Tumor-normal sequencing to reveal frequency of false positive error rates with tumor-only sequencing in GI cancers and with RNA sequencing as a new standard of precision molecular profiling.

Abstract

224 Background: Tumor-only sequencing to identify somatic tumor genomic variants increases the risk of mistakenly identifying germline mutations as somatically derived driver mutations. Simultaneous bioinformatics analysis of both normal germline and the tumor genome with RNA analysis is necessary for precise identification of molecular targets for cancer therapy. The objective of this study was to compare the accuracy and precision of tumor somatic SNV calling with a 45-gene GI cancer molecular actionability panel, analyzing tumor tissue alone versus analyzing tumor DNA with normal germline DNA and tumor RNA. Methods: The study included 204 cancer patients with 11 GI cancer types with whole genome sequencing of tumor and normal genomes. True positive and false positive rates were measured for single nucleotide variants (SNVs). RNA sequencing was available for 139 patients to determine SNVs expressed in messenger RNA. SNVs were identified using well-established and published bioinformatics methods. False positive results were SNVs identified as being somatic origin that were of germline origin. Results: 92% SNVs identified from sequencing tumor genomes alone were of germline origin (false positives) rather than true somatic mutations. Filtering the SNVs using a public database with reported population allele frequencies < 0.001 still resulted in a false positive rate of 41%. Furthermore, 71% of GI cancer patients had at least one false positive SNV after filtering. RNA analysis showed 10% of true somatic variants were not expressed and 17% of patients had at least one true somatic variant that was not expressed. Conclusions: Simultaneous sequencing and analysis of both normal germline and tumor genomes is necessary for accurate identification of molecular targets. Analysis of tumor genome alone results in a high false positive rate. Higher precision is achieved with simultaneous tumor-normal DNA and RNA sequencing analysis. Treatment decisions based on tumor-only DNA analysis or in the absence of RNA analysis can result in administration of ineffective therapies and increasing risk of negative drug-related side effects.