A pilot study was designed to evaluate the efficacy of high-dose FUDR administered through the hepatic artery for the treatment of cancer involving the liver. Three dose schedules were used beginning with a dose of 0.5 mg FUDR/kg/day for 2 weeks followed by normal saline infusion for 2 weeks (schedule A). Elevation of serum bilirubin was the sole indication to deescalate to schedule B (0.3 mg FUDR/kg/day for two weeks followed by saline infusion for 4 weeks). Tolerance to this schedule escalated the patient to schedule C (0.5 mg FUDR/kg/day for 2 weeks followed by normal saline infusion for 4 weeks). Eighteen patients were treated, sixteen with metastatic colon cancer, one with metastatic leiomyosarcoma, and one with hepatoma. The patient with hepatoma developed progressive disease after one cycle of therapy. Of the 17 patients with metastatic cancer only 5 patients failed therapy yielding a 70% response rate. High-dose FUDR was well tolerated with only six patients requiring deescalation to schedule B. Elevation of alkaline phosphatase and glutamic oxaloacetic transaminase was universal. Two patients developed peptic ulceration. Sclerosing cholangitis was not observed. We conclude that high-dose FUDR administered through the hepatic artery is as safe as conventional dose infusion therapy but probably not more effective. The safety of high-dose FUDR infusion therapy suggests that sclerosing cholangitis is association with hepatic arterial infusion therapy is not related to the FUDR dose.