Pilot Project 1

Johns Hopkins University, Baltimore MD

Investigators

Abstract

Prostate cancer is a leading cause of cancer-related death in men. Novel technologies can improve access to high-quality care—particularly for patients at increased risk of aggressive disease or living in areas with a high prostate cancer burden. Current diagnostic methods, including PSA testing and biopsy, are often imprecise, leading to underdiagnosis and misclassification of disease severity. To address this, we propose a novel, non-invasive imaging modality using molecular chemical exchange saturation transfer (CEST) MRI on standard 3T scanners, combined with a machine learning-driven metabolic imaging processing pipeline. This technology will enable more accurate and cost-efficient detection and prognosis of prostate cancer aggressiveness. We aim to improve spatial resolution in molecular MRI using deep learning super-resolution models trained on simulated high-resolution metabolic images. These simulations will be derived from organ-scale vascular networks and metabolic conversion rates, allowing for subject-specific modeling of energy metabolism. Our specific aims include: 1. Mapping energy metabolism in aggressive prostate tumors using dynamic glucose-enhanced CEST MRI, 2. Developing a machine learning-based super-resolution model using subject-specific metabolic simulations, and 3. Validating patient-specific metabolic and perfusion models using clinical MRA and anatomical data. This pilot project contributes to the scientific priorities of H-H U54 by developing novel, scalable imaging technologies to improve prostate cancer detection and management, while also supporting research capacity in advanced molecular imaging and computational modeling across the partner institutions.