The optimum conditions (duration and concentration) of a fixed dose, intra-arterial melphalan infusion in relation to its antitumor effect and toxicity in the liver were investigated in a rat colon tumor model (CC531) of liver metastases. We studied the difference in tumor and liver uptake, as well as antitumor effect and hepatotoxicity after 5- and 20-min hepatic arterial infusion (HAI) of a fixed melphalan dose. Melphalan content in perfusate, liver, and tumor tissue was analyzed by high-performance liquid chromatography. The antitumor effect and hepatotoxicity in rats treated either systemically or with 5- and 20-min HAI, with a fixed dose melphalan (4.4 μmol), were assessed 2 weeks after treatment. No difference in melphalan content of tumor/ liver tissue or antitumor effect was observed between rats treated with 5- and 20-min HAI. Hepatotoxicity was strongly affected by perfusion duration/concentration, however. Rats treated with 5-min HAI weighed significantly less, and liver toxicity parameters were significantly increased compared with those of all other groups; eight of nine rats showed severe cholangiofibrosis. Body weights and liver toxicity parameters of the rats treated with 20-min HAI were not statistically different from the control group. In conclusion, duration of HAI with 4.4 μmol of fixed dose melphalan did not affect tumor uptake and antitumor effect, but the resulting increase in melphalan concentration had major impact on hepatobiliary toxicity. Therefore, in a clinical setting, caution should be taken when infusion duration and concentration of melphalan are changed.