Back

Tumor-infiltrating natural killer cell profiling for therapeutic stratification in patients with resectable non-small cell lung cancer

Ito, A.; Kang, S. W.; Jang, H.-J.; Shim, J. S.; Lee, C.; Ranchod, P.; Lee, G.; Mitra, S.; Jeon, J.; Syarif, A. H.; Zhang, Y.; Byun, J.; Han, Y.; Malo, J.; Patel, M.; Ripley, R. T.; Groth, S. S.; Kheradmand, F.; Burt, B. M.; Polverino, F.; Amos, C. I.; Lee, H.-S.

2026-01-02 immunology
10.64898/2026.01.01.696292 bioRxiv
Show abstract

ObjectiveNeoadjuvant chemoimmunotherapy has improved outcomes in resectable non-small cell lung cancer (NSCLC), yet its real-world implementation is often challenged by surgical delays, immune-mediated fibrosis, and postoperative complications. Smokers with NSCLC, despite a high risk for surgical morbidity, show enhanced responses to neoadjuvant chemoimmunotherapy. This study aimed to identify smoking-associated immune cell determinants that could guide treatment strategies. MethodsSingle-cell RNA sequencing (scRNAseq) was performed on 61 lung tissues from non-smokers and smokers to identify smoking-related immune compositions. The scRNAseq data from 19 invasive lung adenocarcinomas were used to validate their presence in the tumor-immune microenvironment. Bulk RNA sequencing data from 102 resected NSCLC and 24 NSCLC patients treated with neoadjuvant chemoimmunotherapy followed by surgery were used for in silico cellular deconvolution and outcome analyses. ResultsAmong 135 lung cellular phenotypes, two natural killer (NK) cell subsets were strongly associated with smoking and chronic obstructive pulmonary disease (COPD) severity. "Stress-responsive" NK (NKSR) cells exhibited immature features and cytokine-responsive features, and "Adaptive and immunoregulatory NK" (NKAIR) cells showed mature features and elevated multiple immune checkpoint expression. High intratumoral NKSR cells correlated with improved survival after surgery, particularly in current smokers. Conversely, tumors with low NKSR and high NKAIR cells responded more favorably to neoadjuvant chemoimmunotherapy. ConclusionsIntratumoral NK cell phenotyping may aid in therapeutic stratification in patients with NSCLC. NKSR cell preservation predicts benefit from upfront surgery, while NKAIR cell enrichment indicates improved response to neoadjuvant chemoimmunotherapy. These NK cell profiles may help optimize treatment by balancing therapeutic benefit and risk. Graphic Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=112 SRC="FIGDIR/small/696292v1_ufig1.gif" ALT="Figure 1"> View larger version (50K): org.highwire.dtl.DTLVardef@168bc03org.highwire.dtl.DTLVardef@10fa947org.highwire.dtl.DTLVardef@132a92dorg.highwire.dtl.DTLVardef@100e4a4_HPS_FORMAT_FIGEXP M_FIG C_FIG Central Picture O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=81 SRC="FIGDIR/small/696292v1_ufig2.gif" ALT="Figure 2"> View larger version (21K): org.highwire.dtl.DTLVardef@1a2d3b4org.highwire.dtl.DTLVardef@f8f84borg.highwire.dtl.DTLVardef@17510bcorg.highwire.dtl.DTLVardef@9da763_HPS_FORMAT_FIGEXP M_FIG C_FIG NK cell profiling for treatment stratification in resectable NSCLC Central MessageTumor-infiltrating NK cell profiling identifies distinct immune phenotypes that can inform personalized treatment strategies in resectable NSCLC. Perspective statementAbundance of intratumoral stress-responsive NK cells predicts favorable outcomes after upfront surgery, whereas enrichment of adaptive and immunoregulatory NK cells suggests greater benefit from neoadjuvant chemoimmunotherapy. These findings support incorporating NK-cell profiling into preoperative decision-making to optimize therapy selection and balance risks in the high-risk surgical population.

Published in The Journal of Thoracic and Cardiovascular Surgery · not in our set (fewer than 10 published preprints to learn from) · training set

Matching journals

The top 9 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.