Transarterial radioembolization (TARE) with Yttrium-90 microspheres is an established locoregional treatment for hepatocellular carcinoma, but radiation-induced biliary complications remain underreported. This study aimed to determine the prevalence, temporal patterns, and predictors of TARE-related cholestasis, and to provide histological proof of our hypothesis that TARE can cause cholestasis through direct radiation injury to the bile ducts.

We retrospectively analyzed all patients who underwent TARE at a tertiary care center between January 2020 and December 2024. Demographics, tumor type, procedural details (unilobar, bilobar, segmental administration), applied radiation dose, follow-up imaging, and endoscopic findings were reviewed. TARE-related cholestasis was defined as new-onset cholestasis without preexisting or post-transplant occurrence. Risk factors were evaluated using logistic regression and Cox proportional hazards models. Patients who developed biliary strictures underwent ERCP with stent placement and/or balloon dilatation as indicated. In patients with unclear strictures post-TARE, cholangioscopy with targeted biopsies was performed during ERCP to histologically characterize radiation-associated tissue changes.

Among 258 patients (29.8% female; median age 67±12.0 years) undergoing 331 TARE procedures, 19 underwent subsequent ERCP. After excluding alternative causes (n=7) and parenchymal liver damage (n=1), 11 patients had confirmed clinically significant TARE-related biliary strictures. The median interval from TARE to stricture was 19.4 months (applied median activity 2.28 GBq, median dose 121.5 Gy). On cross-sectional imaging, stricture patterns included lobar (52.7%), central (27.3%), and segmental (20.0%) distributions, all localized to irradiated liver segments. Overall, radiologic cholestasis developed in 55 of 295 evaluable procedures (18.6%). Independent predictors of imaging-detected cholestasis included increasing age (HR 1.04/year, p=0.003), cholangiocarcinoma versus HCC (HR 5.66, p<0.001), and bilobar treatment (HR 2.36, p=0.036). Median imaging-based cholestasis-free survival was 829 days. All 11 patients with clinically significant strictures underwent endoscopic stent placement; 6 received additional balloon dilatation. Notably, 10 of 11 patients showed stable disease or tumor response at ERCP, indicating stricture development independent of progression. Critically, cholangioscopy-guided biopsies demonstrated glass microspheres within the stricture tissue, providing direct histological proof that TARE causes cholestasis through radiation-induced bile duct injury.

Radiologically detectable cholestasis occurs in 18.6% of TARE procedures as a delayed complication. The histological demonstration of glass microspheres within biliary strictures confirms our hypothesis of direct radiation-induced bile duct injury. ERCP with stenting remains the primary therapeutic approach for clinically significant strictures. Cholangioscopy with targeted biopsy should be considered for definitive diagnosis when distinguishing radiation-induced strictures from malignant obstruction.