Abstract P4-04-09: Comparing tumor mutation burden detection between whole exome and target enrichment sequencing among Taiwanese breast cancers

Abstract

Abstract Purpose: The ability to detect tumor mutation burden (TMB) is increasingly demanding and has become a pre-requisite for breast cancer personalized therapies, especially for those interrogating immune-modulating agents. The optimized quantification method determining TMB, however, remains inconclusive, and the current study compared tumor mutation yields among Taiwanese breast cancers, either with whole exome or target enrichment sequencing. Methods: Sporadic breast cancers were prospectively recruited. Target-enrichment sequencing was performed with Illumina SolexaTM technology with read length of 150 paired ends and was analyzed with Agilent SureCallTM. There were 56 targets comprising 990 regions with a total regional size of 173.999 kbp. For whole exome sequencing, SureSelectTM Human All Exome V6r2 was adopted. The minimum coverage was set to 20 and the minimum alternative reads was set to 10 to enhance sequencing reliability. Samtools version 1.2 was utilized as variant caller for both whole exome and targeted sequencing. Novel and synonymous variants, as well as those located outside coding regions were removed. TMB was calculated as the number of somatic mutations per megabase (mb). Results: A total of 61 and 52 Taiwanese breast cancers underwent target enrichment and whole exome sequencing, respectively. The number of somatic mutations ranged from 36 to 292 (median: 205), equivalent to 207~1678 (median: 1178) mutations/mb for targeted sequencing. On the other hand, there were 9947 to 10981 somatic mutations (median: 10727) when the whole exome was sequenced, resulting in TMB between 333 and 366 (median: 358). TMB for ER+/HER-, ER+/HER2+, ER-/HER2+, and ER-/HER2- was 1322 (n=31), 713 (n=11), 793 (n=10), and 1230 (n=9) for targeted sequencing and was 358 (n=29), 358 (n=12), 354 (n=5), and 357 (n=6) for whole exome sequencing. The most frequent nonsense mutations were NOTCH1 (S255*), RET (Q87*), KRAS (G57*), JAK3 (Y267*), and SMO (Y61*), while MAP2K1 (G148C), FNACA (G175V), BSG (H254P), ABL1 (T57P), and NOTCH1 (H416P) were the most common missense mutations. Discussion and conclusions: High-throughput parallel massive sequencing can identify large numbers of variants, dependent both on the size of the sequenced regions and the variant caller algorithm utilized. Although Samtools identified more variants against reference than other methods, there should be no differential between whole exome and targeted sequencing. The number of somatic mutations, as well as the variability in TMB, was much higher in target enrichment approach than whole exome sequencing, highlighting that different TMB threshold should be established before wide clinical applications. Uneven distributions of “hotspot” mutation regions across genome as well as hundred-fold of sequencing length of whole exome versus targeted sequencing sequencing may contribute to such discrepancy. Citation Format: Huang C-C, Huang C-S. Comparing tumor mutation burden detection between whole exome and target enrichment sequencing among Taiwanese breast cancers [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P4-04-09.