Patient selection for hepatic resection of colorectal metastases.

OBJECTIVE: To determine the major factors governing patient outcome after hepatic resection of metastatic colorectal cancer and to formulate criteria for optimal resection. PATIENTS: We reviewed records of 74 patients (44 men, 30 women) who underwent resection of colorectal liver metastases. MAIN OUTCOME MEASURES: Sex, age, primary tumor location; Dukes tumor stage; disease-free interval after primary resection (synchronous vs metachronous); location, number, size, and distribution of liver metastases; operative complications; and mortality.
RESULTS: The primary tumor location was rectosigmoid in 46 patients and the colon in the others. The tumor stage was Dukes A in one patient, Dukes B in 16, Dukes C in 31, and Dukes D (synchronous metastases) in 26. The disease-free interval was less than 12 months in 38 patients and 12 months or more in 36. The location of the metastases was the right lobe in 42 patients, left lobe in 22, and bilateral in seven. The cancer was unilobar in 55 patients and bilobar in 18. Surgical resection included wedge resection in 23 patients, segmentectomy in 30, lobectomy in seven, and trisegmentectomy in 12. The number of lesions resected was one in 50 patients, two or three in 18, and four or more in five. Nine patients had repeated liver resections because of recurrence. There were five postoperative deaths within 60 days (7%), four of which occurred after extended resection and were complicated by delayed liver failure and multisystem failure. An additional death occurred at 65 days after an apparently uneventful initial convalescence. Overall median survival was 35 months; actuarial 5- and 10-year survival rates were 24% and 12% respectively. There were significant relationships with survival (P<.05) for the number of metastases (three or fewer vs four or more), bilobar vs unilobar metastases, and extent of liver resection (wedge and segmental vs lobectomy and trisegmentectomy). A multiple logistic regression model (multivariate analysis) showed a significant correlation with survival (P<.05) for distribution of metastases (bilobar vs unilobar) and extent of resection (wedge and segmental vs lobectomy and trisegmentectomy).
CONCLUSION: Patient selection for hepatic resection of colorectal cancer metastases based on standard clinical and tumor outcome variables should be expected to achieve long-term survival with low morbidity and mortality in bilobar disease or extended resection should generally be avoided, especially in medically compromised patients.