One hundred patients with hepatic metastases from colorectal cancer treated by resection: analysis of prognostic determinants.

One hundred patients with hepatic metastases from colorectal cancer underwent 'radical' liver resection from 1980 to 1989. At least 1 cm of normal parenchyma surrounded the tumour and no microscopic invasion of resection margins was evident. The disease was staged according to our own staging system. Lobectomy was performed in 50 patients and non-anatomical resection in the remainder. The postoperative mortality rate was 5 per cent and the major morbidity rate was 11 per cent. The actuarial 5-year survival rate for patients in stages I, II and III was 42 per cent, 34 per cent and 15 per cent respectively (P less than 0.001). The overall actuarial 5-year survival rate was 30 per cent. The prognostic importance of various patient and tumour variables was evaluated by univariate analysis and then by multivariate analysis. Age of patient, site of primary, disease-free interval between treatment of primary and of hepatic metastases, preoperative carcinoembryonic antigen levels, and number of metastases, did not relate to prognosis, while sex (P = 0.024), stage of primary (P = 0.026), extent of liver involvement (P less than 0.001), distribution of metastases (P = 0.01) and type of surgery (P = 0.028) significantly affected prognosis as single factors. Multivariate analysis revealed that only the extent of liver involvement and stage of the primary tumour were independent predictors of survival. We conclude that liver resection is effective in selected patients with hepatic metastases from colorectal cancer. In resectable patients it is not yet possible to formulate a clear prognosis based on clinical factors. The extent of liver involvement and the staging system used may be significant, although not absolute, indicators of outcome.