Prognostic value of morphometry and DNA flow-cytometry features of invasive breast cancers detected by population screening: comparison with control group of hospital patients.

Using the prognostic value of morphometric and flow-cytometric features, a group of patients with invasive breast cancers detected with population screening (PS, n = 70) has been evaluated and compared with a random control group in 2 hospitals (H group, n = 225) diagnosed in the same period. The results show that the PS patients had smaller tumors, less positive lymph nodes, better differentiated tumors with a lower mitotic activity index (MAI) and lower values of the morphometric prognostic index (MPI). Furthermore, the women more frequently had diploid tumors and tumors with small nuclei. The second purpose was to evaluate whether quantitative microscopical features, in comparison with other prognostic features such as size of primary tumor, nodal status and histologic grade, are as strong prognosticators in PS tumors as in H-detected breast cancers. In comparison with H tumors, morphometric and flow-cytometric features, as well as tumor size, had the same prognostic value for the PS tumors. In contrast, nodal status was not significant within the PS group, and the same phenomenon was found in a subgroup of H patients with similar sized tumors. Of all quantitative microscopical features (MPI, MAI, mean nuclear area (MNA) and DNA Index (DI], the MAI had the strongest prognostic value. DI showed additional prognostic value to the MAI for patients with small tumors and with small tumor-cell nuclei, because a diploid pattern in these cases (this combination occurred in 21 patients of the total group = 30%) was correlated with a 95% 10-year survival rate. Histologic grade, although significant within the large H group, was of no prognostic value within the PS group, and also not as in the H sub-group with small tumors. It is concluded from morphometric and DNA flow-cytometric criteria that these prognostic features in invasive breast cancers detected by PS were all more favorable than in randomly detected hospital breast cancers. This may account for the reported better survival rate of PS patients. Furthermore, the prognosis of patients with small invasive breast cancers detected by population screening can be more accurately deduced by quantitative microscopical features than by axillary-lymph-node status.

Evaluation of contextual analysis for computer classification of cervical smears.

A procedure for automated analysis of cervical smears has been implemented in an image cytometry system. Smears are described exclusively in terms of global and contextual information extracted by pattern-recognition algorithms and represented by a vector of proportions of cellular object types. Linear discriminant functions, based on a Fisher criterion, are derived to classify smears with a cross-section of diagnoses into two broad categories, normal and abnormal. Results obtained from 83 smears indicate 78% correct classification. In contrast to most automated systems, good classification results were obtained in normal smears with benign changes caused by inflammation and with postmenopausal atrophia and in abnormals with mild dysplasia. These findings suggest that contextual analysis may be sensitive to subtle changes in cellular morphology and to progressive patterns of dysplasia. When used with standard isolated cell analysis, contextual analysis may provide additional complementary information for automated cervical prescreening.