Purpose: To compare the stability of epirubicin-iodized oil emulsions prepared with ionic or nonionic contrast medium and to compare the efficacy of these emulsions in a prospective, randomized, controlled trial of transarterial chemoembolization (TACE) for hepatocellular carcinoma (HCC). Materials and Methods: Epirubicin-iodized oil emulsions prepared with ionic and nonionic contrast media was evaluated for stability with light microscopy and magnetic resonance imaging. One hundred ninety-seven patients with inoperable HCC were randomized to receive TACE with epirubicin, prepared either with ionic (control group, n = 99) or nonionic (experimental group, n = 98) contrast medium. Tumor response was graded according to iodized oil retention (grade 1 = >90% retention, grade 2 = 50%-90% retention, and grade 3 = <50% retention), as characterized with computed tomography. Survival probabilities were calculated with the Kaplan-Meier method. Results: The epirubicin-iodized oil emulsions prepared with ionic contrast medium were less stable, exhibiting rapid separation of the oil and aqueous phases, compared with emulsions prepared with nonionic medium. Ninety-one patients in the control group and 87 in the experimental group underwent follow-up CT. Thirty-seven of the 91 patients in the control group (41%) had grade 1 tumors, 41 (45%) had grade 2 tumors, and 13 (14%) had grade 3 tumors. Forty-eight of the 87 patients in the experimental group (55%) had grade 1 tumors, 22 (25%) had grade 2 tumors, and 17 (20%) had grade 3 tumors. The number of patients with grade 1 tumors was significantly higher in the experimental group than in the control group (P = .02); however, there was no difference in patient survival (P = .94). Conclusions: Epirubicin-iodized oil emulsions prepared with nonionic contrast medium are more stable and are associated with lower tumor grade in patients with inoperable HCC. The choice of solvent, however, does not appear to have an effect on patient survival.