Introduction: Active surveillance (AS) is now incorporated into many prostate cancer (PCa) management guidelines to reduce overtreatment, but it requires frequent monitoring because of the risk of disease progression. Improved risk-stratification tools are needed to distinguish indolent from aggressive disease early in surveillance so that the window for curative treatment is not missed. At many institutions, multiparametric MRI (mpMRI) and MRI-ultrasound (MRI-US) fusion biopsies are standard of care for interval monitoring of patients on AS. Here, we show that a quantitative mpMRI system, the Habitat Risk Score (HRS), identifies differential lesion growth on longitudinal mpMRI and correlates strongly with quantitative analysis of digital pathology from radical prostatectomy (RP) specimens. Study Methods: Patients from “MRI-Guided Biopsy Selection of PCa Patients for Active Surveillance versus Treatment: The Miami MAST Trial” (ClinicalTrials.gov: NCT02242773; total accrual = 208) with confirmed PCa underwent 12-36 months of AS prior to RP. The mpMRI exams consisted of T2-weighted (T2W), Dynamic Contrast Enhanced (DCE)-MRI, and Diffusion Weighted Imaging (DWI) acquired according to PI-RADSv2 recommendations. The HRS approach automatically identifies suspicious prostate lesions on mpMRI by assigning a 10-point pixel-by-pixel risk score presented as a heat map, overlaid on the T2W. Longitudinal mpMRI images, biopsy pathology at 12, 24 or 36-months of surveillance and RP specimen pathology were analyzed. The temporal changes in HRS mpMRI volumes were tracked and HRS maps from the last time point (prior to RP) were examined for correlation with H&E and digital pathology. Results: Detailed HRS mpMRI review of 34 MAST participants with biopsy progression followed shortly (within six months) by RP showed that 55% exhibited mpMRI evidence of histopathological progression that was not detected on their last surveillance biopsy. Standard-of-care prostate mpMRI, including that used in MAST, is limited by subjective interpretation, and even expert radiologists may miss significant lesions. In contrast, HRS provides an objective, quantitative assessment of prostate mpMRI images. Findings on HRS correlated closely with both expert pathology review and quantitative digital pathology, with the strongest spatial concordance observed between HRS maps and digital pathology. Conclusion: Our results indicate the utility of HRS mpMRI to detect tumor growth in AS patients. The differential growth in the lesions can be used to guide tissue selection for genomic testing. The study found excellent correlation between HRS and H&E and digital pathology from RP specimens. HRS volumes may serve as quantitative biomarkers for early detection of progression.