Prediction of clinically significant prostate cancer using polygenic risk models in Asians.

PURPOSE: To develop and evaluate the performance of a polygenic risk score (PRS) constructed in a Korean male population to predict clinically significant prostate cancer (csPCa).
MATERIALS AND METHODS: Total 2,702 PCa samples and 7,485 controls were used to discover csPCa susceptible single nucleotide polymorphisms (SNPs). Males with biopsy-proven or post-radical prostatectomy Gleason score 7 or higher were included for analysis. After genotype imputation for quality control, logistic regression models were applied to test association and calculate effect size. Extracted candidate SNPs were further tested to compare predictive performance according to number of SNPs included in the PRS. The best-fit model was validated in an independent cohort of 311 cases and 822 controls.
RESULTS: Of the 83 candidate SNPs with significant PCa association reported in previous literature, rs72725879 located in PRNCR1 showed the highest significance for PCa risk (odds ratio, 0.597; 95% confidence interval [CI], 0.555-0.641; p=4.3×10-45). Thirty-two SNPs within 26 distinct loci were further selected for PRS construction. Best performance was found with the top 29 SNPs, with AUC found to be 0.700 (95% CI, 0.667-0.734). Males with very-high PRS (above the 95th percentile) had a 4.92-fold increased risk for csPCa.
CONCLUSIONS: Ethnic-specific PRS was developed and validated in Korean males to predict csPCa susceptibility using the largest csPCa sample size in Asia. PRS can be a potential biomarker to predict individual risk. Future multi-ethnic trials are required to further validate our results. © The Korean Urological Association, 2022.

INTRODUCTION

The incidence of prostate cancer (PCa) is constantly on the rise, currently second in cancer diagnosis and sixth in cancer mortality in men worldwide as of 2020 [1]. This trend is more apparent in Asia where PCa has traditionally had low prevalence, with literature pointing to the gradual transition to a westernized high-fat intake diet and obesity increase as well as broader implementation of nationwide PCa screening as notable factors [2]. As such, genetic origins of PCa pathogenesis have been of keen clinical interest, and with the approval of PARP (poly [adenosine diphosphate-ribose] polymerase) inhibitors as novel therapy for actionable genetic mutations including BRCA 1/2, the National Comprehensive Cancer Network (NCCN) now recommends germline testing for PCa not only in men with a positive family history but also in high- to very-high risk regional or metastatic PCa. Alterations in homologous DNA repair and mismatch repair genes are strongly linked to 3- to 6-fold increase in overall risk [3].

However, not all PCa can be explained with nominal genes of high penetrance, but rather can be attributed to the cumulative effect of small variant alleles that confer varying levels of individual risk to disease [34]. With the pivotal study that identified 5 singular single nucleotide polymorphisms (SNPs) having strong association with metastatic PCa [5], the search for PCa susceptible variants have culminated in over 100 SNPs found in large-scale genome-wide association studies (GWASs), attributing to 60% of heritability [6]. Associations from germline mutations provide valuable information in predicting lifetime trajectory, allowing individualized medical care at an early stage of screening and diagnosis. Polygenic risk score (PRS) takes advantage of common polymorphisms to assess a person’s genetic risk of disease, and when used with other clinical variables, has added value in anticipating tumor aggressiveness and escaping unnecessary biopsy [78].

Current literature on PRS in PCa are scarce compared to the vast number of reported SNPs with variable levels of performance, limiting application in clinical practice. Most large scale models are also constructed in males of mainly Caucasian and European ancestry, resulting in lower predictive power when replicated in Asian cohorts likely due to non-shared genomic loci [4]. As such, ethnic-specific PRS is required to account for genetic variations. In this study, we developed and validated a PRS model predicting clinically significant PCa (csPCa) in a Korean population.

MATERIALS AND METHODS

1. Study population and genotyping

Participants of this study include PCa patients who received radical prostatectomy (RP) and/or core transrectal ultrasound prostate biopsy at one of the four tertiary hospitals in the Republic of Korea (Seoul National University of Bundang Hospital, Seoul National University Hospital, Chungbuk National University Hospital, and the Catholic University of Korea, St. Vincent’s Hospital). All participants provided informed consent. This multi-center study was approved by the Institutional Review Board of Seoul National University Bundang Hospital (SNUBH) (approval number: B-1607/355-302). Gleason score (GS) of PCa was reviewed based on biopsy and/or RP specimens. For selection of csPCa cases, patients with GS 7 or higher were included for this study. Intact genomic DNA (200 ng) extracted from blood or saliva samples was genotyped using the Korea Biobank array (K-CHIP) following the manufacturer’s instructions.

