Prognostic models for predicting death after hepatectomy in individuals with hepatic metastases from colorectal cancer.

BACKGROUND: Appropriate cancer treatment policies should include an accurate estimate of a patient's baseline risk of death, determined by the tumor load. Few prognostic models have reached the stage at which they can be used to select patients who would benefit from hepatectomy for metastatic colorectal cancer. Establishing the prognosis of patients with hepatic metastases from colorectal cancer is an important part of their evaluation and treatment.
METHODS: Pre- and post-treatment prognostic nomograms were developed using Cox regression and multiple imputation from 578 patients with hepatic metastases from colorectal cancer who were candidates for hepatectomy.
RESULTS: The preoperative nomogram included the following prognostic factors: primary histology, number of metastatic lymph nodes associated with the primary lesion, number of hepatic tumors, extrahepatic disease, and prehepatectomy carcinoembryonic antigen level. Plots of predicted versus actual outcomes suggested that the nomogram was well calibrated for predicting death after hepatectomy. The concordance index of the nomogram was 0.66, higher than those of other models for hepatic metastatic colorectal cancer in the literature. A postoperative nomogram was also prepared.
CONCLUSIONS: These models have improved predictive ability in individuals with hepatic metastases from colorectal cancer, which may be helpful in counseling patients and making treatment decisions.