Evaluating Tumor Heterogeneity in an Evolutionary Context
Elaine R. Mardis1, Malachi Griffith1, Obi Griffith1, Nathan Dees1, Krishna Kanchi1, Li Ding1, Christopher Miller1, Scott Smith1, Chuck Perou2, Lisa Carey2, Adam Kibel4, Peter Humphrey3, Christopher Maher1,3, Matthew J. Ellis3, and Richard K. Wilson1
1The Genome Institute at Washington University School of Medicine, St. Louis MO
2Lineberger Comprehensive Care Center, University of North Carolina at Chapel Hill
3Department of Medicine, Washington University School of Medicine
4Brigham and Women’s Hospital, Dana Farber Cancer Institute, Boston MA
Individual tumors are known to have genomic heterogeneity, namely the collection of cancerous cells comprising the tumor mass possess different yet related genomes. Using high depth NGS data, resulting from directed hybrid capture of variant sites in each genome we can model the genomic heterogeneity of the cancer cells used to isolate DNA for whole genome sequencing. By carefully characterizing related presentations of each patient’s cancer, we can build an understanding of the way different therapies impact the genomic heterogeneity in patients who do, or do not, respond to the therapy. We also can understand the evolutionary changes in heterogeneity that result from progression of pre-neoplastic lesions to invasive cancer. Lastly, we can model how tumor evolution plays out in patients with multiple metastatic lesions following a primary cancer diagnosis. All of these approaches depend on having banked the appropriate specimens, the use of highly sensitive variant detection, and the appropriate interpretation of the high coverage data sets generated for each cancer case. I will present examples of each type of tumor evolution described above, using examples from our whole genome sequencing studies.