The control genotype and phenotype data were obtained from the Korean Genome and Epidemiology Study (KoGES) conducted by the Center for Genome Science, the Korea National Institute of Health (KNIH) [9]. The genotype data were produced using K-CHIP array. The control samples were further selected for analyses who are male, 60 years or older and had never been diagnosed with any cancer. More detailed information about the external validation cohort is available in a previously published article [9].

2. Quality control of genotype data and genotype imputation

All sample and marker quality control (QC) was performed using PLINK software (v1.90 beta) [10]. Samples were excluded based on the following criteria: i) low genotype call rate (<95%), ii) excessive heterogeneity, iii) genetic relatedness, and iv) sex inconsistencies. Markers i) that are not SNPs, ii) MAF <5%, iii) low call rate (<95%), or iv) significantly deviated from Hardy–Weinberg equilibrium (p<1.0×10-6) were excluded. Imputation of missing genotypes in the PCa cases were conducted using the Michigan Imputation Server which utilizes Eagle v2.4 for phasing and Minimac 4 with the 1000 Genomes Project Phase 3 as the reference genome panel for genotype imputation [1112]. Genotypeimputed SNPs were filtered with genotype quality (R2>0.8) for the further analyses. For control samples, the genotype data that had been imputed were provided by KNIH, after phasing using Eagle v2.3 and genotype imputation using IMPUTE 4 with 1000 Genomes project phase 3 and Korean reference genome (397 samples) as the reference panel [1112]. The imputed SNPs were filtered based on INFO score >0.8 and MAF >0.1. The imputed genotype data for cases and controls were merged after SNPs commonly found in both cases and controls were kept for the further analyses. The post-imputation QC was performed based on the same criteria applied in the pre-imputation QC. After sample and marker QC, the resulting 11,320 samples (3,013 cases and 8,307 controls) and 4,724,872 SNPs were included for the downstream analyses.

3. Statistical analysis

A total of 10,187 samples (2,702 cases and 7,485 controls) were used for discovery of PCa-associated SNPs and their summary statistics. We performed association tests between csPCa and previously reported PCa-associated SNPs. Previously reported SNPs were extracted after literature review of GWAS studies on PCa. Logistic regression models were applied to test associations for each of those SNPs. Based on the association results, SNPs were filtered based on statistical significance of p<0.001. Linkage disequilibrium (LD)-clumping was performed to extract a set of lead SNPs within each LD block.

Based on the association results, we obtained odds ratio (OR) of PCa-associated SNPs on the risk of PCa development. PRS was calculated as log(OR)-weighted sum of the number of risk alleles as following: β

The individual PRS was computed upon the training set for internal validation and the test set for external validation.

To test performance of PRS models in predicting risk of PCa development, the test set (311 cases and 822 controls) was used. The predictive performance of PRS was evaluated using the area under the ROC (receiver operating characteristics) curve (AUC) [13]. The AUC was compared according to the number of SNPs included for PRS models. Improvement in AUC between ROC curves was tested using Delong’s method [6]. The optimal PRS cutoff value was determined at the maximal Youden’s Index (J, sensitivity+specificity-1) value.

For calculation of individual risk, we fitted the logistic regression model to obtain the regression coefficient of PRS on PCa risk and calculated the individual risk with OR = exp(βPRS*(PRS, where βPRS is the regression coefficient of PRS on PCa risk, PRS is the ith individual’s PRS, µPRS is the mean PRS of the controls.

RESULTS

The mean age of PCa cases and controls were 67.8 and 64.7 years, respectively. The mean BMI of cases and controls was 24.6 kg/m2 and 24.1 kg/m2, respectively. The mean prostate-specific antigen (PSA) level of PCa cases were 35.1 ng/mL with standard deviation of 209.4 ng/mL. Most of PCa cases had GS 7 (n=2,105), followed by GS 9 or higher (n=497) and GS 8 (n=411) (Table 1).

Table 1

Characteristics of the study population

Variable		All (n=11,320)		Training set (n=10,187)		Test set (n=1,133)
		Case (n=3,013)	Control (n=8,307)	Case (n=2,702)	Control (n=7,485)	Case (n=311)	Control (n=822)
Age (y)		67.8±7.5	64.7±3.6	67.6±7.5	64.7±3.6	68.8±7.7	64.9±3.6
BMI (kg/m2)		24.6±2.7	24.1±2.7	24.6±2.7	24.1±2.7	NA	24.2±2.8
PSA (ng/mL)		35.1±209.4	NA	36.7±221.1	NA	21.3±31.0	NA
Gleason score							-
	7	2,105	-	1,928	-	177	-
	8	411	-	338	-	73	-
	≥9	497	-	436	-	61	-

Values are presented as mean±standard deviation or number only.

BMI, body mass index; PSA, prostate-specific antigen; NA, not available.

