PURPOSE: The aim of this study was to investigate the relationship between voxel-based dosimetric variables derived from Y-90 PET/MRI and hypertrophy observed in the left lobe after radioembolization and to investigate if there is any difference in hypertrophy induced by glass versus resin microspheres. METHODS: Voxel-based dosimetry-derived variables and their relationship with the change of the standardized future liver remnant (ΔFLR) was investigated with linear regression models. To compare and evaluate the discriminatory power of the dosimetric variables, ROC analyses were utilized. ΔFLR and kinetic growth rate (KGR) induced with glass and resin microspheres were compared using the Mann-Whitney U test. RESULTS: In this retrospective study, data of the 40 patients treated with Y-90 microspheres were evaluated. Among the several dosimetric variables, the mean perfused volume normal tissue dose (pDnorm), perfused normal tissue V90 (pV90), and pV100 values for glass microspheres
and the mean whole liver normal tissue dose (Dnorm), pDnorm, whole liver normal tissue V30 (nV30), nV40, and pV40 for resin microspheres had the highest relationship with ΔFLR. In the ROC analysis for glass microspheres, the optimal cut-offs to predict ΔFLR >
5% were 60.55 Gy for Dnorm, 94.21 Gy for pDnorm, 28.07% for pV90, and 24.98% for pV100. For resin microspheres, corresponding values were 23.20 Gy for Dnorm, 37.40 Gy for pDnorm, 31.50% for nV30, 24.50% for nV40, and 43.60% for pV40. No significant difference was observed between glass and resin microsphere-induced median ΔFLR, KGR values and atrophy of the right lobe. CONCLUSION: Following Y-90 radioembolization therapy with glass and resin microspheres applied to the right lobe of the liver, ΔFLR is correlated with pDnorm and Dnorm, but is also significantly related to various nV and pV values. In addition, the hypertrophy and kinetic growth rates observed with glass and resin microspheres were largely similar.