Quantitative MRI to automatically delineate intraprostatic tumor for radiation therapy

PROJECT SUMMARY Prostate cancer (PCa) is the most common cancer among men. Radiation therapy is an integral part of the standard of care for the treatment of PCa. Local recurrences of PCa after radiation therapy usually originate from the primary tumor site. A dose escalation to the clinically significant tumor (csT) foci in intermediate- and

high-risk patients is reported to result in significantly improved biochemical disease-free survival (bDFS) without increasing radiation toxicity. Contouring of the intraprostatic tumor is currently based on multiparametric (mp-) magnetic resonance imaging (MRI) and the interpretation of the MRI results are based

on Prostate Imaging Reporting & Data System (PI-RADS). However, relatively large variability in performance of mp-MRI, including that of PI-RADS, continue to pose barriers to identify clinically significant prostate tumor as treatment target and improve patient outcome. In this project, we propose to address the issues in both the

MRI technique and scoring system aspects and develop a tool that can automatically delineate intraprostatic csT and be easily transferred to clinical use. To achieve this goal, in Aim 1, we will build a quantitative MRI (QMRI) toolset including magnetic resonance fingerprinting (MRF), arterial spin labeling (ASL), and diffusion-

weighted MRI (DW-MRI). This toolset provides 3-dimensional high-resolution quantitative maps directly associated with tumor physiology and pathology features. In Aim 2, based on the QMRI output, we will develop an automated PCa tumor probability model and segmentation method that can automatically differentiate areas

with clinically significant prostate cancer from other prostatic tissues. Aim 3 will be a validation study. We will validate the performance of our newly developed method against biopsy results and compare it with the performance of physician's contour of csT foci based on mp-MRI and PI-RADS as the current standard of care

(SOC). We expect that the results of this project will provide a completely new set of QMRI tools and tumor segmentation method that improves current SOC mp-MRI and PI-RADS. This will allow us to accurately identify areas with clinically significant PCa in the intact prostate for boost irradiation, and achieve maximal

local control in prostate cancer radiation therapy without increasing radiation toxicity.