We identified 83 SNPs that were previously reported to be PCa-associated in one or more studies (Table 2). rs72725879 located within PRNCR1 was most significantly associated with PCa risk (OR, 0.597; 95% confidence interval [CI], 0.555–0.641; p=4.3×10-45). Other 38 SNPs also showed negative associations between minor allele and PCa risk. Of the 83 SNPs, 23 SNPs were located within 8q24.21 region, spanning PRNCR1, CASC8, PCAT2, PCAT1, LOC105375751, and CCAT2, 8 SNPs were located within HNF1B (17q12) and 5 SNPs were located within 11q13.3. After LD clumping, 31 SNPs located within 26 distinct genomic loci had been extracted as candidate SNPs for PRS construction (Table 3).

Table 2

SNPs associated with development of PCa

SNP	CHR	BP	Minor allele	OR (95% CI)	p-value	Locus
rs72725879	8	128103969	C	0.597 (0.555–0.641)	4.30E-45	PRNCR1
rs4242384	8	128518554	C	1.761 (1.628–1.906)	4.70E-45	8q24.21
rs7843031	8	128533473	T	1.790 (1.650–1.942)	1.10E-44	8q24.21
rs7837688	8	128539360	T	1.784 (1.645–1.935)	3.50E-44	8q24.21
rs1456315	8	128103937	C	0.584 (0.541–0.630)	8.10E-44	PRNCR1
rs4582524	8	128528435	G	1.748 (1.613–1.894)	2.30E-42	8q24.21
rs10090154	8	128532137	T	1.703 (1.570–1.846)	5.10E-38	8q24.21
rs11986220	8	128531689	A	1.690 (1.559–1.832)	4.60E-37	8q24.21
rs13254738	8	128104343	A	0.633 (0.590–0.680)	4.60E-36	PRNCR1
rs1447295	8	128485038	A	1.598 (1.476–1.730)	4.30E-31	CASC8
rs56005245	8	128113426	T	1.435 (1.342–1.533)	1.80E-26	8q24.21
rs1016343	8	128093297	T	1.417 (1.328–1.512)	7.00E-26	PRNCR1/PCAT2
rs12682344	8	128106784	G	1.441 (1.342–1.547)	6.60E-24	8q24.21
rs6983561	8	128106880	C	1.439 (1.341–1.545)	8.80E-24	8q24.21
rs16901979	8	128124916	A	1.413 (1.317–1.516)	6.30E-22	8q24.21
rs10505483	8	128125195	T	1.413 (1.317–1.516)	6.90E-22	8q24.21
rs11263763	17	36103565	G	0.700 (0.650–0.753)	1.00E-21	HNF1B
rs8064454	17	36101586	A	0.700 (0.651–0.753)	1.00E-21	HNF1B
rs11651052	17	36102381	A	0.702 (0.652–0.755)	1.60E-21	HNF1B
rs7501939	17	36101156	T	0.704 (0.654–0.757)	5.10E-21	HNF1B
rs13252298	8	128095156	G	0.712 (0.662–0.765)	2.80E-20	PRNCR1/PCAT2
rs4871009	8	128108416	T	1.367 (1.277–1.465)	4.40E-19	8q24.21
rs4430796	17	36098040	G	0.743 (0.692–0.797)	9.40E-17	HNF1B
rs2005705	17	36096300	A	0.746 (0.695–0.802)	1.20E-15	HNF1B
rs339331	6	117210052	C	0.775 (0.724–0.829)	1.40E-13	RFX6
rs339351	6	117200434	A	0.778 (0.727–0.832)	2.80E-13	RFX6
rs1512268	8	23526463	T	1.276 (1.195–1.363)	3.80E-13	LOC107986930
rs1160267	8	23529521	G	1.272 (1.191–1.358)	7.90E-13	LOC107986930
rs12549761	8	128540776	G	0.762 (0.706–0.823)	3.20E-12	8q24.21
rs11649743	17	36074979	A	0.781 (0.729–0.838)	4.00E-12	HNF1B/LOC105371754
rs10503733	8	23534018	T	1.246 (1.167–1.331)	6.00E-11	LOC107986930
rs11125927	2	62752975	G	1.247 (1.165–1.336)	2.30E-10	2p15
rs140783917	10	122834482	T	0.748 (0.683–0.818)	2.50E-10	LOC105378521
rs7821330	8	23522452	C	1.228 (1.150–1.310)	6.30E-10	LOC107986930
rs7489409	13	73716861	C	1.213 (1.139–1.292)	1.50E-9	13q22.1
rs10896449	11	68994667	G	1.404 (1.253–1.574)	5.30E-9	11q13.3
