Circulating tumor DNA dynamics and mutational profiles as indicators in neoadjuvant chemoimmunotherapy for resectable or borderline resectable non-small cell lung cancer

Abstract Background Neoadjuvant chemoimmunotherapy has reshaped the treatment landscape for non-small cell lung cancer (NSCLC), yet critical gaps remain in real-time monitoring and personalized adaptation. Methods In this observational study, patients were evaluated via imaging and multidisciplinary review for subsequent surgery or radiotherapy after neoadjuvant chemoimmunotherapy. Serial plasma samples for circulating tumor DNA (ctDNA) dynamics and mutational analyses were collected. The endpoint was event-free survival (EFS). A pognostic mutational signature was developed using the Least Absolute Shrinkage and Selection Operator (LASSO) regression. Results Among the 138 patients, 46.4% and 52.2% underwent surgery and radiotherapy, respectively. Compared to patients with persistently positive ctDNA, the ctDNA non-shedder exhibited a numerically longer EFS (hazard ratio [HR] 0.26, 95% CI: 0.06–1.25, p = 0.062), and the clearance group exhibited a statistically longer EFS (HR 0.33, 95% CI: 0.11–0.97, p = 0.040). Patients with positive ctDNA before surgery/radiotherapy showed a shorter EFS (HR 3.38, 95% CI: 1.26–9.08, p = 0.010), and positive ctDNA was correlated with a numerically shorter EFS (HR 9.72, 95% CI: 0.98–95.96, p = 0.051) in patients assessed as stable disease (SD) of high heterogeneity in the exploratory subgroup analysis. For patients with post-C1 negative ctDNA, no EFS difference was observed between those receiving 3–4 versus 2 cycles of chemoimmunotherapy (HR 1.54, 95% CI: 0.56–4.26, p = 0.41). For patients with post-C1 positive ctDNA, the post-C1 quantity was associated with clearance (AUC = 0.72, 95% CI: 0.55–0.89, p = 0.019). After balancing clinical characteristics, LASSO-Cox regression selected four genes (SLX4, IRF4, TP53, and SOX2) to construct a prognostic signature for EFS. The AUC at 36 months was 0.75, 0.71, 0.74 in the training, validation and full cohort. The high-risk group demonstrated a shorter EFS ( p = 0.0002). Conclusions Longitudinal ctDNA assessment identified high-risk subgroups in neoadjuvant NSCLC, particularly by refining prognosis in heterogeneous SD populations. An exploratory gene-based mutational signature provided potential molecular risk stratification.