Not all major pathological responses are equal: prognostic heterogeneity and risk stratification after neoadjuvant chemoimmunotherapy in resectable NSCLC.

Background: Major pathological response (MPR) is a widely used efficacy endpoint for neoadjuvant chemoimmunotherapy (NCIT) in resectable non-small cell lung cancer (NSCLC). However, whether MPR without pathological complete response (pCR) represents a prognostically homogeneous population remains unclear, and optimal postoperative management strategies require further exploration. This study evaluated the prognostic heterogeneity, risk stratification, and recurrence patterns of patients achieving MPR. Methods: This dual-center, retrospective cohort study included 588 patients with resectable NSCLC who received NCIT followed by radical surgery. The primary endpoint was recurrence-free survival (RFS). Patients were categorized into mutually exclusive pCR (n=200), non-pCR MPR (≤10% residual viable tumor without pCR; n = 111), and Non-MPR (n=277) groups. Kaplan-Meier curves and multivariable Cox proportional hazards regression were used to evaluate prognostic factors and develop a risk-stratification system within the MPR subgroup. Additionally, the effect of postoperative adjuvant therapy was evaluated via exploratory subgroup analyses, and anatomical recurrence patterns were systematically compared across response groups. Results: Recurrence-free survival (RFS) differed among the three groups, with MPR representing an intermediate-risk state. Within the MPR subgroup, baseline cT3-4 stage, postoperative ypN positivity, and lower harvested lymph node count were independently associated with RFS. A simplified system effectively divided MPR patients into low-risk (RFS comparable to pCR) and high-risk groups (RFS approaching Non-MPR). Adjuvant therapy showed no statistically significant RFS benefit in pathological response subgroups. Furthermore, recurrence patterns did not significantly differ among response groups, with brain and bone remaining the predominant distant metastatic sites. Conclusions: Non-pCR MPR after NCIT is a prognostically heterogeneous state in NSCLC. Stratifying these patients using baseline cT stage, ypN status, and lymph node count refines postoperative risk assessment, highlighting the continued necessity of thorough lymph node dissection. Although adjuvant therapy did not show a definitive RFS benefit in our cohort, its potential value, especially in high-risk MPR patients, warrants prospective validation and should not be interpreted as evidence of lack of benefit. Furthermore, the observed recurrence patterns across response groups, with brain and bone being frequently observed metastatic sites, provide preliminary insights into metastatic distribution after NCIT and may serve as a reference for future surveillance studies.