rs7463326	8	128027954	A	0.788 (0.723–0.858)	4.40E-8	PCAT1/LOC105375751
rs7929962	11	68985583	T	1.375 (1.226–1.543)	5.70E-8	11q13.3
rs376592364	11	69011693	T	1.383 (1.229–1.557)	7.60E-8	11q13.3
rs10086908	8	128011937	C	0.796 (0.732–0.867)	1.30E-7	LOC105375751
rs8023793	15	66942093	C	0.837 (0.783–0.895)	1.50E-7	LINC01169
rs12270641	11	69012244	A	1.374 (1.219–1.549)	2.10E-7	11q13.3
rs2659124	19	51354597	A	0.845 (0.792–0.902)	4.30E-7	LOC105372441
rs58235267	2	63277843	C	0.843 (0.788–0.903)	8.90E-7	OTX1
rs10993994	10	51549496	T	1.161 (1.091–1.235)	2.40E-6	MSMB
rs11228583	11	69009114	T	1.348 (1.190–1.527)	2.60E-6	11q13.3
rs77167534	2	173319930	T	0.833 (0.771–0.899)	3.30E-6	ITGA6
rs9600079	13	73728139	T	1.158 (1.087–1.233)	5.00E-6	13q22.1
rs7591218	2	43637998	G	0.850 (0.791–0.912)	6.60E-6	THADA
rs2735839	19	51364623	A	0.862 (0.808–0.920)	7.30E-6	19q13.33
rs1983891	6	41536427	T	1.155 (1.083–1.233)	1.30E-5	FOXP4
rs4714485	6	41536587	G	1.154 (1.081–1.231)	1.50E-5	FOXP4
rs2242652	5	1280028	A	0.835 (0.769–0.906)	1.70E-5	TERT
rs11817544	10	80236999	A	0.837 (0.772–0.909)	2.00E-5	10q22.3
rs12621278	2	173311553	G	0.849 (0.787–0.916)	2.30E-5	ITGA6
rs146618443	2	173309803	T	0.851 (0.789–0.918)	3.10E-5	ITGA6
rs1038822	2	43738173	C	0.863 (0.804–0.926)	4.40E-5	THADA
rs75204040	2	62780325	G	1.141 (1.070–1.215)	5.00E-5	2p15
rs10505477	8	128407443	A	1.139 (1.069–1.213)	5.60E-5	CASC8
rs6983267	8	128413305	G	1.138 (1.068–1.212)	6.00E-5	CASC8/CCAT2
rs817872	9	110144887	C	1.143 (1.070–1.221)	7.20E-5	LOC107987111, RAD23B
rs1465618	2	43553949	C	0.866 (0.806–0.930)	8.00E-5	THADA
rs56159348	11	76267331	G	0.836 (0.764–0.915)	1.00E-4	11q13.5, EMSY
rs74634457	15	66835704	G	1.202 (1.095–1.320)	1.10E-4	ZWILCH
rs4554825	10	80244623	T	0.853 (0.787–0.925)	1.20E-4	10q22.3
rs6660538	1	163295678	C	0.878 (0.821–0.940)	1.70E-4	NUF2
rs2252004	10	122844709	A	0.872 (0.812–0.937)	2.00E-4	10q26.12
rs4711748	6	43694598	T	1.125 (1.057–1.197)	2.10E-4	POLR1C, RP1-261G23.5
rs2659051	19	51345568	C	0.882 (0.825–0.943)	2.20E-4	LOC105372441
rs2238776	22	19757892	A	0.889 (0.835–0.947)	2.40E-4	TBX1
rs10807290	6	43710381	T	0.889 (0.835–0.947)	2.40E-4	POLR1C
rs6955627	7	92577760	T	0.879 (0.820–0.942)	2.60E-4	7q21.2
rs2660753	3	87110674	T	1.135 (1.060–1.216)	2.90E-4	3p12.1
rs17023900	3	87134800	G	1.163 (1.070–1.264)	3.80E-4	3p12.1
rs1283104	3	106962521	G	1.119 (1.051–1.192)	4.70E-4	DUBR
rs7153648	14	61122526	C	1.149 (1.062–1.244)	5.60E-4	14q23.1
rs9284813	3	87152169	G	1.130 (1.054–1.211)	6.10E-4	LINC00506
rs3110641	17	36047417	A	1.124 (1.051–1.202)	6.90E-4	HNF1B/LOC105371755
rs143745027	3	87144017	G	1.155 (1.062–1.256)	7.80E-4	LINC00506
rs8005621	14	61106699	G	1.160 (1.064–1.266)	7.90E-4	SALRNA1
rs12500426	4	95514609	A	1.113 (1.045–1.185)	8.00E-4	PDLIM5
rs6545977	2	63301164	A	0.881 (0.817–0.949)	8.50E-4	2p15
rs56103503	1	154980351	C	0.849 (0.771–0.936)	9.50E-4	ZBTB7B

SNP, single nucleotide polymorphism; PCa, prostate cancer; CHR, chromosome; BP, base pairs; OR, odds ratio; CI, confidence interval.

