Fig. 1From: Development and validation of an AI-enabled digital breast cancer assay to predict early-stage breast cancer recurrence within 6 yearsPDxBr AI-Digital Grade Features. Representative images of A–B tumor-infiltrating lymphocyte feature that detects and quantitates lymphocytes (highlighted in blue) in the periepithelial area of invasive tumor (dark overlay); C-D nuclear pleomorphism feature that quantitates the difference between the largest epithelial nuclei (outlined in red) and the average nuclei (outlined in green) in the invasive tumor; and E–F mitotic count feature, which uses clustering methods to identify the regions of highest proliferation (yellow outline—equivalent to 10 high-power fields)Back to article page