With increasing use of radiation for hepatocellular carcinoma (HCC) through transarterial radioembolization (TARE) and stereotactic body radiation therapy (SBRT), there is concern for increased radiation-related complications when using SBRT after TARE. This study compares safety of SBRT after segmental TARE versus transarterial chemoembolization (TACE).

A retrospective review identified patients receiving SBRT after TACE or TARE for HCC from 2011 to 2017. TARE was delivered subselectively to individual segments using yttrium-90 with Theraspheres. Patients were assessed over time for Child-Turcott-Pugh (CTP)/albumin-bilirubin (ALBI) scores, and Common Terminology Criteria for Adverse Events version 4.0 grade ≥3 events. Linear mixed models were used to examine the trend of CTP and ALBI over time and compare groups. Secondary endpoints were objective response rate via modified Response Evaluation Criteria in Solid Tumors (RECIST), local control, and overall survival.

Ninety-nine patients met criteria with median follow-up of 9.8 months (range, 0.9-47): 31 had SBRT after segmental TARE and 68 patients post-TACE. The groups were well balanced with regard to etiology of HCC, baseline CTP and ALBI scores, and SBRT dose, but there were significant differences in baseline Eastern Cooperative Oncology Group performance status, Barcelona Clinic Liver Cancer staging, and median follow-up. There was a significant increase in post-SBRT CTP and ALBI scores (P < .0001) for both groups. However, there was no significant difference in rise in CTP (P = .11) or ALBI score (P = .82) over time between SBRT post-TACE versus post-segmental TARE. There was no significant increase in ≥grade 3 toxicity postsegmental TARE. There was also no significant difference in local controls (P = 1.0) and overall survival (P = .26) between cohorts, but objective response rate was worse post-TARE.

SBRT after segmental TARE with Theraspheres appears to have acceptable tolerability and is effective compared with SBRT after TACE. Longer follow-up with larger numbers is needed to verify these data.