Abstract Colorectal cancer (CRC) recurrence after curative surgery remains a major challenge to long-term patient outcomes. While individual risk factors have been extensively studied, their combined effects on recurrence patterns are poorly understood. This study explored recurrence predictors using integrated analysis of clinical, pathological, treatment-related, and molecular factors to better characterise recurrence. A retrospective cohort of 1100 patients undergoing curative-intent CRC surgery at Ninewells Hospital, Dundee (2018–2023) was analysed. Recurrence data were stratified by tumour site, focusing on the first 3 years post-surgery (where 80% of patients recur). Clinical variables, TNM staging, resection margins, systemic therapy, and molecular markers (MLH1, KRAS, NRAS, BRAF, MSI) were evaluated using Kaplan-Meier and Cox regression analyses. Multivariate modelling combined these factors to reflect real-world clinical complexity—an approach underutilised in current literature. Recurrence occurred in 17% of patients. Clinical patterns largely matched established literature—liver and lung predominated as metastatic sites. However, integrated molecular analysis revealed novel insights: KRAS mutations independently predicted worse recurrence-free survival (HR 2.37), particularly with aggressive lung metastases. BRAF mutations trended toward omental/peritoneal recurrence. MSI-high status correlated with significantly improved outcomes (P 0.05). MLH1 hyper-methylation emerged as an under-explored prognostic marker associated with poorer recurrence profiles. This study establishes a clinically relevant framework for predicting CRC recurrence by integrating routinely collected data. The identification of key independent predictors supports more personalised, risk-adapted follow-up strategies. The findings here can also support the development of predictive algorithms as well as the incorporation of emergent technologies such as circulating tumour DNA monitoring.
Yadav et al. (Sun,) studied this question.