Objectives: The aim of this study is to derive and validate a reliable indicator for predicting an increased risk of postoperative mortality in elderly patients undergoing curative resection for colorectal cancer (CRC).

Design: This study is of multicentre retrospective design.

Setting and participants: A total of 1227 CRC patients undergoing curative resection (age ≥65 years) from three distinct cohorts were retrospective enrolled. Participant cohorts consisted of the derivation (n=845), external validation (n=95) and localised validation (n=287) groups. The carcinoembryonic antigen (CEA) to lymphocyte ratio (CLR) was derived from the derivation cohort and subsequently validated in two additional cohorts. The observed end point was all-cause death during the follow-up period postoperation.

Results: In the derivation cohort, CLR demonstrated an independent association with all-cause mortality. In the two validation cohorts, CLR also presented a strong discriminatory ability in predicting postoperative all-cause death, with the area under the curve (AUC) of 0.68 in the external cohort and 0.78 in the localised cohort. Survival analyses revealed that CRC patients with CLR ≤2.53 tended to have better overall survival than those with CLR >2.53 (p<0.05 for all cohorts). Multivariate Cox proportional hazard models indicated that CLR ≤2.53 was significantly associated with reduced mortality risk in the derivation (HR: 0.405, p<0.001), external validation (HR: 0.519, p=0.039) and localised validation cohorts (HR: 0.167, p<0.001).

Conclusions: Preoperative CLR serves as a reliable predictor of all-cause death following curative resection in elderly patients with CRC. Individuals with CLR exceeding 2.53 are inclined to a lower overall survival probability.