Analytical and clinical validation of the BostonGene tumor portrait assay.

Abstract

e15085 Background: Analysis of the genetic and transcriptomic profile of solid tumors via next-generation sequencing (NGS) assays is fundamental to propel precision medicine into clinical practice. NGS technology applied to tumor analysis allows for the characterization of somatic alterations, clonality, altered gene expression, and other parameters using a small amount of tissue. Therefore, to uncover cancer-promoting and suppressing activity of the tumor and the tumor microenvironment (TME), we developed the BostonGene TUMOR PORTRAIT assay, integrating whole-exome sequencing (WES) and mRNA sequencing (RNA-seq). Here, we demonstrate the analytical and clinical validity of the assay. Methods: The accuracy, reproducibility, and robustness of the BostonGene assay were evaluated using reference genomic DNA, reference RNA, well-characterized cell lines, and fresh frozen (FF) tumor and normal tissue containing known single nucleotide variants (SNVs), indels, copy number alterations (CNAs), gene fusions and a reference RNA Spike-In mix containing known levels of specific transcripts. The analysis was performed using the BostonGene automated pipeline. Additionally, we demonstrated high concordance of gene expression measured using two orthogonal techniques, RNA-seq and RT-PCR. Results: The BostonGene TUMOR PORTRAIT assay demonstrated high specificity (>98.1%, >99.8%, >96.9%) and sensitivity (>98.3%, >99.2%, >97.1%) for the detection of SNVs, indels, and CNAs, respectively, with low false-positive and false-negative rates. The assay demonstrated a 100% concordance in the mutation (SNV/indel) call rate across all replicates, and a 100% concordance in the mutation (SNV/indel) and copy number variation (CNV) call rate across all runs. The measurement of gene expression from RNA-seq was achieved with high accuracy (>0.96%) and low variation across genes (<6.0%), demonstrating the ability of the assay to provide transcriptomic information. Furthermore, gene fusions were detected in RNA-seq data with high sensitivity (>95.8%) and specificity (>99%) in a reproducible manner. Using the integrated pipeline that utilizes both WES and RNA-seq, we were able to accurately compute all disease-relevant molecular parameters including tumor genomics, tumor transcriptome phenotype, expression of clinically actionable therapeutic targets, tumor microenvironment composition, and tumor clonality within the single BostonGene TUMOR PORTRAIT report. Conclusions: Analytical and clinical validation results show that the BostonGene TUMOR PORTRAIT assay, based on the sequencing of DNA and RNA, provides a comprehensive and accurate view of the tumor molecular profile, identifying all clinically actionable genetic, transcriptomic, and TME targets. Validation of the BostonGene TUMOR PORTRAIT assay provides a solid foundation for the future development of precision oncology.