PURPOSE: To evaluate patterns of recurrence and factors which influence them in radiofreqency (RF) ablation for the treatment of hepatocellular carcinoma (HCC). MATERIALS AND METHODS: Between May 1999 and March 2000, 69 patients with 82 HCCs underwent RF ablation for complete necrosis. They were diagnosed by tissue biopsy or tumor marker, and the results of triphasic spiral CT. The indications were that nodular lesions were clearly visualized at sonography, less than 5 cm in size and less than four in number, and that patients had no history of previous treatment. Local therapeutic efficacy such as complete necrosis and marginal recurrence, and new lesions were evaluated by means of triphasic spiral CT performed at least six months after the completion of ablation. We then analyzed the correlation between local therapeutic efficacy and various influential factors such as tumor size, whether the tumor was attached to the portal vein, gross morphology, Child-Pugh classification, and alpha- fetoprotein level before the procedure, as well as the correlation between new lesions and influential factors which included the alpha-fetoprotein level before the procedure, Child-Pugh classification, and multiplicity per person. RESULTS: During a mean follow-up period of 8.95 (range, 6-14) months after RF ablation, the rate of complete necrosis and of marginal recurrence was 91% and 12%, respectively. When a tumor was larger and was attached to a large branch of the portal vien, the incidence of incomplete necrosis and marginal recurrence was greater. The occurrence rate of new lesion was 19.4%. When the alpha-fetoprotien level before the procedure was higher and a tumor was multiple in number, new lesions occurred more frequently. CONCLUSION: Sufficient knowledge of patterns of recurrence and the factors which influence them might improve the therapeutic effects of RF ablation in patients with HCC.