Implications of Selection Bias Due to Delayed Study Entry in Clinical Genomic Studies
Samantha Brown, Jessica A. Lavery, Ronglai Shen, Axel Martin, Kenneth L. Kehl, Shawn M. Sweeney, Eva M. Lepisto, Hira Rizvi, Caroline G. McCarthy, Nikolaus Schultz, Jeremy L. Warner, Ben Ho Park, Philippe L. Bédard, Gregory J. Riely, Deborah Schrag, Katherine S. Panageas, Shawn M. Sweeney, Margaret Foti, Yekaterina B. Khotskaya, Michael V. Fiandalo, Benjamin Groß, Nikolaus Schultz, Brooke Mastrogiacomo, Mahdi Sarmardy, Marilyn M. Li, Adam Resnick, Angela J. Waanders, Jena Lilly, Richard D. Carvajal, Raúl Rabadán, Matthew Ingham, Susan Hsaio, Jean Abraham, James D. Brenton, Oscar M. Rueda, Carlos Caldas, Mikel Valgañón, Dilrini De Silva, Chris Boursnell, Raquel Rodríguez-García, Ezequiel Rodriguez, Birgit Nimmervoll, Ethan Cerami, Matthew D. Ducar, Priti Kumari, Neal I. Lindeman, Laura MacConnaill, John A. Orechia, Deborah Schrag, Priyanka Shivdasani, Eliezer M. Van Allen, Jason M. Johnson, Pasi A. Jänne, Eva M. Lepisto, Michael J. Hassett, Sindy Pimentel, Parin Sripakdeevong, Katherine A. Janeway, Jason M. Johnson, Matthew Meyerson, Daniel M. Quinn, Oya Cushing, Kevin M. Haigis, Diana Miller, Kenneth L. Kehl, Alexander Gustav, Angela C. Tramontano, Simon Arango Baquero, Jonathan L. Bell, Michelle Green, Shannon J. McCall, Michael Datto, Fabien Calvo, Fabrice André, Meurice Guillaume, Semih Doğan, Lacroix Ludovic, Jean Scoazec, Monica Ardenos, Gilles Vassal, Stefan Michels, Victor E. Velculescu, Alexander S. Baras, Christopher D. Gocke, Julie R. Brahmer, Charles L. Sawyers, David B. Solit, Stuart M. Gardos, Mike Berger, Marc Ladanyi, Gregory J. Riely, S. Joseph Sirintrapun, Katherine S. Panageas, Ari Caroline, Stacy B. Thomas, Andrew Zarski, Ahmet Zehir, Alexia Iasonosa, John Philip, Samantha Brown
Abstract
IMPORTANCE: Real-world data sets that combine clinical and genomic data may be subject to left truncation (when potential study participants are not included because they have already passed the milestone of interest at the time of study recruitment). The lapse between diagnosis and molecular testing can present analytic challenges and threaten the validity and interpretation of survival analyses. OBSERVATIONS: Effects of ignoring left truncation when estimating overall survival are illustrated using data from the American Association for Cancer Research (AACR) Project Genomics Evidence Neoplasia Information Exchange Biopharma Collaborative (GENIE BPC), and a straightforward risk-set adjustment approach is described. Ignoring left truncation results in overestimation of overall survival: unadjusted median survival estimates from diagnosis among patients with stage IV non-small cell lung cancer or stage IV colorectal cancer were overestimated by more than 1 year. CONCLUSIONS AND RELEVANCE: Clinicogenomic data are a valuable resource for evaluation of real-world cancer outcomes and should be analyzed using appropriate methods to maximize their potential. Analysts must become adept at application of appropriate statistical methods to ensure valid, meaningful, and generalizable research findings.