PREDICTING PROSTATE CANCER RISK OF PROGRESSION WITH MULTIPARAMETRIC MAGNETIC RESONANCE IMAGING USING MACHINE LEARNING AND PERITUMORAL RADIOMICS

摘要

Embodiments facilitate stratification of a patient according to prostate cancer (PCa) risk. A first set of embodiments relates to training of a machine learning classifier to compute a probability that a patient has a low-risk of PCa progression based on intratumoral radiomic features and peritumoral radiomic features extracted from multi-parametric magnetic resonance imaging (mpMRI) images. A second set of embodiments relates to classifying a patient as low-risk of PCa progression, or high-risk of PCa progression, based on radiomic features extracted from mpMRI imagery of the patient.