Glass versus resin microspheres in ⁹⁰Y radioembolization for hepatocellular carcinoma: toxicity patterns under uniform voxel-based dosimetry.

Abstract

Objectives: To compare safety outcomes and toxicity patterns of glass and resin microspheres in patients undergoing yttrium-90 (⁹⁰Y) transarterial radioembolization (TARE) for hepatocellular carcinoma (HCC), using uniform voxel-based dosimetry for both platforms and toxicity comparisons adjusted for treatment extent and tumor burden.

Materials and methods: This retrospective single-center study included 103 consecutive patients with HCC and preserved liver function treated with ⁹⁰Y-TARE between June 2020 and June 2024 using glass-based (n = 70) or resin-based (n = 33) microspheres, selected according to routine practice and product availability. Absorbed doses for both cohorts were derived from voxel-based ⁹⁹ᵐTc-MAA SPECT/CT dosimetry. Adverse events (AEs) were graded with CTCAE version 5.0. The primary endpoint was overall AE incidence; secondary endpoints were specific toxicity patterns and dose-toxicity associations, examined with multivariable logistic regression adjusting for treatment extent and tumor volume. Progression-free survival (PFS) and overall survival (OS) were compared with the log-rank test.

Results: Overall AE incidence did not differ between glass and resin microspheres (glass 37.1% vs. resin 30.3%; p = 0.497), and all AEs were grade 1-2. Tumor and non-tumorous liver absorbed doses were similar. Bilirubin toxicity was more frequent with glass microspheres (glass 28.6% vs. resin 9.1%; p = 0.027), whereas hemoglobin decline was more common with resin microspheres (glass 10.0% vs. resin 30.3%; p = 0.010); both exploratory differences persisted after adjustment for treatment extent and tumor volume. Lung shunt fraction, a baseline patient-level parameter, was higher with resin (glass 2.6% vs. resin 4.1%; p = 0.043). Higher lung absorbed dose was associated with AE occurrence overall, without a significant lung dose × microsphere-type interaction (p = 0.41). Neither PFS (median, glass 9 vs. resin 12 months; p = 0.868) nor OS (glass 23 vs. resin 29 months; p = 0.929) differed significantly.

Conclusions: Glass and resin microspheres showed no significant difference in overall AE incidence-without formally establishing equivalence-but differed in their patterns of mild laboratory toxicity independently of treatment extent and tumor burden. These exploratory findings support careful, individualized dosimetric assessment and require confirmation in prospective multicenter studies, but do not provide criteria for microsphere selection.