Prostate cancer is the most common type of cancer in Norwegian men, but many tumors are slow-growing and do not require treatment. Today, MRI is good at detecting suspicious lesions, yet it cannot reliably distinguish aggressive tumors from low-grade ones. As a result, many men undergo repeated invasive biopsies. New PET tracers targeting PSMA improve tumor localization and may correlate with cancer aggressiveness, offering potential for better assessment.
This project aims to develop a method to predict Gleason Score non-invasively by applying machine learning to PET and MRI data. The work involves early static and dynamic PSMA PET imaging, tracer kinetic modelling, deep learning, and validation of PET-based measurements of PSMA internalization using ex-vivo cellular methods.
If successful, the project could reduce the number of biopsies, improve diagnostic accuracy, offer full 3D assessment of the prostate, shorten clinical workflows, and help identify patients who would benefit most from PSMA-based radioligand therapy.