Validation of post-treatment PET-based dosimetry software for hepatic radioembolization of Yttrium-90 microspheres.
Maughan N, Garcia-Ramirez J, Arpidone M, Swallen A, Laforest R, Goddu S, Parikh PJ, Zoberi J. Validation of post-treatment PET-based dosimetry software for hepatic radioembolization of Yttrium-90 microspheres.. Medical physics 2019; .
PURPOSE: Yttrium-90 ((90) Y) microsphere radioembolization enables selective internal radiotherapy for hepatic malignancies. Currently, there is no standard post-delivery imaging and dosimetry of the microsphere distribution to verify treatment. Recent studies have reported utilizing the small positron yield of (90) Y (32 ppm) with PET to perform treatment verification and dosimetry analysis. In this study, we validated a commercial dosimetry software, MIM SurePlan() LiverY90 (MIM Software Inc., Cleveland, OH), for clinical use. METHODS: A MATLAB-based algorithm for (90) Y 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 0.4cm/5% gamma test was performed on PET-derived dose maps from each algorithm of uniform digital and non-uniform physical phantoms filled with (90) Y chloride solution. Average dose (Davg ) and minimum dose to 70% (D70) of a given volume of interest (VOI) were compared for the digital phantom, the physical phantom, and 5 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 Davg and D70 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 Davg , and 88.9% of the VOIs had a difference <5% for D70. CONCLUSIONS: 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 (90) Y radioembolization. This article is protected by copyright. All rights reserved.
ePub ahead of print