Table 3

Candidate SNPs for PRS construction

SNP	CHR	BP	>Minor allele	β (95% CI)	p-value	SNPs in LD block	Locus
rs72725879	8	128103969	C	-0.517 (-0.589 to -0.445)	4.30E-45	rs1016343, rs13252298, rs1456315, rs13254738, rs12682344, rs6983561, rs4871009, rs56005245, rs16901979, rs10505483	PRNCR1
rs4242384	8	128518554	C	0.566 (0.487 to 0.645)	4.70E-45	rs1447295, rs4582524, rs11986220, rs10090154, rs7843031, rs7837688, rs12549761	8q24.21
rs11263763	17	36103565	G	-0.357 (-0.431 to -0.284)	1.00E-21	rs3110641, rs11649743, rs2005705, rs4430796, rs7501939, rs8064454, rs11651052	HNF1B
rs339331	6	117210052	C	-0.255 (-0.323 to -0.188)	1.40E-13	rs339351	RFX6
rs1512268	8	23526463	T	0.244 (0.178 to 0.310)	3.80E-13	rs7821330, rs1160267, rs10503733	LOC107986930
rs11125927	2	62752975	G	0.221 (0.153 to 0.290)	2.30E-10	rs75204040	2p15
rs140783917	10	122834482	T	-0.291 (-0.381 to -0.201)	2.50E-10	rs2252004	LOC105378521
rs7489409	13	73716861	C	0.193 (0.130 to 0.256)	1.50E-9	rs9600079	13q22.1
rs10896449	11	68994667	G	0.339 (0.226 to 0.454)	5.30E-9	rs7929962, rs11228583, rs376592364, rs12270641	11q13.3
rs7463326	8	128027954	A	-0.239 (-0.324 to -0.153)	4.40E-8	rs10086908	PCAT1/LOC105375751
rs8023793	15	66942093	C	-0.178 (-0.244 to -0.111)	1.50E-7	rs74634457	LINC01169
rs2659124	19	51354597	A	-0.168 (-0.233 to -0.103)	4.30E-7	rs2659051, rs2735839	LOC105372441
rs58235267	2	63277843	C	-0.171 (-0.239 to -0.103)	8.90E-7	rs6545977	OTX1
rs10993994	10	51549496	T	0.149 (0.087 to 0.211)	2.40E-6	-	MSMB
rs77167534	2	173319930	T	-0.183 (-0.260 to -0.106)	3.30E-6	rs146618443, rs12621278	ITGA6
rs7591218	2	43637998	G	-0.163 (-0.234 to -0.092)	6.60E-6	rs1465618, rs1038822	THADA
rs1983891	6	41536427	T	0.144 (0.080 to 0.209)	1.30E-5	rs4714485	FOXP4
rs2242652	5	1280028	A	-0.180 (-0.262 to -0.098)	1.70E-5	-	TERT
rs11817544	10	80236999	A	-0.177 (-0.259 to -0.096)	2.00E-5	rs4554825	10q22.3
rs10505477	8	128407443	A	0.130 (0.067 to 0.193)	5.60E-5	rs6983267	CASC8
rs817872	9	110144887	C	0.134 (0.068 to 0.200)	7.20E-5	-	LOC107987111, RAD23B
rs56159348	11	76267331	G	-0.179 (-0.269 to -0.089)	1.00E-4	-	11q13.5, EMSY
rs6660538	1	163295678	C	-0.130 (-0.197 to -0.062)	1.70E-4	-	NUF2
rs4711748	6	43694598	T	0.118 (0.055 to 0.180)	2.10E-4	rs10807290	POLR1C, RP1-261G23.5
rs2238776	22	19757892	A	-0.118 (-0.181 to -0.055)	2.40E-4	-	TBX1
rs6955627	7	92577760	T	-0.129 (-0.198 to -0.060)	2.60E-4	-	7q21.2
rs2660753	3	87110674	T	0.127 (0.058 to 0.196)	2.90E-4	rs17023900, rs143745027, rs9284813	3p12.1
rs1283104	3	106962521	G	0.112 (0.050 to 0.176)	4.70E-4	-	DUBR
rs7153648	14	61122526	C	0.139 (0.060 to 0.218)	5.60E-4	rs8005621	14q23.1
rs12500426	4	95514609	A	0.107 (0.044 to 0.170)	8.00E-4	-	PDLIM5
rs56103503	1	154980351	C	-0.164 (-0.261 to -0.067)	9.50E-4	-	ZBTB7B

SNP, single nucleotide polymorphism; PRS, polygenic risk score; CHR, chromosome; BP, base pairs; CI, confidence interval; LD, linkage disequilibrium.

