Background. The effects of preoperative transcatheter arterial chemoembolization (TAE) for hepatocellular carcinoma (HCC) remain a matter of controversy. Methods. Seventy-two patients with HCC were entered in the study; the patients did not have the risk factors for disease recurrence of tumor larger than 5 cm in diameter, the presence of venous invasion, or intrahepatic metastasis. Only patients with 3 years of follow-up after curative resection were selected. Forty-six underwent TAE (Group I) and 26 did not undergo TAE (Group II). Group I was divided into three subgroups according to the degree of tumor necrosis: IA, complete necrosis; IB, partial necrosis; and IC, no necrosis. Group II was divided into two subgroups: IIB, partial necrosis; and IIC, no necrosis. Results. Preoperative TAE did not improve the average disease-free survival rates of the group as a whole. For patients undergoing TAE, the survival rate of Group IB was significantly worse than that of Groups IA or IC. The survival rate of Group IB was worse than that of Group II, but the difference was not significant. In Group II, the survival of Group IIB was worse than that of Group IIC. Histologically, residual tumor cells lacking mutual contact were detected in some patients in Group IB. Conclusion. These results indicate that partial tumor necrosis caused by preoperative TAE or spontaneous tumor necrosis per se might facilitate postoperative disease recurrence. This may occur because in patients with partial necrosis, the remaining tumor cells are less firmly attached and more likely to be dislodged into the bloodstream during hepatic resection.