. The third step consists of testing unknown cases against these morphonuclear data banks. This is carried out by means of linear discriminant analysis, which enables the various cell nuclear types in the stereotactic biopsy to be quantified.The present methodology makes it possible to investigate whether infiltrating tumor cells are present in or absent from the parenchymal brain area surrounding a glial tumor. It can therefore contribute additional information to that contributed by computed tomography and/or magnetic resonance imaging with respect to the precise delineation of the volume of a brain tumor. This delineation must be as precise as possible to allow total surgical resection of the tumor and prevention of its recurrence.

We investigated whether a relationship exists in terms of growth pattern and hormone sensitivity in 18 gastrointestinal neoplastic cell lines. Hormones studied included gastrin, epidermal growth factor, estradiol and luteinizing hormone-releasing hormone.The growth patterns were assessed by means of computer-assisted microscope analysis of Feulgen-stained nuclei combined with the mathematical Delaunay triangulation and Voronoi paving techniques. This methodology enabled four variables characterizing the cell colony patterns to be computed. The information contributed by these variables was analyzed by means of discriminant analysis and the decision tree technique.Each phenotype (sensitivity level) exhibited distinct growth pattern (or cell colony) characteristics in the case of each hormone and/or growth factor under study. Furthermore, the sensitivity of the gastrointestinal cell lines to a given hormone (or growth factor) appeared to be peculiar to the hormone (or growth factor).A direct relationship seems to exist between growth pattern and hormone sensitivity levels in gastrointestinal cancers, particularly colorectal.

The aim of the present work is to show that decision tree induction algorithms are a useful tool for extracting reliable information from data series, with the objective of assisting pathologists in identifying specific diagnostic and prognostic markers in various types of tumor pathologies. In terms of accuracy, we show that the decision tree technique exceeds other more sophisticated techniques, such as multilayer neural networks. Furthermore, because of the case with which decision tree results can be interpreted (logical classification rules), new methodologies can be readily developed to further assist in analyzing complex data that mix heterogeneous features. In this paper, we illustrate such capabilities in the context of different complex diagnostic and/or prognostic problems in tumor pathology relating to bladder, astrocytomas, and adipose tissues.