RESUMO
BACKGROUND: The best management of liver metastases from colorectal cancer is still debated and little is known about the true impact of treatments on survival. MATERIALS AND METHODS: The study involved 122 patients (77 males), aged 64.0 ± 11.0 years (range: 27.8-86.1) at diagnosis of liver metastatization (synchronous in 59). All underwent chemotherapy and at least one procedure of radiofrequency ablation; 53 also had partial hepatic resections. Demographics, tumor characteristics and survival outcomes from liver metastatization were analyzed with univariate and multivariate techniques. This analysis was performed also taking into account relative survival as the best estimate of specific survival. RESULTS: The analysis with observed survival selected the categorized number of involved lymph nodes in the colorectal specimens as the only statistically significant predictor, while the analysis with relative survival also showed site of the primary tumor (above the sigmoid colon or otherwise) and number of liver metastases as significant factors. The standardized mortality ratio was 9.673 (95% CI: 7.668-11.663) and a total of 201.85 years of life were lost in comparison with the survival of the reference population. CONCLUSIONS: The computation of relative survival better than observed survival selected a more adequate number of predictors, making investigation of even limited series of patients with confounding factors reliable. The finding that prognosis was mainly dependent on the anatomical presentation of the primary tumor and of liver metastases instead of treatments could explain the still contrasting opinions on the role of the available therapies in this field.