Abstract 4109: Clonal evolution of metastatic colorectal cancer

• Published in: Cancer research (Chicago, Ill.) Ill.), 2015-08, Vol.75 (15_Supplement), p.4109-4109
• Main Authors: Dang, Ha X., Grossman, Julie, White, Brian S., Strand, Matthew, Larson, David E., Walker, Jason, Pittman, Elizabeth, Fleming, Timothy, Goedegebuure, Peter S., Fulton, Robert S., Miller, Christopher A., Griffith, Malachi, Lim, Kian H., Ley, Timothy J., Wilson, Richard K., Mardis, Elaine R., Lockhart, A.Craig, Fields, Ryan C., Maher, Christopher A.
• Format: Article
• Language: English
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.AM2015-4109

Introduction: Death from colorectal cancer (CRC) occurs via sequelae of metastases. Our lack of understanding of the mechanisms driving metastatic formation is a critical barrier to the identification and direct targeting of critical genes and pathways. This is further complicated by tumor heterogeneity and subclonal architecture. To reconstruct the patterns of tumor evolution and metastasis in CRC, we have conducted the first comprehensive clonality analysis of ten patients. Methods: Primary tumor, metastases in multiple liver segments, and matched normal tissues were procured from consented patients during operative resection. Deep exome (∼200x coverage) and whole genome sequencing (∼50x coverage) were used to identify somatic mutations and estimate variant allele frequency (VAF) for somatic single nucleotide variants (SNVs). Clonal architecture and evolution models were derived from the SNVs by VAF-based clustering, clonal ordering, and phylogeny analysis. Results: Non-silent somatic alterations were enriched in genes known to be involved in CRC and other major cancers, including APC, TP53, KRAS, PIK3CA and TCF7L2. Each patient had a founding clone originating from the primary tumor (carrying non-silent mutations in at least one cancer driver gene) that survived to metastasis, possibly following evolution and acquisition of additional somatic mutations. Branched evolution was common and spatially-distinct liver metastases within the same patient sometimes arose from different (sub)clones in the primary tumor. Unique subclones appeared to arise in all metastatic samples, and in some cases, were shared among various metastases of the same patient. This suggests that one metastasis seeded another or an ancestor common to those metastases was present in the primary tumor or elsewhere, but not observed due to spatial heterogeneity. In several cases, mutations in the dominant clone of the primary tumor were absent from metastases, suggesting these were subclonal events in more aggressive cancer cells that arose in the primary tumor after metastasis. These additional somatic events may involve (possibly novel) cancer driver genes. Conclusions: Understanding the genomic events driving tumor evolution and metastasis is critical for explaining why existing therapies fail and determining optimal treatment strategies. Our analyses have outlined several clonal evolution patterns in metastatic CRC. We are currently using ultra-deep targeted and multi-region sequencin