Abstract 1316: Germline genetic signals across multiple aggressive prostate cancer phenotypes

• Published in: Cancer research (Chicago, Ill.) Vol. 77; no. 13_Supplement; p. 1316
• Main Authors: Tai, Caroline G., Emami, Nima C., Hoffmann, Thomas J., Sakoda, Lori C., Jorgenson, Eric, Habel, Laurel A., Shan, Jun, Ranatunga, Dilrini K., Chao, Chun R., Ghai, Nirupa R., Aaronson, David, Presti, Joseph, Schaefer, Catherine, Risch, Neil, Eeden, Stephen K. Van Den, Witte, John S.
• Format: Journal Article
• Language: English
• Published: 01-07-2017
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.AM2017-1316

Prostate cancer (PrCa) is documented to be highly heritable. But many known prostate cancer loci are not associated with aggressive prostate cancer phenotypes such as high-grade cancer or biochemical failure. Previous efforts have been largely underpowered or not replicated.
We conducted a study of 6,321 men with PrCa and 5,663 controls of European ancestry recruited from Kaiser Permanente (KP) matched by age and self-reported race. Phenotype data combined patient-level demographic, clinical, and tumor data from KP databases with data from 1994 to 2016. High-grade PrCa was defined using Gleason score: low-grade defined as ≤ 7 (3+4), high-grade defined as ≥ 7 (4+3). Biochemical failure was defined for cases using post-treatment prostate specific antigen (PSA) values: post-surgery failure if PSA ≥ 0.4 ng/mL; post-radiation failure if PSA value ≥ 2 ng/mL + nadir; and no failure if PSA did not reach specified threshold. Early-onset disease was defined as age < 55 years at diagnosis. Men were genotyped for over 650,000 SNPs using an ethnic-specific Affymetrix Axiom array and imputed to 1000 Genomes. We tested genetic associations for multiple aggressive phenotypes (410 early-onset cases, 1324 biochemical failure cases, and 886 high grade cases) in both a case-only (e.g., high vs. low grade) and case-control (e.g., high grade vs. no prostate cancer controls) analysis using logistic regression adjusting for ancestry principal components, age, BMI, and kit for batch effects. PrCa specific mortality was too rare to be evaluated.
We identified several known and novel loci. Case-control analyses highlighted known prostate cancer loci at 8q24, 10q11, 11q13, 16q22, and 19q13. Previous reports of case-only analyses also identified 8q24, 20p13 and the gene, KLK3 in 19q13, to be associated with high-grade PrCa; we confirmed these signals in our study. Previous family-based studies identified 15q11 as a familial or early-onset prostate cancer locus.
These aggressive-specific variants could perform better in predicting clinically actionable prostate cancers than variants for overall prostate cancer and can improve the predictive ability of pre-diagnostic PSA.
Citation Format: Caroline G. Tai, Nima C. Emami, Thomas J. Hoffmann, Lori C. Sakoda, Eric Jorgenson, Laurel A. Habel, Jun Shan, Dilrini K. Ranatunga, Chun R. Chao, Nirupa R. Ghai, David Aaronson, Joseph Presti, Catherine Schaefer, Neil Risch, Stephen K. Van Den Eeden, John S. Witte. Germline genetic signals across multiple aggressive prostate cancer phenotypes [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 1316. doi:10.1158/1538-7445.AM2017-1316