Representative Sequencing: Unbiased Sampling of Solid Tumor Tissue
Kevin Litchfield, Stacey Stanislaw, Lavinia Spain, Lisa L. Gallegos, Andrew Rowan, Désirée Schnidrig, Heidi Rosenbaum, Alexandre Harlé, 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 B.K. Watkins, Emilia Lim, Nicolai J. Birkbak, Philippe Lamy, Iver Nordentoft, Lars Dyrskjøt, Lisa Pickering, Stephen Hazell, Mariam Jamal-Hanjani, Chris Abbosh, Kai‐Keen Shiu, John Bridgewater, Daniel Hochhauser, Martin Förster, Siow Ming Lee, Tanya Ahmad, Dionysis Papadatos-Pastos, Sam M. Janes, Peter Van Loo, Katey S.S. Enfield, Nicholas McGranahan, Ariana Huebner, Sergio A. Quezada, Stephan Beck, Peter J. Parker, Henning Walczak, Tariq Enver, Robert E. Hynds, Mary Falzon, Ian Proctor, Ron Sinclair, Chi-wah Lok, Zoe Rhodes, David Moore, Teresa Marafioti, Elaine Borg, Miriam Mitchison, Reena Khiroya, Giorgia Trevisan, Peter Ellery, Mark Linch, Sebastian Brandner, Crispin T. Hiley, Selvaraju Veeriah, Maryam Razaq, Heather Shaw, G. Attard, Mita Afroza Akther, Cristina Naceur‐Lombardelli, Lizi Manzano, Maise Al-Bakir, Simranpreet Summan, Nnenna Kanu, Sophia Ward, Uzma Asghar, Emilia Lim, Faye Gishen, Adrian Tookman, Paddy Stone, Caroline Stirling, Andrew Furness, Kim Edmonds, Nikki Hunter, Sarah Sarker, Sarah Vaughan, Mary Mangwende, Karla Pearce, Lavinia Spain, Scott T.C. Shepherd, Haixi Yan, Benjamin Shum, Eleanor Carlyle, Steve Hazell, Annika Fendler, Fiona Byrne, Nadia Yousaf, Sanjay Popat, Olivia Curtis, Gordon Stamp, Antonia Toncheva, Emma Nye
Abstract
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.