Presentation Authors: Mark Preston*, Kathryn Wilson, Boston, MA, Travis Gerke, Tampa, FL, Maxine Chen, Adam Kibel, Lorelei Mucci, Kathryn Penney, Boston, MA
Introduction: Baseline PSA during midlife predicts lethal prostate cancer (PCa) with accuracy and can be used for risk-stratified screening. We sought to investigate how adjustment of baseline PSA levels for single nucleotide polymorphisms (SNPs) related to circulating PSA levels might improve prediction of aggressive PCa.
Methods: We used the Physicians Health Study (PHS) and the Southern Community Cohort Study (SCCS) to identify men aged 40-70 yrs who had baseline PSA and SNP data. In the PHS, there were 392 white men with maximum 30 yr follow-up consisting of 145 aggressive PCa cases (Gleason >6, metastatic or fatal) with 121 controls plus 126 non-aggressive PCa cases as comparison. In the SCCS, there were 319 black men with maximum 12 yr follow-up consisting of 211 PCa cases and 108 controls. Baseline PSA>10 ng/ml were excluded. Median and percentile PSA groupings were established from among controls as in prior publications. In PHS, median PSA was 1.10, 75th percentile was 2.2, and 90th was 3.40. In SCCS, median PSA was 0.9, 75th percentile was 1.4 and 90th was 2.8. We created both unweighted and weighted genetic risk scores (GRS) based on SNPs identified to be associated with PSA in the Hoffman et al. Nat Genet. 2017 GWAS of PSA levels. GRS was used to adjust PSA values and then we assessed logistic regression models and AUC results.
Results: In PHS, compared to men with PSA 75th was OR 3.1 (95%CI 1.8, 5.3. When including the GRS, the adjusted OR for men with PSA > 75th improved to OR 3.4 (95%CI 1.9, 5.9) compared to men with a PSA 75th was OR 27.9 (95%CI 12.0, 64.9). The GRS adjusted OR for men with PSA > 75th improved to OR 30.3 (95%CI 12.8, 71.7) compared to a PSA
Conclusions: We found that the GRS adjusted baseline PSA modestly improved prediction of total and aggressive prostate cancer especially in men with moderately elevated PSA levels. It is possible that adjustment for genetic determinants of PSA may remove variation in levels unrelated to future cancer risk.
Source of Funding: Prostate Cancer Foundation