| Popis: |
Detection of tumor-derived circulating cell-free DNA (ccfDNA) in plasma from glioblastoma patients remains elusive. The vast intra-tumor genetic heterogeneity of glioblastoma has limited targeted searches using a priori molecular profiling from a focal tissue sample. Recent data supports the isolation of shorter ccfDNA fragments relative to ccfDNA’s principal mononucleosomal peak (< 150 bp vs. 167 bp, respectively) to both enrich for tumor-derived ccfDNA and reduce false positives associated with next-generation sequencing (NGS). Here, we sought to determine if size selection affords detection of glioblastoma-derived ccfDNA. The ccfDNA from 11 healthy controls and 10 glioblastoma patients was molecularly barcoded, PCR amplified, and short ccfDNA fragments (< 150 bp) isolated using an automated gel-based technology. Both the original and size-selected ccfDNA samples were capture enriched using a custom-designed, glioblastoma-targeted NGS panel (128 genes, 128 kb) followed by paired-end sequencing. A consensus sequence was determined for each group of PCR duplicates with an identical barcode. The number of PCR duplicates used to derive a consensus sequence has been termed family size (FS). Non-reference alleles (NRAs) from consensus sequences in exons were tabulated. In healthy controls, there was a 15-fold reduction (P < 0.001) in NRAs (i.e., false positives) in the short ccfDNA fraction at FS ≥ 20 compared to FS ≥ 1. At FS ≥ 20 in the short ccfDNA fraction, there were significantly more NRAs in glioblastoma patients (i.e., potential tumor-derived variants) compared to healthy controls (1,065 ± 406 vs. 174 ± 75 NRAs, respectively; P < 0.001). In glioblastoma patients, there were significantly more NRAs in the short ccfDNA fraction compared to the original unselected ccfDNA sample at FS ≥ 20 (1,065 ± 406 vs. 165 ± 135 NRAs, respectively; P < 0.001). Selection for short ccfDNA fragments coupled with molecular barcoding detects glioblastoma-derived ccfDNA through concomitant improvements in both sensitivity and specificity during untargeted searches employing panel-based NGS. |
