Prostate Cancer
Prostate cancer remains one of the most prevalent cancers among men worldwide. While many of these cases are considered low-risk, a subset of patients develop clinically significant prostate cancer (csPCa)—a more aggressive form that, if left untreated, can lead to serious health complications or death. Timely and accurate detection is therefore essential for improving outcomes in these patients.
MRI in Prostate Cancer
Over the past two decades, magnetic resonance imaging (MRI) has become an important tool in the diagnosis and management of csPCa. MRI is now widely recommended as a pre-biopsy diagnostic tool, offering information that helps guide clinical decisions. However, the effectiveness of mpMRI is heavily dependent on expert interpretation. Inter-reader variability and a considerable rate of false positives continue to pose challenges, while the growing demand for prostate imaging continues to burden healthcare.
The Role of Artificial Intelligence
Artificial intelligence (AI) has demonstrated high potential in medical imaging, with successful applications in detecting cancers of the lung, breast, and other organs. In the context of prostate cancer, AI offers a compelling opportunity to improve image interpretation, reduce diagnostic variability, and increase efficiency. However, despite its promise, the clinical integration of AI for csPCa diagnosis is still in its early stages, largely due to a lack of robust evidence.
The MAGIC group collaborates closely with the Diagnostic Image Analysis Group (DIAG) to pioneer research at the intersection of AI and prostate cancer imaging. A key focus is the development of high-quality, annotated imaging datasets with carefully curated reference standards. These datasets serve as the foundation for training and benchmarking state-of-the-art AI models.
Beyond model development, the group explores how AI can be potentially integrated into clinical workflows. This includes assessing both assistive tools that support radiologists in decision-making, as well as semi-autonomous systems that may one day operate independently under clinician supervision. The ultimate goal is to improve diagnostic accuracy, reduce workload, and improve access to high-quality prostate cancer care.
