Ethanol injection has been for many years the standard approach as percutaneous treatment of non-surgical early hepatocellular carcinoma in cirrhotic patients. Radiofrequency ablation has been proposed as an alternative approach and some retrospective and prospective comparative trials indicated its better performance for the local control of the disease in comparison to ethanol injection. Some prospective randomized studies from Japan and from Taiwan reported also an increase of overall survival in the patients treated by Radiofrequency. Aim of this study is to compare the two techniques in a prospective randomized trial in western patients.

Currently early hepatocellular carcinoma (HCC) complicating liver cirrhosis is often treated by percutaneous ablation techniques under ultrasound guidance: their performance is easy, safe and efficient. They are less invasive and expensive than surgical removal, representing a potential local cure for a larger number of patients than those who could be treated by a liver resection or offered a liver transplantation.

Percutaneous ethanol injection (PEI), the first non-surgical technique introduced in clinical practice, has been widely used since the late eighties. PEI induces a chemical and ischemic coagulative necrosis in 70-80% of small (<3cm) HCC lesions. Necrosis obtained by this way is generally restricted to the neoplastic lesion itself; the chemical ablation may be moreover hampered by internal scars, that limit the uniform spreading of ethanol and by possible uncontrolled flows outside the lesion, with the ultimate result of a reduced necrotic effect and persistence of residual vital nests of neoplasia in as many as 33% of the cases. Nevertheless, the survival after PEI seems not different from surgery in retrospective series and this technique was indicated as the standard percutaneous treatment for early HCC in the 2001 guidelines of the European Association for the Study of the Liver (EASL).

In 1993 radio-frequency ablation (RF) was proposed as a new technique for the ablation of small HCCs. RF induces heat-generated coagulative necrosis of the neoplastic lesion and surrounding liver tissue. However, RF is not suitable for lesions situated close to large vessels or hollow viscera, as the first can decrease heat generation and the latter can be damaged by the procedure itself; it is more cumbersome, needs anaesthesiology assistance in most of the cases and, accordingly to early reports, is aggravated by a higher complication rate and higher costs. A major advantage of RF is the achievement of complete ablation of the neoplastic lesion with less sessions than PEI. This advantage impacts on the "quality of life" of the patient, who prefers to be cured in a single session rather than by the multisession approach of PEI.

When the present study began in January 2001, no randomised controlled trial comparing RF and PEI had been published and no data were available on mid-long term local tumor progression and on comparative survival rate after either treatment. Some retrospective studies about the primary effectiveness had been published.

The aim of our study was to compare the effectiveness of RF versus PEI in a randomised controlled trial: our primary end point was to evaluate the local control (sustained complete response) after 1 year from the treatment of every lesion defined as HCC in the single patient at the beginning of the trial; secondary end points included the primary effectiveness (early complete response)of the treatment after 1-2 months,the overall survival at four years, the complication rate, and the costs.

Inclusion Criteria:

• Included were cirrhotic patients in Child-Pugh class A/B with 1 to 3 HCC nodes of ≤ 30mm in diameter.

Exclusion Criteria:

• Excluded were patients without liver cirrhosis, in Child-Pugh class C, with a platelets count <40000, INR >1.75, PTT >40sec, hypo vascular HCC, lesions not detectable by ultrasonography (US), lesions close (≤1cm) to the gallbladder, hepatic hilum, colon or stomach, with venous invasion or metastatic disease