Background: The potential for next generation sequencing (NGS) analysis in personalized diagnosis and treatment of cancer is advancing rapidly. Measurement of mutation fraction less than 5% may be clinically relevant in analysis of circulating tumor DNA and/or small cytologic samples. However, due to uncertainty introduced by variation in sample quality, the mutation fraction lower limit of detection (LOD) reported by clinical laboratories does not typically extend beyond this 5% threshold. Even with a more conservative LOD criterion, (e.g., 20% mutation fraction), poor harmony may result from inter-laboratory variation in skills to assess sample quality as well as specimen characteristics that are not evident even to expert clinical pathologists. Thus, a critical unmet need is methods that harmonize reporting of mutation fraction measured by NGS by accurately estimating specimen-specific confidence limits for each measurement. Methods: We designed a mixture of spike-in competitive internal standards for tumor somatic mutations in 32 currently actionable genes that enable quality-controlled measurement by targeted NGS using either PCR-amplicon or hybrid-capture enrichment. We also developed a data analysis pipeline suitable to separate target and internal standard sequencing reads prior to analysis of target copy number and sequencing errors. Through participation in the FDA-led SEquencing Quality Control Project Phase 2 (SEQC2), where a set of highly characterized cell line mixtures was developed to assess cross-site reproducibility and mutation detection accuracy of multiple targeted sequencing platforms, we aim to comprehensively evaluate the utility of the spike-in mixture with various targeted NGS panels to measure 95% confidence limits around each somatic mutation fraction measurement. Results: In pilot studies, combining the mixture of internal standards with specimens prior to NGS library preparation enabled reliable measurement of 95% confidence limits around each somatic mutation measurement. Further, using the internal standards to control for specimen adequacy enabled reliable measurement of actionable mutation fraction at as low as 0.01% in suitable specimens. Conclusion: We expect that the full cross-site SEQC2 study will provide further support for the role of competitive spike-in mixtures to provide reliable 95% confidence limits around each measurement of somatic mutation fraction in a clinical specimen. Disclaimer: The views expressed here are those of the authors only and do not necessarily express the views/policies of the FDA.
Citation Format: James C. Willey, Thomas M. Blomquist, Erin L. Crawford, Joshua Xu, Leihong Wu, Thomas Morrison. Inter-laboratory harmonization of next generation sequencing somatic mutation assays for cancer response prediction [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 1623.