David H Hwang1, Elizabeth P Garcia2, Matthew D Ducar3, Edmund S Cibas1, Lynette M Sholl1,2. 1. Department of Pathology, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts. 2. Center for Advanced Molecular Diagnostics, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts. 3. Center for Cancer Genome Discovery, Dana-Farber Cancer Institute, Boston, Massachusetts.
Abstract
BACKGROUND: Next-generation sequencing (NGS) fails for many small biopsies (BXs) because of a low overall DNA concentration or tumor percentage. Cytology smears and liquid-based preparations (LBPs), or smears/LBPs, often contain abundant tumor cells and may provide adequate material for molecular testing when other materials are insufficient. This study examined the performance of smears/LBPs on a clinical NGS assay. METHODS: This study retrospectively reviewed quality metrics from consecutive smear/LBP, core BX, and cell block (CB) cases run on a hybrid-capture NGS assay interrogating 309 cancer-related genes. The following quality metrics were compared: adequacy rate, initial DNA concentration, postshearing fragment size, post-library preparation fragment size, fragment size difference, insert size, total reads, passing-filter reads aligned, percent passing-filter unique reads aligned, mean target coverage, percentage of loci with >100× coverage, percent duplication rate, percent selected bases, and percent usable bases on bait. RESULTS: Twenty-three of 26 smears/LBPs (88%) were successfully sequenced, whereas 77 of 87 core BXs (89%) and 29 of 30 CBs (97%) were. The mean target coverage, median insert size, and percent usable bases were significantly higher in the smear/LBP category. The postshearing fragment size and the percent duplication were significantly lower for smears/LBPs. CONCLUSIONS: The adequacy rate of cytology smears/LBPs for NGS is comparable to that of core BXs or CBs. Increased values for the mean insert size, mean target coverage, and percent usable bases, along with a lower duplication rate, suggest that smears/LBPs provide higher quality DNA than formalin-fixed material. Cytology smears/LBPs can serve as a valuable source of material for molecular testing. Cancer Cytopathol 2017;125:786-94.
BACKGROUND: Next-generation sequencing (NGS) fails for many small biopsies (BXs) because of a low overall DNA concentration or tumor percentage. Cytology smears and liquid-based preparations (LBPs), or smears/LBPs, often contain abundant tumor cells and may provide adequate material for molecular testing when other materials are insufficient. This study examined the performance of smears/LBPs on a clinical NGS assay. METHODS: This study retrospectively reviewed quality metrics from consecutive smear/LBP, core BX, and cell block (CB) cases run on a hybrid-capture NGS assay interrogating 309 cancer-related genes. The following quality metrics were compared: adequacy rate, initial DNA concentration, postshearing fragment size, post-library preparation fragment size, fragment size difference, insert size, total reads, passing-filter reads aligned, percent passing-filter unique reads aligned, mean target coverage, percentage of loci with >100× coverage, percent duplication rate, percent selected bases, and percent usable bases on bait. RESULTS: Twenty-three of 26 smears/LBPs (88%) were successfully sequenced, whereas 77 of 87 core BXs (89%) and 29 of 30 CBs (97%) were. The mean target coverage, median insert size, and percent usable bases were significantly higher in the smear/LBP category. The postshearing fragment size and the percent duplication were significantly lower for smears/LBPs. CONCLUSIONS: The adequacy rate of cytology smears/LBPs for NGS is comparable to that of core BXs or CBs. Increased values for the mean insert size, mean target coverage, and percent usable bases, along with a lower duplication rate, suggest that smears/LBPs provide higher quality DNA than formalin-fixed material. Cytology smears/LBPs can serve as a valuable source of material for molecular testing. Cancer Cytopathol 2017;125:786-94.