Predictors of survival for locally advanced rectal cancer patients with poor response to neoadjuvant chemoradiotherapy

Patients with locally advanced rectal cancer (LARC) who respond poorly to neoadjuvant chemoradiotherapy (nCRT) represent a high-risk subgroup with inferior survival, but specific prognostic factors for this group are unclear. This study aimed to identify independent predictors of survival outcomes exclusively in poor responders. This retrospective study included patients with LARC who exhibited a poor pathological response (ypT3-4 and/or ypN+) to nCRT between 2008 and 2024. The primary endpoint was overall survival (OS). Secondary endpoints were disease-free survival (DFS), locoregional recurrence-free survival (LRRFS), and distant metastasis-free survival (DMFS). Prognostic factors were analyzed using univariate and multivariate Cox proportional hazards models, with a significance threshold of p < 0.05. Among 295 poor responders (mean age 55.48 ± 12.51 years, 63.4% male), 89.8% had ≥ypT3 and 38.6% had ypN+ disease after nCRT. The median follow-up time was 42 months. The median OS was 67.4 months (95% CI: 57.9–98.2), with an estimated 3-year OS rate of 74.7%. In univariate analysis for OS, significant predictors were: age (HR 1.021, p = 0.013), distance from AV (HR 0.93, p = 0.025), lymph node (LN) ratio (HR 3.36, p = 0.002), ypN2 status (HR 2.08, p = 0.015), involved surgical margin (HR 2.59, p = 0.002), and TRG 2–3 vs. 0–1 (HR 2.24, p = 0.031). Based on multivariate analysis, independent predictors for OS were age (HR 1.044, p = 0.002), involved margin (HR 6.18, p < 0.001), ypN2 status (HR 3.02, p = 0.012), and a higher number of dissected LNs (HR 0.92, p = 0.010). In multivariate analysis, surgical margin, ypN2 status, and number of dissected LNs remained independent predictors for DFS, LRRFS, and DMFS. Additionally, the presence of PNI was a predictor of worse DFS, and advanced age predicted poorer DFS and DMFS. In LARC patients who respond poorly to nCRT, survival is dominantly determined by surgical outcomes and the burden of resistant disease. Specifically, an involved resection margin, ypN2 status, and a low LN count are independent predictors of worse survival across all endpoints. These findings indicate that the quality of surgical resection is a critical modifiable factor for this high-risk group. Therefore, management should prioritize these patients for intensified adjuvant systemic therapy. Future efforts should develop preoperative tools to identify such high-risk patients earlier, enabling more aggressive neoadjuvant treatment.