Resection of colorectal liver metastases: 25-year experience.

The aim of this retrospective study was to analyze survival and prognostic factors in 111 consecutive patients undergoing curative resection of liver metastases from colorectal cancer. In addition, the time periods 1971-1984 and 1985-1995 were compared; criteria for first liver resection did not change with time, whereas the attitude toward re-resection was more aggressive during the latter period. Operative mortality was 6% during 1971-1984 and 0% during 1985-1995 (3.6% for all patients). The crude 5-year actuarial survivals were 19% and 35% for patients operated during 1971-1984 and 1985-1995, respectively (25% for the whole period). Relapse at any site was observed in 52 patients (81%) operated during the first period and in 29 patients (67%) operated during the second period; re-resection was performed in 12 (23%) and 15 (52%) of these patients, respectively. Five-year survival after hepatic re-resection was 29% (no operative mortality). In the univariate analysis, significant determinants for long-term survival were, in descending order, a clear resection margin, high degree of fibrosis around the tumor, absence of extrahepatic metastases (including metastases to the liver hilum), use of an ultrasound dissector, low preoperative serum carcinoembryonic antigen (CEA) level, year of resection (1985-1995), and low/moderate grade of liver tumor. There were no 5-year survivors when extrahepatic metastases were present, the liver tumor(s) had a low differentiation or satellites, or the resection margin was involved with tumor. In the multivariate analysis, the determinants were grade of liver tumor, absence of extrahepatic tumor, few intraoperative blood transfusions, low preoperative serum CEA level, and year of resection (1985-1995). It is concluded that: (1) an increased rate of hepatic re-resection was partly responsible for the improved outcome after liver resection for colorectal metastases during recent years; (2) patients with extrahepatic metastases did not benefit from liver resection; and (3) surgery should be performed with a clear resection margin and minimal blood loss.