Prevalence of pathogenic/likely pathogenic variants in the 24 cancer genes of the ACMG Secondary Findings v2.0 list in a large cancer cohort and ethnicity-matched controls

• Main Authors: Jung Kim, Wen Luo, Mingyi Wang, Talia Wegman-Ostrosky, Megan N. Frone, Jennifer J. Johnston, Michael L. Nickerson, Melissa Rotunno, Shengchao A. Li, Maria I. Achatz, Seth A. Brodie, Michael Dean, Kelvin C. de Andrade, Fernanda P. Fortes, Matthew Gianferante, Payal Khincha, Mary L. McMaster, Lisa J. McReynolds, Alexander Pemov, Maisa Pinheiro, Karina M. Santiago, Blanche P. Alter, Neil E. Caporaso, Shahinaz M. Gadalla, Lynn R. Goldin, Mark H. Greene, Jennifer Loud, Xiaohong R. Yang, Neal D. Freedman, Susan M. Gapstur, Mia M. Gaudet, Donato Calista, Paola Ghiorzo, Maria Concetta Fargnoli, Eduardo Nagore, Ketty Peris, Susana Puig, Maria Teresa Landi, Belynda Hicks, Bin Zhu, Jia Liu, Joshua N. Sampson, Stephen J. Chanock, Lisa J. Mirabello, Lindsay M. Morton, Leslie G. Biesecker, Margaret A. Tucker, Sharon A. Savage, Alisa M. Goldstein, Douglas R. Stewart
• Format: Article
• Language: English
• Published: BMC 2018-12-01
• Series: Genome Medicine
• Subjects: ACMG secondary findings; Familial cancer exome; Population study; Variant classification
• Online Access: http://link.springer.com/article/10.1186/s13073-018-0607-5

Abstract Background Prior research has established that the prevalence of pathogenic/likely pathogenic (P/LP) variants across all of the American College of Medical Genetics (ACMG) Secondary Findings (SF) genes is approximately 0.8–5%. We investigated the prevalence of P/LP variants in the 24 ACMG SF v2.0 cancer genes in a family-based cancer research cohort (n = 1173) and in cancer-free ethnicity-matched controls (n = 982). Methods We used InterVar to classify variants and subsequently conducted a manual review to further examine variants of unknown significance (VUS). Results In the 24 genes on the ACMG SF v2.0 list associated with a cancer phenotype, we observed 8 P/LP unique variants (8 individuals; 0.8%) in controls and 11 P/LP unique variants (14 individuals; 1.2%) in cases, a non-significant difference. We reviewed 115 VUS. The median estimated per-variant review time required was 30 min; the first variant within a gene took significantly (p = 0.0009) longer to review (median = 60 min) compared with subsequent variants (median = 30 min). The concordance rate was 83.3% for the variants examined by two reviewers. Conclusion The 115 VUS required database and literature review, a time- and labor-intensive process hampered by the difficulty in interpreting conflicting P/LP determinations. By rigorously investigating the 24 ACMG SF v2.0 cancer genes, our work establishes a benchmark P/LP variant prevalence rate in a familial cancer cohort and controls.