Diagnostic Validation of a Comprehensive Targeted Panel for Broad Mutational and Biomarker Analysis in Solid Tumors

Abstract

Simple Summary The analysis of tumor-associated genetic variants and biomarkers is critical for therapy choice, as specific mutations allow for a personalized treatment. Because more and more mutation-treatment combinations become available, screening should be performed on many genes simultaneously. The use of large and comprehensive gene panel screenings in molecular diagnostics, however, requires an extensive and thorough validation to demonstrate the correctness of all clinically relevant data. Here, we describe such validation using a large number of samples and confirmed effective detection of several types of mutations for different validation parameters. Samples of tumor patients thus can be reliably tested with a comprehensive assay to maximize their personalized treatment regimen. The use of targeted Next Generation Sequencing (NGS) for the diagnostic screening of somatic variants in solid tumor samples has proven its high clinical value. Because of the large number of ongoing clinical trials for a multitude of variants in a growing number of genes, as well as the detection of proven and emerging pan-cancer biomarkers including microsatellite instability (MSI) and tumor mutation burden (TMB), the currently employed diagnostic gene panels will become vastly insufficient in the near future. Here, we describe the validation and implementation of the hybrid capture-based comprehensive TruSight Oncology (TSO500) assay that is able to detect single-nucleotide variants (SNVs) and subtle deletions and insertions (indels) in 523 tumor-associated genes, copy-number variants (CNVs) of 69 genes, fusions with 55 cancer driver genes, and MSI and TMB. Extensive validation of the TSO500 assay was performed on DNA or RNA from 170 clinical samples with neoplastic content down to 10%, using multiple tissue and specimen types. Starting with 80 ng DNA and 40 ng RNA extracted from formalin-fixed and paraffine-embedded (FFPE) samples revealed a precision and accuracy >99% for all variant types. The analytical sensitivity and specificity were at least 99% for SNVs, indels, CNVs, MSI, and gene rearrangements. For TMB, only values around the threshold could yield a deviating outcome. The limit-of-detection for SNVs and indels was well below the set threshold of 5% variant allele frequency (VAF). This validated comprehensive genomic profiling assay was then used to screen 624 diagnostic samples, and its success rate for adoption in a clinical diagnostic setting of broad solid tumor screening was assessed on this cohort.