Genetic Risk Assessment of Lethal Prostate Cancer Using Polygenic Risk Score and Hereditary Cancer Susceptibility genes

doi:10.21203/rs.3.rs-2446801/v1

Download PDF

Research Article

Genetic Risk Assessment of Lethal Prostate Cancer Using Polygenic Risk Score and Hereditary Cancer Susceptibility genes

https://doi.org/10.21203/rs.3.rs-2446801/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

The genetic risk of aggressive prostate cancer (PCa) is hard to be assessed due to the lack of aggressiveness-related single-nucleotide polymorphisms (SNPs). Prostate volume (PV) is a potential well-established risk factor for aggressive PCa, we hypothesize that polygenic risk score (PRS) based on benign prostate hyperplasia (BPH) or PV-related SNPs may also predict the risk of aggressive PCa or PCa death.

Methods

We evaluated a PRS using 27 BPH/PV-associated SNPs, two established PCa risk-related PRS and 10 guideline-recommended hereditary cancer risk genes in the population-based UK Biobank cohort (N = 209,910).

Results

The BPH/PV PRS was significantly inversely associated with the incidence of lethal PCa as well as the natural progress in PCa patients (hazard ratio, HR = 0.92, 95% confidence interval [CI]: 0.85–0.99, P = 0.03; HR = 0.92, 95%CI: 0.86–0.99, P = 0.02). Compared with men at top 25th PRS, PCa patients with bottom 25th PRS would have a 1.40-fold (HR, 95%CI: 1.16–1.69, P = 0.001) increased PCa fatal risk and shorter survival time at 0.37 year (95%CI: 0.14–0.61, P = 0.002). In addition, patients with BRCA2 or PALB2 pathogenic mutations would also have a high risk of PCa death (HR = 3.91, 95%CI: 2.34–6.51, P < 0.001; HR = 4.24, 95%CI: 1.34–13.34, P = 0.01, respectively). However, no interactive but independent effects were detected between this PRS and pathogenic mutations.

Conclusions

Our findings provide a new measurement of PCa patients’ natural disease outcomes via genetic risk ways.

Benign prostatic hyperplasia

Germline mutation

Polygenic risk score

Prostate cancer

Prostate cancer (PCa) is one of the most common and lethal cancers worldwide. It ranked 3rd in the new cancer cases and 8th in the cancer deaths with approximately 1.4 million new cases and 375,00 deaths in 2020[1]. The common use of prostate-specific antigen (PSA) screening has contributed to early detection of PCa and the increased number of PCa incidence in statistics. However, it might not be able to distinguish clinically aggressive cancer from low-risk cases and has relatively low socioeconomic efficacy without selectively applying[2]. Early detection is critical for high-risk individuals to receive effective intervention at disease’s early stage, especially for men with increased risk of clinically significant PCa (advanced or lethal PCa).

Genetic risk assessment is one of the common tools to identify individuals at risk to develop diseases. It usually contains three components: family history (FH), inherited pathogenic (P/LP) variants in cancer susceptibility genes (such as BRCA2, HOXB13), and single nucleotide polymorphism (SNP)-based polygenic risk scores (PRS). The relationship between genetic risk and aggressive PCa is still under exploration. About ~ 3% of general population were observed with a positive FH but it was not associated with clinically significant PCa[3]. In addition, less than 2% of men in the population and ~ 5% of PCa patients would carry a P/LP variant[4]. P/LP variants in BRCA1/2, ATM, CHEK2 (c.1100 delC) were reported to be associated with PCa aggressiveness[5–9]. NCCN guideline also recommended a basic germline testing in hereditary cancer susceptibility genes if patients had a PCa FH, Ashkenazi Jewish ancestry, a personal history of certain cancers and aggressive PCa[10]. However, the cancer-specific lethal risk of these P/LP variants of cancer susceptibility genes was rarely reported in the incidental PCa patients. PCa-risk-associated SNPs based PRS could identify additional individuals at high PCa risk and the clinically relevant endpoint of PCa death with wide range at population level[6, 11]. But it fail to differentiate indolent PCa and advanced PCa[7]. To date, it is difficult to assess the genetic risk of aggressive PCa or disease outcomes due to the lack of aggressive-associated SNPs. Only 8 mixed-ethnical small-scale GWAS reported 16 SNPs associated with aggressive PCa (10 independent SNPs reached genome-wide significant level and 4 SNPs had been validated)[12]. And the standard in the definition of aggressive PCa and the validation were divided across these GWASs.

