Abstract
1308
Objectives: To compare the yttrium-90 (Y-90) dose distribution in the liver predicted with a custom algorithm (CFDose) to the dose distribution imaged with Y-90 PET post radioembolization. Transarterial radioembolization (TARE) is a loco-regional radionuclide therapy based on the delivery of radioactive yttrium-90 microspheres to liver tumors. Accurate pretreatment dosimetry is necessary to determine the Y-90 activity to inject to optimize the dose to the tumor while sparing the rest of the liver. We have developed a patient-specific approach (CFDose) combining computational fluid dynamics (CFD) simulation and Y-90 microsphere decay physics to estimate the volumetric dose distribution in the liver. This study aims at validating CFDose against Y-90 positron emission tomography (PET) based dosimetry for individual patients.
Methods: CFDose consists of 3 steps: the hepatic arterial tree is segmented for each patient; it is then used to simulate the microsphere distribution with computational fluid dynamics (CFD); the predicted microsphere distribution is convolved with a Y-90 dose point kernel to compute the absorbed dose distribution (Figure 1). The absorbed dose is reported for each liver segment affected by the injection and compared to the dose reported by the physician based on the MIRD formalism [1]. The Y-90 distribution was imaged for each patient with PET/CT 2 hours after radioembolization, using 15-minute scans and two bed positions (GE Discovery 690, GE Healthcare). The reconstruction parameters were chosen based on a prior phantom study using the NEMA 2007 phantom filled with 4.5 GBq of Y-90 solution and a sphere-to-background ratio of 7.4. The reconstruction was performed with 3D TOF OSEM with 24 subsets, 2 iterations, and a voxel size of 3.645 mm, and CT-based attenuation correction. PET/CT images were read by a board-certified radiologist. Here we present the qualitative comparison of the predicted liver dose distribution (from CFDose) with the Y-90 activity imaged with PET/CT and the next steps to perform a quantitative comparison of the dose.
Results: The patient presented in Figure 1-2 had a hepatocellular carcinoma between liver segments 7 and 8 and received 2 injections of glass microspheres (TheraSphere®, BTG) in the right hepatic arterial tree. The microsphere distribution predicted by CFDose spread across the tumor, segments 5, 6, 7, and 8 (49%, 11%, 7%, 11%, 22%, respectively) as shown in Figure 1, right. CFDose predicted a total dose of 125 Gy, while the MIRD [1] reported a total dose of 137 Gy. This 9% difference is expected since the MIRD equation for microspheres does not account for their heterogeneous distribution in the liver. The axial, coronal, and sagittal views of the PET/CT acquired 2 hours after treatment show the activity primarily in segments 7 and 8 and some activity in segments 5 and 6 (Figure 2). Conclusion: For one patient, we present a qualitative comparison of the Y-90 liver dose distribution predicted using CFD combined with a Y-90 dose point kernel and the Y-90 activity distribution in the liver imaged with PET/CT after radioembolization. Next steps for all patients currently enrolled in the study (n=5) include i) quantification of the Y-90 activity in the tumor volume and different liver segments using the calibration factor obtained from the phantom scans, ii) computation of the dose distribution based on the quantitative Y-90 PET, iii) registration of the two dose distribution maps based on CT, and iv) comparison of the volumetric distributions in the liver volume. Research Support: NIH R21 CA237686 (ITCR) and CCSG P30 (NCI P30CA093373).