Quantitative and Automated MRI Cancer Risk Maps used for Identification of Prostate Cancer Progression

This study’s objective was to determine whether automated mpMRI cancer risk maps could identify prostate cancer progression during active surveillance. Derived lesion masks were used to analyze factors for progression. A decision tree model for progression was generated with sensitivity = 0.84, specificity = 0.56, and ROC AUC 0.75. The identification results indicate the potential of mpMRI and MRI cancer risk maps to assist in identifying progression during prostate cancer active surveillance.