Background: Validation is very important for diagnostic assays, here we describe a validation process of a comprehensive cancer genomic profiling assay based on massively parallel DNA sequencing. This assay can detect base substitutions, short insertions and deletions, copy number alterations, tumor mutation burden (TMB) and microsatellite instability (MSI) across 543 cancer-related genes from tumor specimens or plasma. 

Methods: Analytical validation was conducted with 20 cell line pools whose mutation were verified by digital droplet PCR (ddPCR). Clinical validation was conducted with Proficiency Test samples and clinical samples whose mutation were verified by ddPCR. Sensitivity, positive predictive value (PPV) and precision of tumor specimens and plasma were assessed across the reportable range of the assay. For measurement of TMB，NGS libraries of a cohort of tumor samples were tested by whole exosome panel (WES) and this assay respectively. And clinical tumor tissue samples whose MSI status were identified by fluorescent PCR-capillary electrophoresis were also tested by this assay. 

Results: The SNV/Indel LOD95 was 0.6% for hot-spot variants, and 1.3% for panel-wised variants of tumor tissue samples with high specificity (PPV >99%). The SNV/Indel LOD95 was 0.4% for hot-spot variants, and 1.1% for panel-wised variants of plasma samples (cfDNA≥15ng), with PPV >99%. LOD95 of CNV was 2.2-2.3, with PPV>99%. Precision of CNV was higher than 95%, precision of SNV/Indel was higher than 97%. Concordance between TMB results tested by WES and this assay was 0.94. MSI results identified by this assay was the same as fluorescent PCR-capillary electrophoresis method. 

Conclusions: In summary, we present the analytical and clinical validation of a comprehensive NGS-based diagnostic assay for comprehensive tumor genomic profiling.