Overall, hepatic resection appears to be an important means of curing patients with metastatic colorectal cancer isolated to the liver. The only absolute contraindication to surgery was the impossibility of a radical removal of tumor: if residual disease will remain after the hepatic resection, this operation is not indicated. A possible second contraindication to surgery is the presence of tumor in the hepatic or celiac lymph nodes. Such metastases from liver metastases signal a biologic grade of tumor that is almost sure to spread to other sites. However, one patient of the 25 in this group did survive long term when positive lymph node groups were dissected. Further clinical experience with this form of the disease along with trials of regional adjuvant therapies such as intraperitoneal chemotherapy may be needed. The presence of extrahepatic metastases at the time of liver resection should be considered a relative contraindication to this surgery, but if the patient can be made clinically disease free, long-term disease-free survival may result. It seems imperative that all patients with hepatic metastases be evaluated by an experienced hepatic surgeon for a curative resection. If the patient has between one and four metastases, a 25 per cent long-term disease-free survival rate can be expected. Patients who have a radical resection of more than four metastases should be considered to be in an experimental group in whom more data are needed. In our current state of knowledge, making such patients clinically disease free is their only chance for long-term survival. Other factors besides the number of metastases that will affect the prognosis of the patient include the disease-free interval between colorectal resection and liver resection, the pathologic margin of resection on the liver specimen, and the presence or absence of mesenteric lymph node metastases from the primary cancer. These factors should be considered when determining the prognosis in a given patient and should be used as stratification variables in prospective trials. However, from our analysis of available data, these factors should not be considered contraindications to hepatic resection.