Perfusion CT for liver radioembolization

/ Abstract Liver cancer remains a huge problem in the US and globally, often diagnosed at advanced stages with limited options. This problem is only going to get worse, with diseases such as non-alcoholic fatty liver disease dramatically rising (an alarming 20 million cases is expected in the next 15 years) and leading to cancer. Responding to this need will require establishing new treatments and existing therapies to deploy them a large scale. Transarterial embolization, also called selective internal radiation therapy (SIRT), is a catheter- based therapy for liver cancer with radioactive yttrium-90 (90Y) microspheres. By injecting the microspheres in the liver arterial blood flow, radiation can be directed at tumors, but the targeting is highly patient-dependent and difficult to plan with current imaging techniques. Tumor targeting is established with a day-long procedure combining exploration of the liver vasculature and nuclear imaging to evaluate the risk of microsphere leakage to the lungs (characterized by the lung shunt fraction). This “workup” uses cone-beam CT for vascular imaging and single photon emission computed tomography (SPECT) to image the 99mTc macro- aggregated albumin (MAA) distribution, thus shuttling the patient back and forth between the interventional radiology suite and the nuclear medicine ward. Done 1-2 weeks before treatment, the workup is as long and complex as the treatment itself and represents a huge burden for the patient and healthcare system due its risk, extent, and high cost. Its efficacy to optimize targeting and predict the lung shunt fraction remains limited. We hypothesize that contrast-enhanced perfusion CT can be an alternative to the workup to map the vasculature, predict the lung shunt fraction (LSF), and provide the additional benefit of predicting the 90Y dose distribution. This hypothesis is based on our preliminary results obtained with 4-phase abdominal CT. To test the validity of this approach, three main goals will be evaluated: 1) Perfusion CT allows to visualize the liver arteries in order to select potential 90Y injection sites, 2) perfusion CT can predict the LSF with comparable accuracy as 99m Tc MAA SPECT and 3) novel computational methods based on perfusion can predict the 90Y dose distribution and be used to optimize the injection dosage and locations. As a reference, we will use 90Y positron emission tomography (PET) post treatment to measure the radiation dose distribution and LSF. We will also collect standard-of-care pretreatment 99m Tc MAA SPECT. We are developing a new tensor-based approach that analyzes the blood flow velocity from dynamic perfusion CT to predicts the dose distribution. If successful, treatment planning could be accurately done from an outpatient and non-invasive perfusion CT instead of a day-long workup in interventional radiology and nuclear medicine for some patients. This new paradigm could unlock the use of SIRT in many patients too sick to undergo the workup. While focused on SIRT, these perfusion CT methods could be applied to other embolization techniques and other pathologies for which blood flow plays an important role and for which predictive models remain insufficient. Project Number: 1R21CA300717-01A1 | Fiscal Year: 2026 | NIH Institute/Center: National Cancer Institute (NCI) | Principal Investigator: Emilie Roncali (+2 co-PIs) | Institution: UNIVERSITY OF CALIFORNIA AT DAVIS, DAVIS, CA | Award Amount: $409,141 | Activity Code: R21 | Study Section: Special Emphasis Panel[ZRG1 ISB-V (80)] View on NIH RePORTER: https://reporter.nih.gov/project-details/11258646