Circulating pre-treatment T-Cell Receptor repertoire as a predictive biomarker in advanced or metastatic non-small cell lung cancer patients treated with pembrolizumab alone or in combination with chemotherapy

The circulating T-cell receptor (TCR) repertoire is a dynamic representation of overall immune responses in an individual. We prospectively collected baseline blood from patients treated with first-line pembrolizumab monotherapy or in combination with chemotherapy. TCR repertoire metrics were correlated with clinical benefit rate (CBR), progression-free survival (PFS), overall survival (OS) and immune-related adverse events (irAEs). We built a logistic regression classifier by fitting all four TCR-β repertoire metrics to the ICI CBR data. In the subsequent receiver operating characteristic (ROC) analysis of the resulting logistic regression model probabilities, the best cut-off value was selected to maximise sensitivity to predict CBR to ICI. We observed an association between reduced number of unique clones and CBR among patients treated with pembrolizumab monotherapy (cohort 1) (RR=2.86,95%CI 1.04-8.73, P=0.039). For patients treated with pembrolizumab plus chemotherapy (cohort 2), increased number of unique clones (HR=2.96,95%CI 1.28-6.88,P=0.012) and Shannon Diversity (HR=2.73,95%CI 1.08-6.87, P=0.033), and reduced evenness (HR=0.43, 95%CI 0.21-0.90, P=0.025) and convergence (HR=0.41,95%CI 0.19-0.90, P=0.027) were associated with improved PFS, while only increased number of unique clones (HR= 4.62,95%CI 1.52-14.02, P=0.007) was associated with improved OS. Logistic regression models combining the TCR repertoire metrics improved the prediction of clinical benefit rate (cohort 1 and 2) and was strongly associated with PFS (cohort 1,HR=0.38,95%CI 0.19-0.78,P=0.009) and OS (cohort 2,HR=0.20,95%CI 0.05-0.76,P