Abstract LB-017: Combining tumor mutational burden, CD8+ T-cell abundance and PD1 mRNA expression accurately predicts response to anti-PD1/PDL1 therapy across cancers

Abstract

Abstract Introduction Anti-PD1/PDL1 therapies provide significant survival benefits in many cancers, but the efficacy of these treatments varies considerably across different cancer types. Identifying the underlying variables associated with the cancer type specific response remains an important open research challenge. Here, we systematically evaluate a multitude of variables to determine the key factors that can accurately predict the response to anti-PD1/PDL1 therapy across different cancer types. Approach To this end we analyzed whole exome sequencing (WES) and RNA sequencing (RNAseq) of 10k patients from the Cancer Genome Atlas (TCGA) and the overall response rate (ORR) to anti-PD1/PDL1 therapy of 21 cancer types obtained from previous clinical trials. We considered more than 30 different variables of three distinct classes: those associated with (a) tumor neoantigen landscape, (b) tumor microenvironment and inflammation, and (c) the checkpoint inhibitor targets. We evaluated the performance of each of these variables and their combinations in predicting the ORR to anti-PD1/PDL1 therapy. Prediction accuracy was quantified with Spearman correlation and explained variance quantified through a standard leave-one-out cross-validation. Results Our analysis shows CD8+ T-cell abundance is the most predictive of the response to anti-PD1/PDL1 therapy across cancer types, followed by the tumor mutational burden and the fraction of samples with high PD1 gene expression in each cancer type. Notably these top three predictors cover the three distinct classes considered, and their combination is highly predictive of the ORR to anti-PD1/PDL1 therapy (Spearman R=0.9, P<1E-7), explaining more than 80% of the variance observed across different tumor types. Conclusions and Relevance This is the first systematic evaluation of the different variables associated with PD1/PDL1 therapy response across different tumor types. Our findings show that, overall cancer types, three key variables can explain most of the observed cross-cancer response variability, but that their relative explanatory roles may vary in specific cancer types. Citation Format: Joo Sang Lee, Eytan Ruppin. Combining tumor mutational burden, CD8+ T-cell abundance and PD1 mRNA expression accurately predicts response to anti-PD1/PDL1 therapy across cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr LB-017.