Comparing surgical resection of limited hepatic metastases from colorectal cancer to non-operative treatment.

The survival of two groups of patients, affected by liver metastases (Stage I and II by Gennari et al.) from a previously operated colorectal cancer and treated by surgical resection (Group 1, 39 patients) or chemotherapy with various cytotoxic drugs (Group 2, 31 patients) at the Istituto Nazionale Tumori, Milan, is reported. In comparison with Group 2, Group 1 included more patients with metachronous lesions, with high level of serum bilirubin and with primary tumour originating from the colon. A univariate analysis (log rank test) identified a statistically significant prognostic role of type of treatment (surgery vs chemotherapy) and of the level of serum bilirubin. However the multivariate analysis by the Cox's regression model showed that the only independent statistically significant prognostic factor was type of treatment, since the hazard ratio of surgery vs chemotherapy was 0.490 with a 95% confidence interval of 0.256-0.936. The survival probabilities at 24 and 36 months were respectively 60% and 47% in surgical patients, vs 30% and 23% in those receiving chemotherapy, the difference between the curves being statistically different (P = 0.001). The median survival of Group 1 patients was 30 months whereas the median survival of Group 2 patients was 19 months, a value quite similar to that published in literature for untreated patients with limited metastatic disease-thus indicating that this patients' population was not selected according to unfavourable criteria. These findings suggest a beneficial role of surgical resection in patients with colorectal metastases confined to the liver in Stages I and II.