Quantitative Off-target Detection of Epstein-Barr Virus-derived DNA in Routine Molecular Profiling of Hematopoietic Neoplasms by Panel-based Hybrid-capture Next-Generation Sequencing

Abstract Epstein-Barr virus (EBV) is associated with hematologic and solid tumors. In this study, we utilized a hybridization capture-based next generation sequencing (NGS) platform that targets 400 genes associated with hematological malignancies to detect and quantify non-targeted viral-derived EBV reads that aligned to the EBV reference contig (NC_007605). We evaluated 5,234 samples from 3,636 unique patients with hematological neoplasms and found that 100 samples (1.9%) in 93 unique patients had ≥ 6 EBV reads (range, 6-32,325; mean, 827.5; median, 54). The majority (n=73, 73%) represented known EBV-associated conditions, and the most common was post-transplant lymphoproliferative disorders (n=21, 29%). Documented EBV viremia accounted for a moderate quantity of EBV reads in 4/27 samples corresponding to conditions not known to be EBV-associated, while suspected viremia or low-level activation was likely the etiology in the remaining 23 samples. A good correlation (Spearman r=0.8, 95% CI 0.74-0.85) was found between EBV reads by NGS and systematic semi-quantitative EBER ISH assessment in 162 available samples, particularly at higher level of EBV-involvement. An optimal threshold for significant morphologic EBV-involvement was found to be ≥10 reads by the receiver operating characteristic (ROC) analysis (AUC 0.990 and 95% CI 0.9974-1.000%). Thus, in addition to mutational analysis, hybrid-capture-based NGS panels can be used to detect and quantitate off-target EBV-derived viral DNA, which correlates well with morphology.