AACR Late-Breaking Research Predicting Response to Anti-PD1/PDL1 Therapy beyond Tumor Mutational Burden

SUMMARY: Immunotherapy with checkpoint inhibitors such as anti-PD1/PDL1 antibodies, is rapidly moving to the forefront of cancer treatment. These agents include PD1 targeted therapies such as KEYTRUDA® (Pembrolizumab), OPDIVO® (Nivolumab) and LIBTAYO® (Cemiplimab-rwlc) and PDL1 targeted therapies such as TECENTRIQ® (Atezolizumab), IMFINZI® (Durvalumab) and BAVENCIO® (Avelumab). Treatment with checkpoint inhibitors given as a single agent or in combination with chemotherapy has resulted in significant survival benefit in a variety of solid tumors, as well as hematologic malignancies. The efficacy of checkpoint inhibitors however varies considerably across different cancer types. Understanding tumors and their microenvironment and identifying the underlying variables that predict response to anti-PD1/PDL1 antibodies, has been challenging.

Tumor Mutational Burden (TMB) has recently emerged as a potential biomarker for immunotherapy with anti PD-1/PDL1 antibodies. TMB can be measured using Next-Generation Sequencing (NGS) and is defined as the number of somatic coding base substitutions and short insertions and deletions (indels), per megabase of genome examined. Several studies have incorporated Tumor Mutational Burden (TMB) as a biomarker, using the validated cutoff of TMB of 10 or more mutations/megabase as High, and less than 10 mutations/megabase, as Low. Drawbacks with TMB include sample consumption, higher attrition rate due to sample quality and quantity, and lack of standardization for the different TMB testing assays, with the definition of High TMB varying across studies from 7.4 or more to 20 mutations/megabase.

The Cancer Genome Atlas (TCGA), a landmark cancer genomics program, is a joint effort between the National Cancer Institute and the National Human Genome Research Institute. This program began in 2006 and has molecularly characterized over 20,000 primary cancers and matched normal samples, across 33 different cancer types. After 12 years and contributions from over 11,000 patients, TCGA has deepened our understanding of the molecular basis of cancer, changed the way cancer patients are managed in the clinic, established a rich genomics data resource for the research community and helped advance health and science technologies.

The authors in this study systematically analyzed Whole Exome Sequencing (WES) and RNA sequencing (RNAseq) data of 10,000 patients from the Cancer Genome Atlas, and the Overall Response Rate (ORR) to anti-PD1/PDL1 therapy of 21 different cancer types obtained from previous clinical trials. The researchers took into consideration more than 30 different variables belonging to three distinct classes: a) those associated with tumor neoantigen landscape (Tumor Mutational Burden-TMB) b) tumor microenvironment and inflammation, and c) the checkpoint inhibitor targets (PD1/PDL1). The performance of each of these variables and their combinations was then evaluated in predicting the ORR to anti-PD1/PDL1 therapy.

It was noted that the most important predictor of response to anti-PD1/PDL1 therapy across cancer types was CD8+ T-cell abundance in the tumor microenvironment, followed by the Tumor Mutational Burden, and a high PD1 gene expression in each cancer type in a fraction of samples. These three top predictors encompassed the three distinct classes considered in this analysis, and their combination was highly predictive of the ORR to anti-PD1/PDL1 therapy, and was able to explain more than 80% of the variance observed across different tumor types.

The authors concluded that in this first systemic evaluation of the different variables associated with PD1/PDL1 therapy response across different tumor types, the three top predictors mentioned above can explain most of the observed cross-cancer response variability. Combining tumor mutational burden, CD8+ T-cell abundance and PD1 mRNA expression accurately predicts response to anti-PD1/PDL1 therapy across cancers. Lee JS and Ruppin E. Presented at: 2019 AACR Annual Meeting; March 29 to April 3, 2019; Atlanta, GA.LB-017/9