Established evidence show that a low prostate volume (PV) is significantly associated with clinically significant PCa[13–15]. However, the change of PV or benign prostatic hyperplasia (BPH) happens most likely in men’s 70s[16]. To evaluate the possibility of an enlarged PV or BPH diagnosis at a younger age, we would use a PRS as an instrument to assess genetic risk of BPH/PV. The hypothesis of the present study is that BPH/PV PRS may predict individuals’ risk of BPH/PV as well as the risk of PCa death. We will evaluate whether it could predict PCa death and provide additional assessment over P/LP variants of hereditary cancer susceptibility genes (ATM, BRCA1, BRCA2, CHEK2, PALB2, MLH1, MSH2, MSH6, PMS2, and HOXB13).

Study Population

The current study was conducted in a large cohort from the UK Biobank (UKB) [17]. PCa diagnosis (ICD10: C61) was based on records from national cancer registries and self-reports. Lethal PCa is identified based on records of the primary cause of death from national death registries. PCa as the contributory (secondary) cause of death was therefore considered as an accompanied condition. Self‐reported PCa family history (FH) in father or brothers, ethnic background, age at recruitment, age at death, BPH diagnosis (ICD10: N40) were also obtained. Charlson Comorbidity Index score was calculated based on the presence of 17 conditions, each of which is assigned a weighted score of 1, 2, 3, or 6[18]. All participants provided written informed consent to take part in the study. The study was approved by the institutional review board at Ruijin Hospital, Shanghai, China.

Genotyping and Whole Exome Sequencing

Approximately 97% and 40% of UKB subjects had genotyping data and whole exome sequence (WES) data respectively, which were provided by the UKB (October 20, 2020). The UKB samples were genotyped using UK Biobank Axiom array[19]. Genotyping data was provided with performed imputation and quality control [19]. SNPs were excluded if they had: (1) genotype call rate < 95%; (2) minor allele frequency < 0.01. PLINK 2.0 software was used for calculating 10 principal components (PCs) after linkage disequilibrium pruning[20]. Whole exome sequencing data in VCF form was also obtained. Pathogenic/likely pathogenic (P/LP) mutations in the 10 hereditary cancer susceptibility genes (ATM, BRCA1, BRCA2, CHEK2, HOXB13, PALB2, MLH1, MSH2, MSH6 and PMS2) were annotated following the general sequence variant interpretation guidelines of the ACMG[21].

PRS calculation

Independent genome-wide significant risk variants were selected (n = 27 for BPH [22, 23], Table S2) from the evidence-based published GWAS in the Icelandic BPH/LUTs dataset[22] (n = 113,443) and the Electronic Medical Records and Genomics network (eMERGE)[23] (n = 6,672).We used PRSice2 software to calculate PRS with GWAS clumped results (range 250kb, r² = 0.1) and summary association statistics[24]. Empirical p values were determined by permutation analyses in 10000 rounds. A PRS for each patient was calculated by summing the number of risk alleles at each of the 27 SNPs multiplied by the logarithm of the SNP’s Odds ratio. The previously reported lethal PCa risk related PRACTICAL PRS (No. SNPs = 147)[11, 25] and Polygenic hazard score (PHS, No. SNPs = 46)[26] were also calculated in this way. Details of the evaluation was described in Fig. 1.

Statistical Analyses

Baseline characteristics were illustrated by descriptive statistics. Chi-squared test was used to compare the difference of categorical variables. Student’s t test was applied to evaluate the normally distributed continuous variables while Mann-Whitney U test was used to evaluate non-normally distributed continuous variables and the trend for the ranks of across ordered groups. Crude PCa-specific mortality was calculated from the date of the first diagnosis of PCa, to the date of death or the last date of follow-up updated in the Jan 1st, 2022. Univariable or multivariable analyses were performed to test for the independent effect of factors associated with PCa death using logistic and Cox proportional hazards regression models. Survival time after PCa diagnosis and the differences across subgroups were estimated with restricted mean survival time (RMST)[27]. Likelihood ratio tests were implemented for the interaction between BPH PRS and hereditary P/LP gene mutations, comparing models with or without interaction terms.

All Statistical analyses were performed using R version 4.1.2[28] and a 2-tailed p < 0.05 was considered statistically significant.

