Evaluating multiplex immunofluorescence analysis of the tumor microenvironment for predicting immunotherapy efficacy in NSCLC compared to traditional biomarkers.

Abstract

e20600 Background: Only a subset of patients (pts) with advanced non-small cell lung cancer (aNSCLC) benefits from treatment with immune checkpoint inhibitors. Predictive biomarkers of immunotherapy efficacy are mainly limited to PD-L1 expression and tumor mutational burden (TMB). We investigated the potential of quantitative multiplex immunofluorescence (mIF) to improve prediction of clinical outcomes to immunotherapy in aNSCLC. Methods: Baseline clinicopathologic and genomic data were collected from pts with NSCLC, and treatment outcomes in response to PD-(L)1 inhibitors +/- CTLA-4 or other immune-modulating agents were analyzed. mIF markers included CD8 + , PD-1 + , FOXP3 + , PD-L1 + , and cytokeratin. Immune cell counts were categorized as high or low using cutoff values optimized via the log-rank statistic maximization method for progression-free survival (PFS). Groups were defined by examining overlaps between higher and lower cell densities, based on calculated cut-offs, and analyzing their combinations to enhance outcome stratification. Cox proportional-hazard models estimated hazard ratios (HR) for PFS and overall survival (OS) in univariate and multivariate analyses, significance set at a P < 0.05. Results: Among 117 pts with aNSCLC treated with immunotherapy, the median age was 67, 90.6% (N=87) had history of tobacco use, 85.5% had non-squamous histology (N=100), 86.3% had an ECOG PS of 0-1 (N=101), and 68.8% received immunotherapy as first-line treatment (N=83). PD-L1 TPS ≥ 50% was observed in 24.8% of cases (N=29), and the median TMB (N=112) was 9.89 mut/Mb using the OncoPanel platform. A total of 63.7% (N=72) had a mutation in KRAS , BRAF , EGFR , MET , HER2 , or RET . We identified three distinct immune cell density groups: Low CD8 + /Low PD-1 + (N = 25), High CD8 + /Low PD-1 + (N = 69), and High CD8 + /High PD-1 + (N = 23). Immunotherapy outcomes progressively improved with increasing infiltration of CD8 + and PD-1 + cells. The High CD8 + /High PD-1 + group exhibited the highest overall response rate (ORR, 56%) as well as the longest median PFS (25.6 months) and OS (not reached), compared to the High CD8 + /Low PD-1 + group (ORR 30%; mPFS 6.7 months; mOS 24.5 months) and the Low CD8 + /Low PD-1 + group (ORR 16%; mPFS 4.2 months; mOS 14.7 months) ( P < 0.001). These findings remained consistent in multivariable analyses adjusted for smoking status, ECOG, treatment line, PD-L1 expression, and TMB. Importantly, the High CD8 + /High PD-1 + group was the only biomarker combination to retain statistical significance for both PFS (HR 0.28, P < 0.001) and OS (HR 0.35, P < 0.01). Conclusions: Increasing infiltration of both CD8 + and PD-1 + cells, as assessed by mIF, was associated with improved immunotherapy outcomes in NSCLC. Combining immune profiling with existing biomarkers may improve the identification of pts who can benefit most from immunotherapy.