Nonlinear association between number of lymph nodes resected and survival in T1N0M0 colorectal cancer: a SEER-based analysis.

The number of lymph nodes retrieved (NLNR) is a key prognostic factor in colorectal cancer (CRC), yet its relationship with outcomes in early-stage (T1N0M0) disease remains unclear. We performed a retrospective cohort study using SEER data (2010-2022) including 8,012 T1N0M0 CRC patients treated with radical resection. Patients were dichotomized by NLNR (< 12 vs. ≥ 12) and propensity score matching (PSM) was applied. Survival was evaluated by Kaplan-Meier analysis; multivariable Cox regression examined NLNR as a continuous predictor of overall survival (OS). Restricted cubic spline (RCS) modelling explored nonlinear associations, including within subgroups defined by tumor location: right-sided colon, transverse colon, left-sided colon, and rectal cancer. Subgroup analyses were also performed. Before PSM, the ≥ 12-node group showed superior OS and cancer-specific survival (CSS) (both p < 0.001); after PSM differences were no longer statistically significant (OS p = 0.059; CSS p = 0.090). In multivariable Cox analysis NLNR was independently associated with improved OS (HR 0.992, 95% CI 0.985-0.998). RCS revealed a significant nonlinear relationship with OS (P for nonlinearity = 0.019): each additional node below 25 conferred reduced mortality (HR 0.982 per node, 95% CI 0.972-0.993, p < 0.001), whereas benefit plateaued beyond 25 nodes (HR 1.001, p = 0.343). The nonlinear association between NLNR and OS remained significant in the left-sided colon and rectal cancer subgroups but was not statistically significant among patients with right-sided or transverse colon cancer. Subgroup analyses were broadly consistent; a significant interaction with sex was observed (P for interaction = 0.012). In T1N0M0 CRC, NLNR is an independent predictor of long-term survival and exhibits a distinct threshold effect. This effect is more pronounced in left-sided colon and rectal cancers.