The demographic characteristics and the baseline information of the prospective cohort were shown in Table 1, while the prevalence data were in Table S1. During a mean follow-up time of 8.0 yr, a total of 14,564 males were diagnosed with PCa, of which 959 (8.2%) died of PCa, 1,173 (8.1%) died due to other reasons and 12,193 (83.7%) survived till the latest follow-up in Jan 1^st, 2022. Men who died of PCa were less likely to have a BPH diagnosis (15.7% vs. 19.6%, P=0.004) compared to men with the non-lethal disease. PCa patients without a BPH history would have a higher PCa-specific mortality rate (crude PCa-specific mortality rate in males without BPH was 8.49 per 1,000 men vs. 6.91 per 1,000 men in males with BPH, rate ratio, RR=1.28 ,95% confidence interval, 95%CI: 1.08-1.56, P=0.004). There was no difference in family history of PCa (P>0.9) between the two groups (Table 1, Table S2). Besides the increasing age, a higher score on the Charlson Comorbidity Index (indicating a greater burden of illness) was also associated with an increased lethal PCa risk in three different defined subgroups (P<0.001, Table S2).

A total of 21 SNPs were used to calculate BPH/PV PRS finally (Table S3). And the Odds ratios (ORs) per 1-SD increment in BPH-related PRS represented 1.75 (OR, 95%CI: 1.68–1.83) increased risk to have BPH (significance test P<0.001, empirical P<0.001). Due to mismatched allele codes, we only matched 130 out of 147 SNPs and 24 out of 47 SNPs reported in the PRACTICAL PRS[11, 25] and PHS [26](established lethal PCa-related PRS), respectively.

BPH/PV PRS was not only significantly associated with PCa-specific mortality (Hazard ratio, HR=0.92, 95%CI: 0.86-0.99, P=0.02) in PCa patients, but also associated with incident PCa death (HR=0.92, 95%CI: 0.85-0.99, P=0.03 or P=0.008) at a population level. The PRACTICAL PRS and the PHS performed similarly well in predicting lethal PCa to those previously reported, especially in the prevalence and incidence cohort (both P<0.05). However, they showed no significant relation with lethal PCa outcomes in PCa patients (P>0.05). The survival analysis in the entire PCa cohort revealed that men with lower BPH/PV PRS might have a significantly shorter survival time (Log-rank test P=0.004). Compared with those with top BPH/PV PRS (75-100^th percentile, the highest risk of BPH), PCa patients with lowest BPH/PV PRS (0-25^th percentile, the lowest risk of BPH) would have a 1.40-fold increased risk of PCa death during the follow-up (HR, 95%CI: 1.16-1.69, P=0.001, Table 3), as well as a relatively lower long-term survival probability at 83.7% (20-year survival probability, 95%CI: 80.4-87.1%, P for trend<0.001) and shorter survival time of 0.37 yr (RMST, 95%CI: 0.14-0.61, P=0.002).

The carrier rates of P/LP mutation in the UKB unselected PCa patients (data available in 5,899 PCa patients) for each of the ten genes were shown in Table 4, ranging from 0.02% for MLH1 to 1.64% for CHEK2 (Mutations listed in Table S4). BRCA2 and PALB2 were significantly associated with PCa death, with HR estimated at 3.91 (95%CI: 2.34-6.51, P<0.001) and 4.24 (95%CI: 1.34-13.34, P=0.01). Mutation status was not significantly associated with shorter survival time after PCa diagnosis except for mutations in BRCA2. BRCA2 carriers were 3.23 yr (RMST, 95%CI: 1.55-4.90, P<0.001) earlier to die from PCa. When investigating the interaction between mutation status and BPH/PV PRS, additional analyses were performed in subgroups of patients based on the carrier status (Figure 2). In non-carriers, a decreased BPH/PV PRS was significantly associated with an increased risk of PCa death (Log rank P=0.03, Figure 2.A). However, the trend was not observed among DNA damage repair (DDR) gene carriers (Log rank P=0.28, Figure 2.B). And no interaction effects between BPH/PV PRS and hereditary gene mutations were detected (P_interaction for BRCA2/PALB2: 0.83; P_interaction for DDR: 0.99).

