Does flat epithelial atypia have rounder nuclei than columnar cell change/hyperplasia? A morphometric approach to columnar cell lesions of the breast.

Columnar cell lesions of the breast encompass columnar cell change/hyperplasia (CCC/CCH) and flat epithelial atypia (FEA). These have attracted researchers because emerging data suggest that FEA may represent the earliest histologically detectable non-obligate precursor of breast cancer. However, it is occasionally difficult to distinguish FEA from CCC/CCH because of similar histology. Although the nuclei of FEA are frequently described as relatively round compared with those of CCC/CCH, there are few morphometric studies to support this statement. The aim of this study was to provide objective data as to the nuclear shape in columnar cell lesions. As a shape descriptor, we adopted ellipticity that is defined by the formula 2b/2a, where a is the length of the long axis of the ellipse and b is the length of the short axis. Contrary to circularity, ellipticity reflects the overall configuration of an ellipse irrespective of surface irregularity. Our image analysis included generating whole slide images, extracting glandular cell nuclei, measuring nuclear ellipticity, and superimposing graded colors based on execution of results on the captured images. A total of 7917 nuclei extracted from 22 FEA images and 5010 nuclei extracted from 13 CCC/CCH images were analyzed. There was a significant difference in nuclear roundness between FEA and CCC/CCH with mean ellipticity values of 0.723 and 0.679, respectively (p < 0.001, Welch's t test). Furthermore, FEA with malignancy had significantly rounder nuclei than FEA without malignancy (p < 0.001). Our preliminary results suggest that nuclear ellipticity is a key parameter in reproducibly classifying columnar cell lesions of the breast.