The use of hepatic arterial infusion (HAI) chemotherapy in patients with liver-only colorectal metastases is based on the pharmacologic principle that the regional administration of some drugs can lead to higher drug concentrations at the site of the metastases and avoid systemic toxicity. Early randomized trials resulted in high response rates but did not lead to a survival advantage with HAI. More recent trials have utilized improved surgical techniques and strict guidelines regarding dose reduction or cessation of HAI chemotherapy, resulting in a significant reduction in toxicity. In patients with unresectable liver metastases, two recent European trials using HAI fluorouracil (5-FU) again failed to demonstrate an improvement in survival, but both were plagued by a high complication rate with an associated high proportion of patients failing to receive the assigned treatment. In contrast, the preliminary results of a recent Cancer and Leukemia Group B trial did demonstrate a survival advantage with HAI floxuridine when compared to systemically administered 5-FU. Trials investigating the use of HAI chemotherapy in the adjuvant setting have yielded mixed results. Moreover, in light of improved response rates and overall survival with newer more active chemotherapeutic and novel agents, the absolute role of HAI chemotherapy remains undefined.