Analytical and clinical validation of CancerMaster, an automated targeted NGS panel, for tumor-only precision oncology

Abstract

We developed a CancerMaster-targeted next generation sequencing (NGS) panel covering 524 key genes specifically designed for the multidimensional molecular profiling of solid tumors. We aimed to establish CancerMaster as an asynchronous and parallel one-stop automated analysis pipeline with a reporting system to overcome the delay in existing NGS approaches. This panel can detect common genomic alteration types, including single nucleotide variants/indels and copy number variants (CNVs), fusions, Epstein-Barr virus/human papillomavirus infection, microsatellite instability (MSI), tumor mutational burden (TMB) status, and human leukocyte antigen typing. Using reference materials, CancerMaster demonstrated 100% reproducibility, 99% analytical sensitivity, and high accuracy (94%). Analytical performance for other biomarker classes-including CNVs, fusions, MSI, TMB, and viral detection-was evaluated separately. In a cohort of 668 patients, we identified actionable mutations, including in TP53, KRAS, and PIK3CA, and CNVs, e.g., ERBB2 amplification in gastric (n = 412) and colorectal (n = 66) cancers. MSI and TMB were strongly correlated in all patients (n = 668, r = 0.75; p < 10(- 15)), with consistent results in the gastric (n = 412, r = 0.75; p < 10(- 15)) and colorectal (n = 66, r = 0.87; p < 10(- 15)) cancer groups. Furthermore, directly comparing with the TruSight Oncology 500 (TSO500) panel, CancerMaster demonstrated high concordance while uniquely identifying certain clinically relevant alterations, including an ERBB2 missense mutation. Hence, the CancerMaster panel demonstrated high analytical performance and strong clinical potential for supporting clinical decisions regarding personalized cancer treatment.