Careful patient selection for hepatic resection of colorectal cancer metastases is essential to improve current poor results. Carcinoembryonic antigen level and number of metastases were significant preoperative prognostic indicators of 5-year disease-free survival in patients selected clinically for hepatic surgery. Surgical margin, weight of hepatic tissue resected, carcinoembryonic antigen level, and flow cytometry were significant postoperative prognostic indicators. Patients with a carcinoembryonic antigen level less than 200 ng/mL, 1-cm surgical margins, and less than 1,000 g of liver tissue removed had a greater than 50% estimated 5-year disease-free survival rate. If the metastases were diploid on flow cytometry, an additional survival advantage may have been gained. Inadequate surgical margins led to high rates of liver-only recurrence. Nonhepatic recurrence was unrelated to surgical margins. Intraoperative liver examination by ultrasound during primary colon cancer resection and adjuvant chemotherapy may offer earlier selection of biologically appropriate patients and improved outcome; both recommendations require clinical trials.