In the present study, we derived a PRS based on the 21 reported independent risk variants for BPH in European ancestry, which showed an additional and considerable prognostic value for PCa. We also evaluated 10 guideline‐recommended familial cancer risk genes for their association with PCa fatality in prospective PCa cohort. We found that: (1) BPH/PV PRS was significantly and independently associated with PCa death; (2) known PCa-related PRS (PRACTICAL PRS and PHS) could distinguish individuals’ PCa risk but failed to predict PCa patients’ natural outcome; (3) Only the P/LP mutations in BRCA2 or PALB2 were found to be associated with PCa death in this cohort and did not interact with BPH/PV PRS.

As a well-established risk factor, the association between PV or BPH and PCa has been comprehensively investigated in many studies. Orsted et al.’s nationwide cohort study of 3,009,258 men showed males with BPH were more likely to develop PCa and die of it[29]. However, increased examination (PSA testing, DRE, and biopsies) in patients with BPH could lead directly to PCa detection as well as the consequence of PCa-related death. Other studies conducted that a relatively small PV would be associated with poorer PCa prognosis from prostatectomy or biopsy cohort. For instance, a series of studies on radical prostatectomy specimens showed that tumors arising in larger prostates tend to have favorable pathological features and biochemical progression[13-15]. Moreover, Lorenzo et al. suggested that the mechanical stress induced by BPH might impede prostatic tumor growth[30]. Thus, a measure of PV or BPH status could be a potential index for PCa aggressiveness risk. However, the change of PV or BPH happens most likely in men’s 70s[15]. To evaluate the possibility of a changed PV or BPH diagnosis at a younger age, it would be useful to develop an instrument to provide an indirect measurement in advance. We, hence, focused on SNPs-based PRS which could provide consistent lifetime estimates for different health statuses, including PV and BPH.

Many studies reported PRS using PCa risk-associated SNPs to evaluate PCa predisposition. Those PRSs were mainly designed for identifying individuals who may be predisposed to PCa rather than the likelihood of more aggressive or lethal PCa. PRS for PCa risk was not associated with pathologic Gleason Score, pathological stage, and biochemical recurrence[31]. Thus, it showed the same detection bias as family history or BPH in epidemiology research, ignoring the fact that increased discoveries of PCa could bring elevated number of clinically significant PCa (Table 2 and Table S2). On the other hand, limited aggressive PCa-associated SNPs made it hard to establish a direct SNPs-based PRS for PCa prognosis. Therefore, with a logically sound hypothesis, sufficient clinical evidence, and underline biological mechanisms, we proved that the BPH/PV PRS could be a tool for predicting aggressive PCa beyond other reported PRSs. However, cumulative effect of BPH/PV PRS could be too mild to emerge under the relative few DNA damage repair gene carriers.

Important strengths of this study include the prospective design and the relatively large sample size of the cases from a general population which increases the generalizability of our findings. Meanwhile, several limitations should also be noted. First, the relatively few P/LP mutation carriers affected the subgroup analysis though a similar trend still existed. Further investigation would be worthy in a carrier-only cohort. Second, the statistical power was reduced due to the limited number of available SNPs. If possible, we would repeat and enrich it in another validation cohort in the future. Third, the identification of aggressive PCa and indolent PCa was restricted by the nature of the UK Biobank database. We were unable to identify patients with aggressive phenotypes other than death, such as high Gleason Score, metastatic disease, etc. These phenotypes were equally important in terms of disease management and clinical intervention. However, the current study focusing on PCa death could also provide reliable evidence which might raise future interest in the evaluation for different disease outcomes.

In conclusion, BPH/PV PRS was significantly associated with PCa death and functioned as a potential prognostic assessment tool for PCa. Moreover, a combination of BPH/PV PRS and BRCA2/PALB2 mutation status may help to identify high lethal PCa genetic risk and shorter survival time.

PCa

Prostate cancer

PRS

Polygenic risk score

BPH

Benign prostatic hyperplasia

Prostate volume

PSA

Prostate-specific antigen

Family history

SNP

Single nucleotide polymorphism

RMST

restricted mean survival time

UKB

UK Biobank

Availability of data and material

Data used in this research are publicly available to qualified researchers on application to the UK Biobank (www.ukbiobank.ac.uk).

