Dr. Jason Chiang and Dr. Kyung Sung of the Department of Radiological Sciences at the David Geffen School of Medicine at UCLA and the UCLA Health Jonsson Comprehensive Cancer Center have received a $3.2 million, five-year grant from the National Cancer Institute to develop an artificial intelligence (AI)-enhanced imaging platform designed to improve yttrium-90 (Y90) radioembolization planning for patients with liver cancer.
Y90 radioembolization is a minimally invasive treatment for liver cancer that delivers millions of tiny radioactive beads directly into tumors through the liver’s blood vessels. The beads, called microspheres, become trapped inside the tumor and release radiation that kills cancer cells while limiting damage to healthy liver tissue. This minimally invasive, image-guided cancer treatment is increasingly being used for cancers that start in the liver as well as cancers that have spread there from other parts of the body. For the therapy to work effectively, doctors must carefully control not only the radioactive dose, but also how many beads reach the tumor. Too few may not deliver enough radiation to destroy the cancer, while too many can block blood flow or redistribute to nearby healthy tissue.
Before treatment, doctors use imaging scans to estimate where the radioactive beads will travel inside the liver and tumor. However, current methods cannot fully capture the complex and often unpredictable blood flow within tumors, making it difficult to accurately predict how the beads will distribute during treatment. To address this challenge, Chiang, a physician-scientist who is also a member of the the UCLA Broad Stem Cell Research Center, and Sung, an expert in artificial intelligence and magnetic resonance imaging (MRI), are leading the research team to develop an AI-enhanced imaging approach that uses special dynamic contrast-enhanced (DCE)-MRI scans to better characterize tumor blood flow and predict how Y90 microspheres distribute within liver tumors.
The new NIH funding will support the development and validation of the AI platform to optimize Y90 radioembolization treatment planning. Using specialized hepatic vascular phantoms, the team will first evaluate how Y90 microsphere distribution is affected by arterial flow pattern, catheter position and tumor vascularity. These phantom models will allow investigators to train and validate computational tools linking perfusion patterns derived from DCE-MRI scans to the Y90 microsphere density. The validation process will then be extended to large animal liver tumor models using clinically relevant MRI scanners and imaging protocols.
“By combining advanced MRI techniques with artificial intelligence, we hope to better predict how Y-90 microspheres distribute within liver tumors and improve the precision of radioembolization therapy,” said Chiang. “This work has the potential to optimize treatment effectiveness, reduce unintended radiation exposure, and advance personalized care for patients with liver cancer.”
