Validation of post-treatment PET-based dosimetry software for hepatic radioembolization of Yttrium-90 microspheres

Purpose Yttrium-90 (Y-90) microsphere radioembolization enables selective internal radiotherapy for hepatic malignancies. Currently, there is no standard postdelivery imaging and dosimetry of the microsphere distribution to verify treatment. Recent studies have reported utilizing the small positron yield of Y-90 (32 ppm) with positron emission tomography (PET) to perform treatment verification and dosimetry analysis. In this study, we validated a commercial dosimetry software, MIM SurePlan (TM) LiverY90 (MIM Software Inc., Cleveland, OH), for clinical use. Methods A MATLAB-based algorithm for Y-90 PET-based dosimetry was developed in-house and validated for the purpose of commissioning the commercial software. The algorithm is based on voxel S values and dosimetry formalism reported in MIRD Pamphlet 17. We validated the in-house algorithm to establish it as the ground truth by comparing results from a digital point phantom and a digital uniform cylinder to manual calculations. Once we validated our in-house MATLAB-based algorithm, we used it to perform acceptance testing and commissioning of the commercial dosimetry software, MIM SurePlan, which uses the same dosimetry formalism. A 0.4 cm/5% gamma test was performed on PET-derived dose maps from each algorithm of uniform digital and nonuniform physical phantoms filled with Y-90 chloride solution. Average dose (D-avg) and minimum dose to 70% (D-70) of a given volume of interest (VOI) were compared for the digital phantom, the physical phantom, and five patient cases (27 tumor VOIs), representing different clinical scenarios. Results The gamma-pass rates were 97.26% and 97.66% for the digital and physical phantoms, respectively. The differences between D-avg and D-70 were 0.076% and 0.10% for the digital phantom, respectively, and <5.2% for various VOIs in the physical phantom. In the clinical cases, 96.3% of the VOIs had a difference <5% for D-avg, and 88.9% of the VOIs had a difference Dose calculation results from MIM SurePlan were found to be in good agreement with our in-house algorithm. This indicates that MIM SurePlan performs as it should and, hence, can be deemed accepted and commissioned for clinical use for post-implant PET-based dosimetry of Y-90 radioembolization. (c) 2019 American Association of Physicists in Medicine