Intratumor ethanol injection was studied in the treatment of small hepatocellular carcinoma (HCC). One of the major drawbacks of this technique remains the lack of objective information about its efficiency and the practical conditions of injection. To ensure accurate evaluation of alcoholization, we developed a model based on the tumor obtained after subcutaneous injection of human hepatoma cell lines into nude mice. Each of three cell lines (Hep G2, Hep 3B, PLC/PRF/5) was tested on 24 mice. A total of 0.1 ml of a solution containing 5 x 10(6) cells was injected under the abdominal skin of a male nude mouse weighing 30 g. The Hep G2 cell line appeared to be the most suitable for the model. It enabled us to obtain tumors of 20 mm in diameter within a mean delay (m +/- SD) of 45 +/- 16 days (range: 29-60) with only a 25% failure rate. No visceral spreading of the carcinoma was noticed and the tumors obtained, similar to human HCC, were convenient to be measured, monitored, and treated by alcoholization. To validate this model for alcoholization, 38 tumors ranging in diameter from 10 to 20 mm were treated using either a unique centrotumoral injection (n = 27) or five cross-shaped injections (n = 11). Intratumor absolute ethanol injection resulted in tumoral necrosis which was easily quantified as a percentage of the tumor volume, using a semiquantitative method. It is concluded that the Hep G2 cell line transplanted in nude mice resulted in a relevant model to assess tumoral destruction by alcoholization.