We calculated individual PRS in the test set (311 cases and 822 controls) and tested the performance in predicting PCa risk (Fig. 1). PRS models composed of top 29 SNPs, top 26 SNPs and top 28 SNPs showed comparably superior performance compared to other models (Table 4). The AUC of the prediction model of PRS composed of 29 SNPs was estimated to be 0.700 (95% CI, 0.667–0.734), with a sensitivity and specificity of 0.672 and 0.662, respectively (Fig. 2A).

Fig. 1

Distribution of PRS in cases and controls. Mean PRS in controls: pink dashed line, Mean PRS in cases: blue dashed line, and optimal PRS cutoff: black solid line. PRS, polygenic risk score.

Table 4

Predictive performance of PRS models for development of PCa

Number of top SNPs included for PRS calculation	Mean PRS of controls	Mean PRS of cases	AUC (95% CI)	J	Sensitivity	Specificity	Improvement in AUC (p)
29	-0.009	0.000	0.700 (0.667–0.734)	0.334	0.672	0.662	-
26	-0.014	-0.004	0.700 (0.667–0.734)	0.321	0.659	0.662	0 (0.993)
28	-0.010	-0.001	0.700 (0.666–0.733)	0.320	0.637	0.684	-0.001 (0.726)
27	-0.012	-0.003	0.698 (0.664–0.732)	0.323	0.646	0.676	-0.002 (0.411)
31	-0.008	0.001	0.697 (0.664–0.731)	0.313	0.598	0.715	-0.003 (0.263)
30	-0.007	0.002	0.697 (0.664–0.731)	0.314	0.675	0.639	-0.003 (0.115)
25	-0.013	-0.003	0.696 (0.663–0.730)	0.337	0.640	0.697	-0.004 (0.350)
17	-0.017	-0.004	0.692 (0.658–0.726)	0.310	0.688	0.622	-0.008 (0.292)
24	-0.012	-0.001	0.692 (0.658–0.726)	0.315	0.624	0.691	-0.009 (0.065)
18	-0.018	-0.005	0.690 (0.656–0.724)	0.319	0.666	0.653	-0.010 (0.173)
20	-0.015	-0.003	0.690 (0.656–0.724)	0.331	0.656	0.675	-0.011 (0.106)
22	-0.013	-0.002	0.689 (0.655–0.723)	0.315	0.675	0.640	-0.012 (0.035)
21	-0.012	-0.001	0.688 (0.654–0.722)	0.311	0.630	0.681	-0.012 (0.420)
13	-0.024	-0.008	0.687 (0.654–0.721)	0.305	0.588	0.717	-0.013 (0.168)
16	-0.021	-0.007	0.687 (0.653–0.722)	0.298	0.569	0.729	-0.013 (0.110)
23	-0.014	-0.004	0.687 (0.653–0.721)	0.312	0.624	0.689	-0.013 (0.013)
19	-0.019	-0.007	0.687 (0.653–0.721)	0.316	0.675	0.641	-0.013 (0.059)
12	-0.022	-0.004	0.684 (0.650–0.718)	0.309	0.643	0.665	-0.016 (0.094)
15	-0.019	-0.005	0.684 (0.650–0.718)	0.290	0.540	0.749	-0.016 (0.060)
14	-0.018	-0.003	0.683 (0.649–0.718)	0.293	0.572	0.720	-0.017 (0.061)
11	-0.018	0.000	0.683 (0.649–0.717)	0.292	0.656	0.636	-0.018 (0.090)
10	-0.013	0.006	0.680 (0.646–0.714)	0.284	0.621	0.663	-0.020 (0.064)
8	-0.014	0.009	0.675 (0.641–0.709)	0.283	0.637	0.646	-0.026 (0.030)
5	-0.040	-0.006	0.675 (0.640–0.709)	0.260	0.653	0.607	-0.026 (0.060)
7	-0.028	-0.002	0.674 (0.639–0.708)	0.264	0.778	0.485	-0.027 (0.027)
6	-0.024	0.005	0.672 (0.638–0.707)	0.270	0.595	0.675	-0.028 (0.029)
4	-0.070	-0.031	0.670 (0.636–0.705)	0.285	0.675	0.609	-0.030 (0.042)
9	-0.010	0.011	0.669 (0.635–0.704)	0.273	0.666	0.607	-0.031 (0.006)

PRS, polygenic risk score; PCa, prostate cancer; SNP, single nucleotide polymorphism; AUC, area under the receiver operating characteristics curve.

