Spark-like T cell represents a novel functional state of CD8+T cells associated with anti-PD-1 immunotherapy response and clinical prognosis in non-small cell lung cancer

Background Spark-like T cells constitute a novel functional state of CD8 + T cells and have emerged as key drivers of immunotherapy response in preclinical models, yet their clinical relevance in non-small cell lung cancer (NSCLC) remains undefined. Methods Single-cell RNA sequencing (GSE207422), spatial transcriptomics (GSE271689), bulk RNA expression cohorts (TCGA-NSCLC, GSE37745, GSE41271, and GSE72094), and multiplex immunohistochemistry (mIHC) of a 100-patient cohort treated with anti-PD-1 were integrated to characterize Spark-like T cells and assess their clinical relevance. A 14-gene prognostic signature was constructed via LASSO-Cox regression. Results Spark-like T cells were significantly enriched in anti-PD-1 responders in the single-cell RNA sequencing cohort. mIHC validation in the 100-patient cohort confirmed that high Spark-like T cell intensity in tumors was independently associated with prolonged overall survival (log-rank test, p < 0.001) and partial response to anti-PD-1 therapy (p = 0.0075). Spatial transcriptomics indicated strong co-localization of Spark-like T cells with M1 macrophages (Spearman r = 0.524, p = 3.4 × 10 -42 ), suggesting their spatial co-enrichment with pro-inflammatory myeloid cells. In silico knockout identified TNF, ENTPD1 (CD39), and IL2 as cooperative regulators of Spark-like T cell effector function. The 14-gene Spark-like T cell-related signature (VDAC1, ITGB1, GALNT2, SEMA7A, UBB, SLC7A5, FURIN, SNX9, PSMB2, CXCL13, TOX, PPM1M, CCND2, and CLNK) robustly predicted overall survival across all four independent cohorts and remained an independent prognostic factor after adjustment for age, sex, and tumor stage. High-risk patients exhibited an immune-excluded, oncogene-driven tumor microenvironment with reduced infiltration of CD8 + T cells. Conclusions Spark-like T cells are clinically relevant determinants of anti-PD-1 response in NSCLC. mIHC quantification and the 14-gene signature provide promising translational tools for patient stratification and risk assessment.