Texture-Analysis Classification of Intraductal Carcinoma of the Prostate Using Label-Free Multimodal Nonlinear Optical Imaging

Abstract

Intraductal carcinoma of the prostate (IDC-P) is a very aggressive histopathological subtype of prostate cancer (PCa) for which no accurate biomarkers exist. In our work, we apply a multimodal nonlinear optical imaging approach that uses second-harmonic generation (SHG) and stimulated Raman scattering (SRS) imaging to distinguish IDC-P from regular PCa and benign prostate tissue. Images from each tissue type were classified using support vector machine (SVM). The technique classified the images from each region based on first-order statistics and texture-based second-order statistics derived from the gray-level co-occurrence matrix (GLCM) of the images. Our results demonstrate that SVM models trained on either SHG or SRS images accurately classify IDC-P as well as high-grade PCa, low-grade PCa, and benign tissue with a mean classification accuracy of over 89%. Furthermore, a classification model combining both SHG and SRS imaging modalities can accurately classify all tissue types with a mean classification accuracy of 98%.