Detection of Copy-Number Variations in CNS Tumours From Off-Target Reads of Hybrid-Capture Sequencing.

Abstract

Copy number variations (CNVs) play a central role in the classification, grading and prognostication of central nervous system (CNS) tumours. While genome-wide methylation arrays are widely used for CNV profiling, next-generation sequencing (NGS) panels are increasingly implemented in routine diagnostics. We hypothesised that off-target sequencing reads from small hybrid-capture panels not specifically designed for copy-number detection can yield clinically actionable genome-wide CNV profiles. We analysed 60 CNS tumour samples, including glioblastomas, oligodendrogliomas, ependymal tumours, medulloblastomas and choroid plexus tumours using a small-scale custom hybrid-capture panel (< 0,2 Mb) and compared CNV profiles inferred from sequencing reads to those obtained with methylation arrays. Additionally, 58 meningiomas and 6 pilocytic astrocytomas with BRAF fusions were profiled with the same NGS panel. Across 527 chromosomal arm-level alterations, concordance between NGS- and methylation-derived profiles was 95%. All 19 focal amplifications (e.g., EGFR, MDM4 and MYCN) and the majority of homozygous CDKN2A/B deletions were correctly detected. In meningiomas, genome-wide CNV profiling from off-target reads identified WHO-relevant alterations, including CDKN2A/B deletions and 1p/22q co-deletions, supporting molecular upgrading in 9/58 (16%) of histologically lower grade tumours. Focal copy-number variations on Chr7q suggestive of BRAF fusions were observed in 5/6 fusion-positive pilocytic astrocytomas. These findings demonstrate that off-target reads from minimal targeted NGS panels can generate genome-wide CNV profiles, comparable to methylation array data, without the need for additional assays or specialised probe designs.