Fig. 2

(A) ROC curve of PRS models in prediction of PCa risk. (B) Distribution of PRS in cases and controls with two cutoff values to categorize PRS risk groups into moderate, high and very-high risk group. (C) Risk of development of PCa by PRS risk group. AUC, area under the receiver operating characteristics curve; OR, odds ratio; ROC, receiver operating characteristic; PRS, polygenic risk score; PCa, prostate cancer.

From the best performing PRS model composed of top 29 SNPs, we chose two different cutoff values, i) the optimal PRS value at the maximum Youden’s index (-0.05) and ii) 95th percentile PRS cutoff (0.02) (Fig. 2B). We defined those with PRS below the optimal cutoff, those with PRS above 95th percentile cutoff and those with PRS between the two cutoff values as moderate, very-high and high-risk group, respectively. High PRS and very-high PRS group showed 3.80-fold and 5.93-fold increased risk, respectively, for PCa, compared to the moderate PRS group (the reference group) (Fig. 2C). The very-high PRS group had a 4.92-fold increased risk compared to the remaining population.

DISCUSSION

Based on the OR estimates of each SNP in PCa specific to a Korean population, we developed and evaluated PRS in Korean males. Of the 83 SNPs identified with significant association, 31 SNPs were selected for PRS construction. The best performing PRS composed of 29 SNPs reported an AUC of 0.700, with the top 5th percentile having more than a 4-fold increased risk of Grade Group (GG) ≥2 PCa. To note, minor alleles of many SNPs included in this study showed protective effects on PCa risk, which highlights that common risk alleles may have a cumulative impact in the development of PCa.

With trends of increasing prevalence worldwide, familial PCa and heritability is thoroughly researched for potential identification of actionable mutations. Hereditary PCa is caused by inherited genes of high penetrance, including homologous DNA pair genes (e.g., BRCA1/2, ATM, CHEK2, PALB2) and mismatch repair genes (e.g., MLH1, MSH2, MSH6, PMS2), the latter related to Lynch syndrome. While such nominal mutations are easily identified and linked to PCa risk, not all familial PCa can be explained by singular mutations, but rather is accountable by the sum of multiple common variants with small effect size. PRS is a novel and only biomarker that uses the concept of polygenic risk as a single combinatory value estimate of the individual’s genetic propensity of a complex disease to potentially stratify high- and low-risk patients in order to apply different methods of screening and prevention.

While the exact number required to achieve the full potential of PRS is unknown, summary statistics based on sufficient data is required to detect rare but significant variants shared between phenotypes and approximate the populational burden of disease. PCa is reported to have variants with relatively larger effect size and hence has been suggested to require a smaller sample to assess heritability compared to other, more polygenic cancers. While falling short of over 60,000 cases required to explain 80% of heritability as suggested by Zhang et al. (2020) [14], one of the largest cohorts of over 2,700 males in biopsy and RP-proven GS ≥7 PCa was utilized, compared to previous Korean [15] and Japanese [16] PRSs developed from 1,001 and 689 cases, respectively.

Selection of appropriate SNPs for inclusion critically influences PRS performance, due to obvious uncertainties inherent in estimating effect size from a potentially biased population. As such, statistical techniques must be applied prior to PRS development, a process necessary to weed out insignificant variants and to avoid overfitting or under-fitting of risk alleles. p-value thresholds are often implemented as a cut-off, and while p<1×10-5 is most commonly used, it is often an arbitrary value. A strict, parsimonious threshold risks the loss of otherwise powerful variants confounded by nearby SNPs resulting in false negatives, whereas an overly flexible threshold risks inclusion less predictive alleles as well as a resulting SNP combination too large to be feasible in actual practice. Hence, to optimize SNP selection, a relative lenient p-value threshold of p<10-3 was applied to filter a relatively large set of potential candidates, followed by LD clumping to extract lead SNPs and quality control via imputation for missing genotypes as well as SNP combination comparisons, a similar approach described in our previous report [15].

To additionally bolster performance, our study selectively used previously reported, well-established PCa-susceptible SNPs, many found in the 8q24.21 region. 8q24 is commonly considered a “desert” region due to scarcity of genes, but has been the source of numerous PCa susceptible variants, contributing to 8% to 9% of two-fold increase in familial risk [17]. A proto-oncogene Myc is closest, and while polymorphisms in this region have been associated with regulatory enhancers and androgen receptor (AR) response, mechanisms of action are not yet well understood. Nonetheless, strong associations have been made in both European and East Asian populations [1819], with rs7837688 and rs1016343 more common in Korean males [15], a finding comparable in this study as 4th and 12th in statistical significance. Eleven candidate SNPs were associated with PRNCR1 in the 8q24 region, whose upregulation is implicated in aggressive PCa via AR-mediated transcriptional programs [20]. Thirteen were located in HNF1B (17q12) and 11q13.3, both well-documented as susceptible loci in males of European and Asian descent [2122].

