Abstract 1996: Tumor mutational burden is affected by next-generation sequencing enrichment method in highly deaminated samples from solid tumors

Abstract

Abstract Background: Tumor mutational burden (TMB) is currently under investigation as a biomarker for predicting response to anti-cancer immunotherapy. TMB can be defined as the number of somatic nonsynonymous mutations per Mb in a cancer genome, therefore DNA artifacts could lead to inaccurate measurements. Cytosine deamination (C:G>T:A) is a well-known phenomenon in formalin-fixed, paraffin-embedded (FFPE) samples, producing errors in next-generation sequencing (NGS). Herein, we measured TMB in FFPE samples using two comprehensive NGS panels for routine diagnostics implementation. Methodology: TMB assessment was performed on FFPE samples by amplicon-based target enrichment on Thermo Fisher's Oncomine™ Tumor Mutation Load (OTML; n=9) and/or hybrid capture-based on Illumina's TruSight™ Oncology 500 Assay (TSO500; n=8), according to manufacturer's instructions. TMB values were compared for uracil-DNA glycosylase (UDG) treatment prior to PCR amplification and enrichment method. Samples were also compared according to their previous results in a clinically-certified NGS and an in-house microsatellite instability (MSI) assay. TMB >=10 mutations/Mb was considered high, and MSI >=20% was considered unstable. Results and Discussion: The concordance between TMB values for OTML and TSO500 was 75% (n=4) and 62.5% (n=8) with and without UDG treatment, respectively. UDG reduced artifactual C:G>T:A for both enrichment approaches, reinforcing its presence in FFPE DNA; however, there was no alterations on TMB results for both methods (TSO500 n=5; OTML n=7). C:G>T:A variants could be successfully filtered by low allelic frequency (<5%) or low quality (<q20) in non-treated samples. For OTML, a high baseline noise was observed in samples with high deamination scores (>60), leading to TMB overestimation (>350), which could not be corrected even after UDG treatment and manufacture's bioinformatic adjustment. Interestingly, the same samples presented low TMB values for TSO500. Samples with low deamination rates were not influenced by UDG treatment or bioinformatics adjustment regardless of enrichment method. TMB values from a certified NGS was completely concordant for TSO500 (n=4) but only 25% for OTML (n=4). MSI was fully concordant between certified NGS and TSO500 (n=4) and between in-house method and TSO500 (n=4) (value not analyzed by OTML assay). No QC parameter predicted differences between high and low deaminated samples. Conclusions: NGS enrichment methods strongly influenced TMB assessment. Hybrid-based capture showed better performance than amplicon-based. UDG treatment and higher allelic frequency cutoff could minimize false positives at variant calling level, although no significant effect could be observed for general TMB values. Future experiments should cover other alternatives to distinct FFPE DNA damages. Citation Format: Michele A. Pereira, Feliciana L. Marinho, Joice P. Silva, Renato Puga, Maíra C. Freire, Mariana R. Monteiro, Rafael L. Guedes, Luiz H. Araujo. Tumor mutational burden is affected by next-generation sequencing enrichment method in highly deaminated samples from solid tumors [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1996.