In situ detection of PD1–PD-L1 interactions as a functional predictor for response to immune checkpoint inhibition in NSCLC.

Introduction: Immune checkpoint inhibitors (ICIs) have transformed lung cancer treatment, yet their effectiveness seem restricted to certain patient subsets. Current clinical stratification on the basis of programmed death ligand 1 (PD-L1) expression offers limited predictive value. Given the mechanism of action, directly detecting spatial programmed cell death protein 1 (PD1)–PD-L1 interactions might yield more precise insights into immune responses and treatment outcomes. Methods: We applied a second-generation in situ proximity ligation assay to detect PD1–PD-L1 interactions in diagnostic tissue samples from 16 different cancer types, a tissue microarray with surgically resected early-stage NSCLC, and finally diagnostic biopsies from 140 patients with advanced NSCLC with and without ICI treatment. RNA sequencing analysis was used to identify potential resistance mechanisms. Results: In the early-stage NSCLC, only approximately half of the cases with detectable PD-L1 and PD1 expression exhibited PD1–PD-L1 interactions, with significantly lower levels in EGFR-mutated tumors. Interaction levels varied across cancer types, aligning with reported ICI response rates. In ICI-treated patients with NSCLC, higher PD1–PD-L1 interactions were linked to complete responses and longer survival, outperforming standard PD-L1 expression assays. Patients who did not respond to ICIs despite high PD1–PD-L1 interactions exhibited additional expression of stromal immune mediators (EOMES, HAVCR1/TIM-1, JAML, FCRL1). Conclusion: Our study proposes a diagnostic shift from static biomarker quantification to assessing active immune pathways, providing more precise ICI treatment. This functional concept applies to tiny lung biopsies and can be extended to further immune checkpoints. Accordingly, our results indicate concerted ICI resistance mechanisms, highlighting the need for combination diagnostics and therapies.