Prognostic factors and survival after resection of colorectal liver metastasis in the era of preoperative chemotherapy: an 11-year single-centre study.

INTRODUCTION

A variety of factors have been identified in the literature which influence survival following resection of colorectal liver metastases (CRLM). Much of this literature is historical, and its relevance to contemporary practice is not known. The aim of this study was to identify those factors which influence survival during the era of preoperative chemotherapy in patients undergoing resection of CRLM in a UK centre.

METHODS

All patients having liver resection for CRLM during an 11-year period up to 2011 were identified from a prospectively maintained database. Prognostic factors analysed included tumour size (≥5 or <5 cm), lymph node status of the primary tumour, margin positivity (R1; <1 mm), neo-adjuvant chemotherapy (for liver), tumour differentiation, number of liver metastases (≥4), preoperative carcinoembryonic antigen (CEA; ≥200 ng/ml) and whether metastases were synchronous (i.e. diagnosed within 12 months of colorectal resection) or metachronous to the primary tumour. Overall survival (OS) was compared using Kaplan-Meier plots and a log rank test for significance. Multivariate analysis was performed using a Cox regression model. Statistical analysis was performed in SPSS v19, and p < 0.05 was considered to be significant.

RESULTS

432 patients underwent resection of CRLM during this period (67% male; mean age 64.5 years), and of these, 54 (13.5%) had re-resections. The overall 5-year survival in this series was 43% with an actuarial 10-year survival of 40%. A preoperative CEA ≥200 ng/ml was present in 10% of patients and was associated with a poorer 5-year OS (24 vs. 45%; p < 0.001). A positive resection margin <1 mm was present in 16% of patients, and this had a negative impact on 5-year OS (15 vs. 47%; p < 0.001). Tumour differentiation, number, biliary or vascular invasion, size, relationship to primary disease, nodal status of the primary disease or the use of neo-adjuvant chemotherapy had no impact on OS. Multivariate analysis identified only the presence of a positive resection margin (OR 1.75; p < 0.05) and a preoperative CEA ≥200 ng/ml (OR 1.88; p < 0.01) as independent predictors of poor OS.

CONCLUSION

Despite the wide variety of prognostic factors reported in the literature, this study was only able to identify a preoperative CEA ≥200 ng/ml and the presence of tumour within 1 mm of the resection margin as being of value in predicting survival. These variables are likely to identify patients who may benefit from intensive follow-up to enable early aggressive treatment of recurrent disease.