Prostate ISUP classification

This software implements a AI pipeline for the detection of clinically significant prostate cancer (csPCa) from multiparametric MRI images combined with routine clinical data. The specific tasks performed by the software include: radiomics extraction from whole-gland prostate gland and sub-regions based on T2-weighted and DWI from MRI sequences; data harmonization at both image and feature levels to correct for vendor and acquisition batch effects; and classification using machine learning models to stratify patients according to the likelihood of clinically significant prostate cancer. The inference model is available.