Nuclear Morphometry in Flat Epithelial Atypia of the Breast as a Predictor of Malignancy A Digital Image-based Histopathologic Analysis

OBJECTIVE: To identify morphometric features unique to flat epithelial atypia associated with cancer using digital image analysis. STUDY DESIGN: Cases with diagnosis of flat epithelial atypia were retrieved and divided into 2 groups: flat epithelial atypia associated with invasive or in situ carcinoma (n= 31) and those without malignancy (n=27). Slides were digitally scanned. Nuclear features were analyzed on representative images at 20 x magnification using digital morphometric software. RESULTS: Parameters related to nuclear shape and size (diameter, perimeter) were similar in both groups. However, cases with malignancy had significantly higher densitometric green (p =0.02), red (p = 0.03), and grey (p = 0.02) scale levels as compared to cases without cancer. A mean grey densitometric level >119.45 had 71% sensitivity and 70.4% specificity in detecting cases with concomitant carcinoma. CONCLUSION: Morphometry of features related to nuclear staining appears to be useful in predicting risk of concurrent malignancy in patients with flat epithelial atypia, when added to a comprehensive histo pathologic evaluation.