Predictive Models in Colorectal Liver Metastases – Can We Personalize Treatment and Outcome?

Abstract

Background/Aims: Hepatic resection for colorectal liver metastases offers patients the best chance of long-term survival. Survival rates after resection range from 25 to 60%. Predictive models may risk-stratify patients and allow improved selection for surgery or other therapies. This review aims to achieve consensus regarding the predictors of survival after hepatic resection for colorectal metastases and evaluate current predictive models of outcome. Methods: A comprehensive literature review of published studies with more than 500 patients describing models for predicting long-term survival in patients undergoing hepatic resection for colorectal metastases. Results: Five large predictive models have been published to date. The three predictive models developed from the largest series agree over the key independent predictors of poor long-term outcome for patients undergoing liver resection for colorectal metastases. All five models have individual shortcomings, and whilst three have been internally validated, two have been externally validated with conflicting results. The Basingstoke Predictive Index appears to be accurate and internally validated. Conclusions: There is need within the international surgical oncology community for consensus regarding a predictive model to be universally adopted for risk-stratifying patients who may benefit from intensive surveillance and selection for adjuvant therapy and clinical trials.