PURPOSE: To report clinical effectiveness, toxicity profile, and prognostic factors of combined 100 µm ± 25 and 200 µm ± 50 epirubicin-loaded polyethylene glycol (PEG) microsphere drug-eluting embolic transcatheter arterial chemoembolization protocol in patients with hepatocellular carcinoma. MATERIALS AND METHODS: In this prospective, single-center, single-arm study with 18 months of follow-up, 36 consecutive patients (mean age 69.9 y ± 10.8; 26 men, 10 women; 54 naïve lesions) were treated. Embolization was initiated with 100 µm ± 25 microspheres, and if stasis (10 heart beats) was not achieved, 200 µm ± 50 microspheres were administered. Each syringe (2 mL) of PEG microsphere was loaded with 50 mg of epirubicin. Results were evaluated using Modified Response Evaluation Criteria In Solid Tumors with multidetector computed tomography/magnetic resonance imaging at 1, 3-6, 9-12, and 15-18 months. Toxicity profile was assessed by laboratory testing before and after the procedure. Complications were recorded. Postembolization syndrome (PES) was defined as onset of fever/nausea/pain after the procedure. Patient/lesion characteristics and treatment results were correlated with predicted outcome using regression analysis. Child-Pugh score was A in 86.1% of patients (31/36) and B in 13.9% (5/36). RESULTS: In 10 of 21 lesions, < 2 cm in diameter (47.5%) stasis was achieved with 100 µm ± 25 microspheres only, whereas all other lesions required adjunctive treatment with 200 µm ± 50 microspheres. Reported adverse events were grade 1 acute liver bile duct injury (3/39 cases, 7.7%) and PES (grade 2; 3/39 cases, 7.7%). Complete response (CR) at 1, 3-6, 9-12, and 15-18 months was 61.1%, 65.5%, 63.63%, and 62.5%. Objective response (CR + partial response) at 1, 3-6, 9-12, and 15-18 months was 83.3%, 65.85%, 63.63%, and 62.5%. No single factor (laboratory testing, etiology, patient status, hepatic status, tumor characteristics, administration protocol) predicted outcomes except for albumin level at baseline for CR (P < .05, odds ratio = 1.09). CONCLUSIONS: The combined microsphere sizing strategy was technically feasible and yielded promising results in terms of effectiveness and toxicity.