Acknowledgement

We thank the UK Biobank for access to the data (project number: 66813). This work was in supported by grants from National Natural Science Foundation of China (Grant No. 81972645), Innovative research team of high-level local universities in Shanghai, Shanghai Youth Talent Support Program, intramural grant of The University of Hong Kong to Dr. Rong Na, and Shanghai Sailing Program (22YF1440500) to Dr. Da Huang. All the funders had no role in study design, data collection, data analysis, interpretation, and writing of the report

Author Contribution Statement

Conceptualization, Rong Na; Data curation, Da Huang; Formal analysis, Xiaohao Ruan; Investigation, Jingyi Huang; Project administration, Da Huang; Supervision, James Hok-Leung Tsu and Rong Na; Writing – original draft, Xiaohao Ruan; Writing – review & editing, Rong Na.

Ethics approval and consent to participate

All participants provided written informed consent to take part in the study. The study

was approved by the Northwest Multi-centre Research Ethics in Manchester, UK (IRAS project ID: 299116; Application No. 66813)

Competing interests

The authors declare that there are no competing interests.

Consent for publication

All the patients were asked for their consent for publication.

Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, Bray F: Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021, 71:209-249.
Pezaro C, Woo HH, Davis ID: Prostate cancer: measuring PSA. Internal Medicine Journal 2014, 44:433-440.
Ahn J, Moslehi R, Weinstein SJ, Snyder K, Virtamo J, Albanes D: Family history of prostate cancer and prostate cancer risk in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) Study. International Journal of Cancer 2008, 123:1154-1159.
Xu J, Isaacs WB: Incorporation of Polygenic Risk Score into Guidelines for Inherited Risk Assessment for Prostate Cancer. European Urology 2021, 80:139-141.
Na R, Zheng SL, Han M, Yu H, Jiang D, Shah S, Ewing CM, Zhang L, Novakovic K, Petkewicz J, et al: Germline Mutations in ATM and BRCA1/2 Distinguish Risk for Lethal and Indolent Prostate Cancer and are Associated with Early Age at Death. European Urology 2017, 71:740-747.
Shi Z, Platz EA, Wei J, Na R, Fantus RJ, Wang C-H, Eggener SE, Hulick PJ, Duggan D, Zheng SL, et al: Performance of Three Inherited Risk Measures for Predicting Prostate Cancer Incidence and Mortality: A Population-based Prospective Analysis. European Urology 2021, 79:419-426.
Darst BF, Sheng X, Eeles RA, Kote-Jarai Z, Conti DV, Haiman CA: Combined Effect of a Polygenic Risk Score and Rare Genetic Variants on Prostate Cancer Risk. European Urology 2021, 80:134-138.
Wu Y, Yu H, Zheng SL, Na R, Mamawala M, Landis T, Wiley K, Petkewicz J, Shah S, Shi Z, et al: A comprehensive evaluation of CHEK2 germline mutations in men with prostate cancer. The Prostate 2018, 78:607-615.
Wu Y, Yu H, Zheng SL, Feng B, Kapron AL, Na R, Boyle JL, Shah S, Shi Z, Ewing CM, et al: Germline mutations in PPFIBP2 are associated with lethal prostate cancer. The Prostate 2018, 78:1222-1228.
NCCN: NCCN Guidelines Prostate Cancer, NCCN clinical practice guidelines in oncology – Prostate Cancer. Version 4.2022.
Klein RJ, Vertosick E, Sjoberg D, Ulmert D, Rönn A-C, Häggström C, Thysell E, Hallmans G, Dahlin A, Stattin P, et al: Prostate cancer polygenic risk score and prediction of lethal prostate cancer. NPJ Precision Oncology 2022, 6:25.
Pinto AR, Silva J, Pinto R, Medeiros R: Aggressive prostate cancer phenotype and genome-wide association studies: where are we now? Pharmacogenomics 2020, 21:487-503.
Freedland SJ, Isaacs WB, Platz EA, Terris MK, Aronson WJ, Amling CL, Presti JC, Kane CJ: Prostate size and risk of high-grade, advanced prostate cancer and biochemical progression after radical prostatectomy: a search database study. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 2005, 23:7546-7554.