The ultimate goal of this research was to evaluate whether PRS can be utilized in actual clinical practice to avoid unnecessary intervention in low-risk PCa. Therefore, we selected patients with biopsy or pathology-proven csPCa with ≥GG2, excluding males under the age of 60. While generally preferred in the Korean population over active surveillance (AS) or watchful waiting, early intervention and surgery in low-grade PCa risks overtreatment and economic burden. As such, imaging modalities and biomarkers have been introduced to focus biopsy and treatment in csPCa, which generally includes any of histopathology ISUP grade ≥2 (≥GG2) and/or tumor volume ≥0.5 cc [23]. Our study, selecting only males of intermediate- to high-risk PCa category, showed performance comparable to previous literature. Seibert et al. (2018) [7], defining aggressive PCa as any cancer not eligible for surveillance based on the NCCN guidelines (i.e., any ≥GG2, stage T3–4, PSA ≥10 ng/mL, and any nodal or distant metastasis), developed a PRS from 54 SNPs to predict early age at onset of aggressive PCa (z=11.2, p<10-16), with males in the 98th percentile having a 2.9-fold increased risk. Stockholm3 combined a PRS generated from 232 SNPs with 6 plasma proteins and other clinical variables to predict ≥GG2 PCa, outperforming PSA alone with a resulting AUC of 0.74 (95% CI, 0.72–0.75) [8]. Takata et al. (2019) [24] published results from a large Japanese cohort with a development set of 5,088 PCa cases from Biobank Japan and 10,682 controls from multiple institutions, but 1,806 (35.5%) patients had missing GS, and 1,293 (25.4%) had GS ≤6. Our effect size estimates calculated from csPCa better represent genetic polymorphisms that require intervention in actual practice.

Generalizability is key to the performance of a PRS, and population bias from a limited cohort poses a unique challenge in genomic prediction. Previously reported models are generated from cohorts of largely European ancestry and hence often fail to have the same predictive power in Asian males, primarily because unshared genomic loci, as well as variable population-specific effect size. A meta-analysis of multiancestry GWAS data found East Asians to have a 0.73-fold lower risk score compared to their European counterparts, highlighting the probability of potent germline variations among ethnic populations [25]. As such, accurate prediction of PCa risk requires tailoring to each race and ethnicity. Also, with increasing prevalence, even a modest performance is able to substantially stratify absolute populational risk, allowing application into clinical practice of early detection and prevention. AUC of 0.700 in our study is one of the highest discriminatory performances of PRS developed in an Asian cohort to date [4], compared to a previous report by our team conferring an AUC of 0.605 for any PCa utilizing 5 SNPs [15] and those developed in Japanese and Chinese populations with AUCs of 0.60 and 0.659, respectively [1626]. The 4.92-fold increased risk of csPCa in the 95th percentile of our study is comparable to results validated in western counterparts with a 4 to 5-fold increased risk in the 99th percentile [1827].

Despite successful replication of significant PCa susceptible polymorphisms and outperforming some previous models [15162829], our study is not without limitations. First, although we utilized data from 2,702 males with csPCa, it is still a relatively small sample size to accurately represent true populational variation. Second, because both development and validation cohorts consisted of Koreans, future evaluation in other race and ethnicities are required to fully assess predictive performance and generalizability. Due to the inherent selection bias in design, not all significant SNPs may be adequately represented in our data. In addition, while we limited the study population to csPCa, applying our model to males with any PCa at any age may further PRS utility, as both patients and clinicians can have additive information to guide screening schedules and opt for early intervention in AS-eligible males with high genetic risk.

Nonetheless, this study provides valuable evidence for prediction of csPCa based on the cumulative effects of small genetic variants specific to Asian populations. PRS calculated from individual GWAS data is the hallmark of modern precision medicine with the possibility of predicting a patient’s lifetime trajectory for disease, offering the chance for early screening and personalized treatment, as in the case of coronary heart disease [30] where genetics is making great strides in real-world practice. Future acquisition of more sample summary statistics will further potentiate PRS performance in PCa and produce more accurate stratification of risk groups.

CONCLUSIONS

We successfully developed and validated PRS models for csPCa risk developed in a large sample of 2,702 ≥GS7 PCa cases and 7,485 healthy controls. A model utilizing the top 29 SNPs conferred the best predictive performance with an AUC 0.700, with over 4-fold risk predicted in males in above the 95th percentile of PRS. Future prospective, large-scale studies are required to further validate our results.