Background <p>Neoadjuvant chemo-immunotherapy improves outcomes in resectable non-small cell lung cancer (NSCLC), but predicting postoperative recurrence remains challenging. While the precursor exhausted T cell (Texp) index offers prognostic value, the broader tumor immune microenvironment (TIME) contains untapped predictive potential.</p> Methods <p>A secondary analysis was conducted based on single-cell RNA and TCR sequencing data from 138 NSCLC patients treated with neoadjuvant chemo-immunotherapy. Following collinearity filtering of 51 immune cell subtypes and the Texp index, LASSO-Cox regression was applied to develop a prognostic Risk Score for recurrence-free survival (RFS).</p> Results <p>A four-feature Risk Score was identified, incorporating the Texp index, CXCL13⁺ Th1-like CD4⁺ T cells, KLRB1⁺ MAIT CD8⁺ T cells, and terminally exhausted T cells. This score robustly stratified patients into distinct RFS groups and demonstrated independent prognostic value in multivariate analysis (HR = 6.090, <i>p =</i> 0.004). Compared to the baseline Texp index alone, the expanded Risk Score model yielded significantly improved predictive discrimination (mean ΔC-index = 0.096) and demonstrated superior net clinical benefit in decision curve analysis at 12 and 24 months.</p> Conclusion <p>This integrated single-cell immune Risk Score provides enhanced prognostic accuracy over single-metric biomarkers, offering a refined tool for postoperative risk stratification and personalized adjuvant management in NSCLC.</p>