Briganti A, Chun FKH, Suardi N, Gallina A, Walz J, Graefen M, Shariat S, Ebersdobler A, Rigatti P, Perrotte P, et al: Prostate volume and adverse prostate cancer features: fact not artifact. European journal of cancer (Oxford, England : 1990) 2007, 43:2669-2677.
Newton MR, Phillips S, Chang SS, Clark PE, Cookson MS, Davis R, Fowke JH, Herrell SD, Baumgartner R, Chan R, et al: Smaller prostate size predicts high grade prostate cancer at final pathology. The Journal of Urology 2010, 184:930-937.
Ørsted DD, Bojesen SE: The link between benign prostatic hyperplasia and prostate cancer. Nature Reviews Urology 2013, 10:49-54.
Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, Downey P, Elliott P, Green J, Landray M, et al: UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med 2015, 12:e1001779.
de Groot V, Beckerman H, Lankhorst GJ, Bouter LM: How to measure comorbidity. a critical review of available methods. Journal of Clinical Epidemiology 2003, 56:221-229.
Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, Motyer A, Vukcevic D, Delaneau O, O'Connell J, et al: The UK Biobank resource with deep phenotyping and genomic data. Nature 2018, 562:203-209.
Galinsky KJ, Bhatia G, Loh P-R, Georgiev S, Mukherjee S, Patterson NJ, Price AL: Fast Principal-Component Analysis Reveals Convergent Evolution of ADH1B in Europe and East Asia. American Journal of Human Genetics 2016, 98:456-472.
Wei J, Yang W, Shi Z, Lu L, Wang Q, Resurreccion WK, Engelmann V, Zheng SL, Hulick PJ, Cooney KA, et al: Observed evidence for guideline-recommended genes in predicting prostate cancer risk from a large population-based cohort. The Prostate 2021, 81:1002-1008.
Gudmundsson J, Sigurdsson JK, Stefansdottir L, Agnarsson BA, Isaksson HJ, Stefansson OA, Gudjonsson SA, Gudbjartsson DF, Masson G, Frigge ML, et al: Genome-wide associations for benign prostatic hyperplasia reveal a genetic correlation with serum levels of PSA. Nature Communications 2018, 9:4568.
Li W, Klein RJ: Genome-wide association study identifies a role for the progesterone receptor in benign prostatic hyperplasia risk. Prostate Cancer and Prostatic Diseases 2021, 24:492-498.
Choi SW, Mak TS-H, O'Reilly PF: Tutorial: a guide to performing polygenic risk score analyses. Nature protocols 2020, 15:2759-2772.
Schumacher FR, Al Olama AA, Berndt SI, Benlloch S, Ahmed M, Saunders EJ, Dadaev T, Leongamornlert D, Anokian E, Cieza-Borrella C, et al: Association analyses of more than 140,000 men identify 63 new prostate cancer susceptibility loci. Nature Genetics 2018, 50:928-936.
Huynh-Le M-P, Fan CC, Karunamuni R, Thompson WK, Martinez ME, Eeles RA, Kote-Jarai Z, Muir K, Schleutker J, Pashayan N, et al: Polygenic hazard score is associated with prostate cancer in multi-ethnic populations. Nature Communications 2021, 12:1236.
Royston P, Parmar MKB: Restricted mean survival time: an alternative to the hazard ratio for the design and analysis of randomized trials with a time-to-event outcome. BMC medical research methodology 2013, 13:152.
Team RC: R: A language and environment for statistical computing. 2021.
Ørsted DD, Bojesen SE, Nielsen SF, Nordestgaard BG: Association of clinical benign prostate hyperplasia with prostate cancer incidence and mortality revisited: a nationwide cohort study of 3,009,258 men. European Urology 2011, 60:691-698.
Lorenzo G, Hughes TJR, Dominguez-Frojan P, Reali A, Gomez H: Computer simulations suggest that prostate enlargement due to benign prostatic hyperplasia mechanically impedes prostate cancer growth. Proceedings of the National Academy of Sciences of the United States of America 2019, 116:1152-1161.
Wang S-H, Huang S-P, Pan Y-J, Hsiao P-C, Li C-Y, Chen L-C, Yu C-C, Huang C-Y, Lin VC, Lu T-L, Bao B-Y: Association between the polygenic liabilities for prostate cancer and breast cancer with biochemical recurrence after radical prostatectomy for localized prostate cancer. American Journal of Cancer Research 2021, 11:2331-2342.

Table 1. Demographic and clinical characteristics of prostate cancer patients in UK Biobank.

