322 RISK ESTIMATION MODEL WITH MULTIPLE COMMON GENETIC VARIANTS FOR JAPANESE PROSTATE CANCER

Abstract

Introduction and Objectives: We identified multiple Japanese prostate cancer (PC) susceptibility-associated SNPs in our genome-wide association study (Takata et al., Nature Genetics, 2010). Although each of the genetic markers confers a modest effect, the effect to PC risk could be stronger when the effects of multiple variants are combined. Recently, several genetic tests combining with genetic variants associated with risk of PC are constructed in Caucasians. However, different risk estimation model for other race was desired, since there is large genetic heterogeneity of PC risk among various ethnic populations. To establish risk estimation model for Japanese PC, we examined the multiplicative genetic effect on Japanese PC. Methods: We evaluated genotype of sixteen PC-associated SNPs in a Japanese population using 3,001 cases and 5,415 controls from BioBank Japan project, and established risk estimation model using 16 genotypes with logistic regression analysis. Selected 16 SNPs were indicated in Table. 11 SNPs were previously reported SNPs that were replicated in Japanese PC, and 5 SNPs were our newly identified SNPs. Results: As a result, estimated odds ratio (OR) of each SNP were ranged from 1.02 to 1.32. As expected, the estimated combined ORs increased consistently with increasing risk alleles, the OR being 7.6 fold for men in the top 5% of the distribution compared with men in lowest 5% (Figure). With use of receiver-operating-characteristic curve analysis, the area under the curve (AUC) for the model was 66. 5%. In addition, high reproducibility was achieved when we validated this model with independent group set (689 cases and 749 controls from two institutes, AUC was 67.2%). When we applied the model to PC patients whose PSA was < 10 ng/ul, the estimated probability of PC was 29.6% in the group with genetic high-risk, and 12.5% in the group with low risk. Conclusions: The genetic model moderately distinguished Japanese at high risk for PC from ones at low risk. The combination with the genetic risk estimation and PSA test may be useful for more accurate screening of PC. Result of Logistic regression analysis We identified multiple prostate cancer (PC) susceptibilityassociated SNPs in the Japanese population. Takata et al., Nature Genet, 2010 Newly identified loci associated with prostate cancer susceptibility However, There is genetic heterogeneity of PC risk among various ethnic populations. Establishment of risk estimation model using Japanese PC-associated SNPs. Each of the genetic markers confers a modest effect (Odds Ratio: 1.1-1.8), but the effect to PC risk could be stronger when the effects of multiple variants are combined. Several genetic tests with genetic variants associated with risk of PC are constructed in Caucasians. http://www.decodehealth.com/prostate-cancer