Neoadjuvant anti-PD-1 therapy plus chemotherapy versus chemotherapy alone for resectable stage IIA-IIIB non-small cell lung cancer: a two-center real-world propensity score-matched study

To compare pathological and radiological responses, perioperative outcomes, treatment-related adverse events (TRAEs), and exploratory survival outcomes between neoadjuvant chemoimmunotherapy (NCAI; anti-PD-1 therapy plus chemotherapy) and neoadjuvant chemotherapy alone (NCT) in patients with clinically resectable stage IIA-IIIB non-small cell lung cancer (NSCLC) who underwent surgery. This two-center retrospective real-world cohort study screened patients with clinically resectable stage IIA-IIIB NSCLC who underwent surgery between January 1, 2020, and December 31, 2022. Patients who received neoadjuvant therapy followed by surgery were classified into the NCAI or NCT group according to the neoadjuvant regimen received. The primary and key secondary endpoints were major pathological response (MPR) and pathological complete response (pCR), respectively. One-to-one propensity score matching was performed to balance measured pretreatment characteristics. In the matched cohort, associations of treatment with MPR and pCR were estimated using logistic regression with matched-pair cluster-robust standard errors and additional adjustment for residual imbalance in PD-L1 expression and tumor location. Disease-free survival (DFS) and overall survival (OS) were secondary exploratory endpoints and were analyzed using robust Cox proportional hazards models accounting for matched-pair clustering. Sensitivity analyses included alternative matching and modeling specifications, a biomarker complete-case analysis, and overlap weighting. Of 2,708 patients screened, 293 received neoadjuvant therapy followed by surgery (NCAI, n = 177; NCT, n = 116), and 98 matched pairs were analyzed. NCAI was associated with higher odds of MPR (53.1% vs 33.7%; adjusted odds ratio [OR], 2.68; 95% confidence interval [CI] 1.45–4.92; P = 0.002) and pCR (39.8% vs 12.2%; adjusted OR, 8.30; 95% CI 3.31–20.79; P < 0.001). The objective response rate was also higher with NCAI than with NCT (69.4% vs 52.0%; P = 0.013). No statistically significant between-group differences were detected in the evaluated perioperative outcomes or documented TRAEs. During a median follow-up of 25 months, 71 DFS events and 42 deaths were documented in the matched cohort. In exploratory analyses, NCAI was associated with lower hazards of a DFS event (adjusted hazard ratio [HR], 0.58; 95% CI 0.38–0.90; P = 0.014) and death (adjusted HR, 0.40; 95% CI 0.21–0.78; P = 0.007). Effect estimates were generally consistent in direction across sensitivity analyses; the pCR association remained statistically significant in all analyses, whereas the MPR association was attenuated and not statistically significant in the biomarker complete-case analysis. Among patients who underwent surgery after neoadjuvant therapy, NCAI was associated with higher odds of MPR and pCR and a higher objective response rate than NCT. No statistically significant differences were detected in the evaluated perioperative outcomes or documented TRAEs, although these findings do not establish equivalent perioperative safety or toxicity. Exploratory analyses also suggested favorable associations with DFS and OS. Given the retrospective design, potential residual confounding, treatment heterogeneity, limited event counts, and immature survival data, prospective confirmation with longer follow-up is warranted.