Work

Higher proximity of PD-L1+ tumor cells to PD-1+ tissue-resident memory CD8 determines response to neoadjuvant chemoimmunotherapy in non-small cell lung cancer

Year: 2026

Type: article

Abstract: Background: Although neoadjuvant chemoimmunotherapy (NCIT) improves outcomes in resectable non-small cell lung cancer (NSCLC), traditional biomarkers like programmed death-ligand 1 (PD-L1) expression are insufficient for accurate patient selection. This study aimed to investigate the spatial interaction between specific T-cell subsets and PD-L1+ tumor cells using multiplex immunofluorescence (mIF), and to evaluate its value in predicting the efficacy of NCIT in NSCLC patients. Methods: We retrospectively recruited 44 patients with stage IIA–IIIB NSCLC who had NCIT at Shandong Cancer Hospital and Institute from January 2021 to June 2023. Pre-treatment specimens were subjected to multicolor immunofluorescence staining (panel 1: CK/PD-L1/PD-1/CD8/CD103/CD4/FOXP3; panel 2: CK/CD8/CD4/α-SMA/CD31/HIF-1α) to quantify PD-L1+ tumor cells and specific target T cells (CD4/conventional CD4/regulatory CD4, CD8/CD8+ TRM/bystander CD8) and to delineate their spatial distribution. The Mann-Whitney U test, receiver operating characteristic (ROC) curves, and logistic regression were employed to examine the correlation between these indicators and treatment response. Spearman correlation analysis assessed their relationship with tumor microvasculature, cancer-associated fibroblasts (CAFs), and hypoxia-inducible factor-1α (HIF-1α). Results: Among the 44 patients, 52.3% showed a response. Compared with non-responders, responders had closer distances between PD-L1+ tumor cells and CD8+ TRM, CD8, bystander CD8, conventional CD4, CD4, and regulatory CD4 prior to treatment, with decreasing area under the curve (AUC) values (0.728, 0.715, 0.680, 0.644, 0.643, 0.612). Further analysis of CD8+ TRM expressing programmed cell death 1 (PD-1) revealed that PD-1+ CD8+ TRM was the best predictor with an AUC of 0.743. Logistic regression analysis indicated that the closer PD-L1+ tumor cells were to CD8+ TRM, the better the treatment response [odds ratio (OR) =10.43, 95% confidence interval (CI): 1.49–73.08, P=0.02], especially for PD-1+ CD8+ TRM (OR =8.83, 95% CI: 1.81–43.13, P=0.007). Additionally, PD-L1+ tumor cell-CD8+ TRM interactions and PD-L1+ tumor cell-PD-1+ CD8+ TRM interactions were significantly positively correlated with HIF-1α+ CD8 (r=0.36 and 0.35, respectively, P<0.001 for both). Conclusions: T cells interacting with PD-L1+ tumor cells may be characterized as PD-1+ CD8+ TRM cells. A closer distance between the two enhances therapeutic efficacy and may be associated with density of hypoxic CD8 cells.

Source: Translational Lung Cancer Research

DOI: 10.21037/tlcr-2025-aw-1259