Abstract 2514: Inter-test variability in tumor mutational burden (TMB) quantification and identification of TMB thresholds

Abstract

Abstract Introduction: Tumor mutational burden (TMB) is a possible predictive biomarker for response to immune checkpoint inhibitors (ICI). However, there is significant heterogeneity in the calculation of TMB and the thresholds that have been identified to associate with clinical benefit. We analyzed TMB and clinical outcomes across next generation sequencing (NGS) panels to address inter-panel heterogeneity and to identify consistent TMB thresholds. Methods: Non-small cell lung cancer sequencing data generated from three validated clinical platforms (OncoPanel n = 1155, MSK-IMPACT n = 1520, and Foundation Medicine n = 3377) and whole exome sequencing from The Cancer Genome Atlas (TCGA, n = 1144) were obtained. Tumor mutational burden (TMB) was calculated across cohorts as the number of nonsynonymous mutations per length of genome sequenced. The distribution of TMB was plotted for each cohort and compared to TCGA. For the subset of patients treated with ICI (OncoPanel n = 296; MSK-IMPACT,n = 227), durable clinical benefit (DCB) was defined as responsive/stable disease lasting ≥ 6 months and was associated with TMB. Results: There was high concordance across sequencing panels in the mean number of mutations per individual gene (OncoPanel vs TCGA R2 = 0.8914; Foundation vs TCGA R2 = 0.9025; MSK-IMPACT vs TCGA R2 = 0.8579). Absolute TMB values were higher in NGS panels compared to TCGA, likely due to enrichment for oncogenes and differences in germline filtration. Harmonization between NGS panels and WES was attempted by applying a linear transformation of panel TMB values to WES TMB values. Linear transformation constants were variable at the highest and lowest TMB values, and in low TMB subgroups such as EGFR mutants and never smokers. Consistent with prior studies, TMB was higher in patients with DCB compared to those with no durable benefit (NDB) (Oncopanel p-value=0.003; MSK-IMPACT p-value=0.006). Increasing TMB thresholds were associated with increase in DCB rate, likely due to enrichment of the highest TMB responders; DCB rates did not linearly associate with TMB at lower TMB values. Conclusion: There were significant differences in TMB quantification among NGS and TCGA panels, particularly at lower TMBs. Categorization, rather than fitting absolute values, may be an approach for harmonizing distinct panels. In these cohorts, the relationship between benefit of ICI and TMB was not consistent across the TMB spectrum but accentuated in the highest TMB patients. These findings may have implications for the application of TMB as a biomarker in clinical practice. Citation Format: Natalie Vokes, Elizabeth Jimenez Alguilar, Renato Umeton, Anika Adeni, Lynette Sholl, Matthew Hellmann, Hira Rizvi, Mark Awad, Eliezer Van Allen. Inter-test variability in tumor mutational burden (TMB) quantification and identification of TMB thresholds [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 2514.