Image Cytometry of Duct Cells from Benign and Malignant Breast Disease

摘要

We used image cytometry to classify benign and malignant breast lesions, to identify parameters that classify premalignant lesions, and to assess these parameters for a biological association that may be linked with malignancy. Results suggest that quantitative image analysis is able to discriminate ''normal'' from ''abnormal'' breast lesions. Thus image cytometry may provide an objective approach for early detection of malignant changes in the breast. The parameters selected by stepwise discriminant and regression analyses as being most useful for discriminating among breast lesions relate to nuclear size, shape and DNA content. These parameters are robust and easily measured and are similar to those used by others for discriminating benign from malignant lesions. Significantly, no parameter based on nuclear chromatin distribution ranked as highly in this study, unlike our experience with studies of other premalignant lesions. 9 refs., 1 fig., 3 tabs. (ERA citation 12:023618)