Hepatic metastases are a major cause of morbidity and mortality for patients with colorectal cancer (CRC). The rationale for hepatic arterial chemotherapy has both an anatomical and pharmacological basis. Several randomized clinical studies of fluoropyrimidine showed higher response rates in all trials when the drug was given as an hepatic arterial infusion (HAI) versus systemic administration. However, the studies did not accurately define survival for the following reasons: (1) some allowed a crossover; (2) some were too small; and (3) some used inadequate systemic chemotherapy. Patients who have failed to respond to previous systemic chemotherapy have an approximately 50% response rate with HAI treatment. Hepatic toxicity, especially biliary sclerosis, is the dose-limiting toxicity, occurring in 6% to 25% of patients. Adding dexamethasone to HAI fluoropyrimidine decreases the hepatobiliary toxicity. The therapeutic benefit of HAI in one study was also demonstrated by an increased time with normal quality of life. To truly define the role of regional therapy in patients with CRC confined to the liver, the current Cancer and Leukemia Group B (CALGB) study is randomizing patients to HAI versus systemic therapy without a crossover to demonstrate if HAI improves survival and/or quality of life in addition to response rates.