Roland Hustinx : La médecine nucléaire augmentée par l'IA dans les thérapies multimodales du cancer

Saturday, April 25, 2026 at 10:15 a.m. - Public session on the impact of technological advances in cancer therapy Roland Hustinx (ULiège): Nuclear medicine augmented by Artificial Intelligence in multimodal cancer therapies Artificial intelligence contributes routinely and daily to nuclear medicine techniques. AI is integrated, seamlessly for the user, into image reconstruction methods for positron emission tomography (PET scan) and scintigraphy (SPECT). Regarding cancer treatments, which often combine systemic approaches (chemotherapy, targeted therapy, immunotherapy) and radiotherapy using various modalities (external beam radiotherapy, brachytherapy, radioligand therapy), specific tools have been developed using AI to define prognostic indices, contribute to the definition of target volumes, and quantify treatment response. These tools initially relied on relatively simple machine learning techniques, and currently employ deep learning. However, developments are extremely rapid, and a new stage, agentic AI, has already entered the field of radiology and is expected to quickly expand into nuclear medicine. This is a form of AI capable of performing complex and interconnected tasks autonomously and without constant supervision.