Hyperspectral Imaging for Distinguishing Granulomatous Mastitis From Normal Breast Tissue in Core Needle Biopsies.

This study explores the use of hyperspectral imaging (HSI) to distinguish granulomatous mastitis (GM) from normal breast tissue in core-needle biopsy specimens. High-resolution spectral data were captured from paraffin-embedded sections across the 400-1000 nm range. Following preprocessing, normalization, and principal component analysis, one-way analysis of variance revealed ten wavelengths with the greatest diagnostic power. Notably, hemoglobin absorption peaks and lipid-collagen signatures provided the strongest spectral discrimination between GM and healthy tissue. These results underscore HSI's promise as a rapid, real-time adjunct for preoperative and intraoperative evaluation of breast lesions.