Representative sequencing: Unbiased sampling of solid tumor tissue.
journal contributionposted on 2020-11-10, 11:24 authored by Kevin Litchfield, Stacey Stanislaw, Lavinia Spain, Lisa L Gallegos, Andrew Rowan, Desiree Schnidrig, Heidi Rosenbaum, Alexandre Harle, Lewis Au, Samantha M Hill, Zayd Tippu, Jennifer Thomas, Lisa Thompson, Hang Xu, Stuart Horswell, Aoune Barhoumi, Carol Jones, Katherine F Leith, Daniel L Burgess, Thomas BK Watkins, Emilia Lim, Nicolai J Birkbak, Philippe Lamy, Iver Nordentoft, Lars Dyrskjøt, Lisa Pickering, Stephen Hazell, Mariam Jamal-Hanjani, PEACE Consortium, James Larkin, Charles Swanton, Nelson R Alexander, Samra Turajlic
Although thousands of solid tumors have been sequenced to date, a fundamental under-sampling bias is inherent in current methodologies. This is caused by a tissue sample input of fixed dimensions (e.g., 6 mm biopsy), which becomes grossly under-powered as tumor volume scales. Here, we demonstrate representative sequencing (Rep-Seq) as a new method to achieve unbiased tumor tissue sampling. Rep-Seq uses fixed residual tumor material, which is homogenized and subjected to next-generation sequencing. Analysis of intratumor tumor mutation burden (TMB) variability shows a high level of misclassification using current single-biopsy methods, with 20% of lung and 52% of bladder tumors having at least one biopsy with high TMB but low clonal TMB overall. Misclassification rates by contrast are reduced to 2% (lung) and 4% (bladder) when a more representative sampling methodology is used. Rep-Seq offers an improved sampling protocol for tumor profiling, with significant potential for improved clinical utility and more accurate deconvolution of clonal structure.