PURPOSE: To investigate the safety and usefulness of Lipiodol-percutaneous transhepatic ethanolinjection(L-PEI) and to determine the appropriate concentration of Lipiodol during L-PEI. This was achieved byevalvating CT findings and histopathologic changes according to the concentration of Lipiodol, amount of ethanol,and the time interval after injection into normal rabbit liver. MATERIALS AND METHODS: This experimental studyinvolved 18 New Zealand rabbits under US guidance. They were divided into five groups according to injectedmaterials; two rabbits with 0.4cc of normal saline(group I), six with 0.4cc of ethanol in the left hepaticlobe(group II), and 0.4cc of Lipiodol in the right hepatic lobe(group III), five rabbits with 5%Lipiodol-ethanol(5% vol. of Lipiodol+95% vol. of ethanol), 0.2cc in the right hepatic lobe, and 0.4cc in theleft(group IV); and five rabbits with 10% Lipiodol-ethanol as per group IV(group V). CT was performed immediately,one week, two weeks, and three-four weeks after injection, and pathologic specimens were obtained on the thirdday(acute phase) and during the third or fourth week(chronic phase) after injection. RESULTS: On CT, intrahepaticlocalization of the L-PEI injection site was well demonstrated as a focal high attenuated area which graduallydecreased in attenuation on follow up CT. The opacification of the inferior vena cava by Lipiodol, the lineardistribution of Lipiodol along portal veins or fissures, and peritoneal leakage were clearly demonstrated ingroups III-V, though the effects gradually disappeared during follow-up CT. There was no remarkable difference ingross CT attenuation between group IV and group V. The main pathologic findings during the acute phase of group IIwere coagulation necrosis surrounded by macrophage, inflammatory reaction, and early periportal and subcapsularfibrosis. The findings in group IV and V were similar to those in group II and additional fat vacuoleaccumulations in the necrotic area were also seen. During the chronic phase of group II, areas of necrosis wereabsent or smaller and were surrounded or replaced by more organized fibrosis, macrophage or multinucleated giantcell infiltration. Periportal, subcapsular fibrosis was also found. In group IV and V, the findings were similarto those of group II, though additional fat vacuoles in fibrotic or necrotic areas, foreign body reaction to fatvacuole, regenerating nodule and calcification were also observed. CONCLUSION: L-PEI is more useful for thedetection by CT of an injection site than PEI alone, and with regard to CT and histopathologic findings, there wasno significant difference between the 5% and 10% Lipiodol-ethanol groups. Compared to PEI, L-PEI provoked nosighificant additional hepatic injury; only fatty change and foreign body reaction were noted. Thus, L-PEI is moreuseful than PEI for the management of HCC.