Characteristics	Lethal PCa	Non-lethal PCa	All	P value²
Male Participants, No. (%)	959 (6.6)	13,605	14,564
Age at diagnostic, mean (SD), yr	65.8 (6.2)	66.4 (6.5)	66.4 (6.5)	0.013
Age at death, mean (SD), yr	71.6 (5.3)	73.5 (5.1)¹	72.7 (5.3)	<0.001
BPH diagnosis, No. (%)				0.004
Yes	151 (15.7)	2,662 (19.6)	2,813 (19.3)
No	808 (84.3)	10,943 (80.4)	11,751 (80.7)
Family history of PCa, No. (%)				0.949
Yes	114 (11.9)	1,618 (11.9)	1,732 (11.9)
No	799 (83.3)	11,416 (83.9)	12,215 (83.9)
Missing	46 (4.8)	571 (4.2)	617 (4.2)
Charlson comorbidity index scores				<0.001
0	108 (11.3)	4444 (32.7)	4552 (31.3)
1-2	112 (11.7)	5911 (43.4)	6023 (41.4)
>2	739 (77.0)	3250 (23.9)	3989 (27.3)
WES data, No. (%)	392 (40.9)	5,507 (40.5)	5,899 (40.5)	0.808

Abbreviation: PCa, Prostate cancer; BPH, Benign prostate hyperplasia; N/A, Not applicable; WES, Whole-exome sequencing.

1. 1173 deaths due to other reasons, not for PCa as primary death cause.

2. Chi-squared test or t test was used to compare lethal PCa group and Non-lethal PCa group.

Table 2. Comparison of multivariate analysis of associations between BPH/PV risk score and established prostate cancer risk score.

Cohort	Risk Score⁴	Event	OR (95%CI)	P value	HR (95%CI)	P value
PCa prevalence¹ (n=209,910)	BPH/PV PRS (21 SNPs)	959	0.93 (0.87 - 1.00)	0.04	0.92 (0.86 - 0.98)	8.29×10^-3
	PRACTICAL PRS (130/147 SNPs)	959	1.69 (1.58 - 1.80)	<0.001	1.56 (1.46 - 1.66)	<0.001
	PHS (24/46 SNPs)	959	1.18 (1.15 - 1.20)	<0.001	1.15 (1.12 - 1.17)	<0.001
PCa incidence² (n=206,194)	BPH/PV PRS	662	0.85 (0.66 - 1.09)	0.19	0.92 (0.85 - 0.99)	0.03
	PRACTICAL PRS	662	1.64 (1.51 - 1.78)	<0.001	1.59 (1.47 - 1.72)	<0.001
	PHS	662	1.17 (1.14 - 1.20)	<0.001	1.15 (1.13 - 1.18)	<0.001
PCa prognosis³ (n=14,564)	BPH/PV PRS	959	0.92 (0.86 - 1.00)	0.04	0.92 (0.86 - 0.99)	0.02
	PRACTICAL PRS	959	1.00 (0.92 - 1.08)	0.94	0.97 (0.91 - 1.03)	0.34
	PHS	959	1.00 (0.99 - 1.02)	0.61	0.99 (0.97 - 1.01)	0.34

Abbreviation: PCa, Prostate cancer; OR, Odds ratio; HR, Hazard ratio; CI, Confidence interval; PHS, Polygenic hazard score; BPH, Benign prostate hyperplasia; PV, prostate volume; SNP, Single nucleotide polymorphism; PRS, Polygenic risk score.

PCa prevalence cohort: the cohort included all patients and used patient’s age at death or at last follow-up (2022 Jan 1^st or lost) as time variable for Cox regression. Multivariate ORs were adjusted for age at last follow-up, family history, Charlson Comorbidity Index score and 10 principal components. Multivariate HRs were adjusted for the same factors except for age.
PCa incidence cohort: the cohort excluded PCa patients before recruitment and used follow-up time as time variable for Cox regression. Multivariate ORs or HRs were adjusted for age at recruitment, family history, Charlson Comorbidity Index score and 10 principal components.
PCa prognosis cohort: the cohort excluded patients without PCa at last follow-up and used survival time as time variable for Cox regression. Multivariate ORs or HRs were adjusted for age at onset, family history, Charlson Comorbidity Index score and 10 principal components.
The risk scores were standardized within each cohort by dividing by standard deviation.

Table 3. Associations between categorized BPH/PV-related PRS and risk of prostate cancer-specific mortality in prostate cancer prognosis cohort.

	75-100^th PRS	50-75^th PRS	25-50^th PRS	0-25^th PRS	*P-trend*
Mean age at PCa death (SD),yr	73.0 (5.1)	71.9 (5.1)	71.6 (5.7)	71.1 (5.5)	0.17
Proportion of deceased	5.5% (199/3,641)	6.6% (241/3641)	7.1% (260/3,641)	7.1% (259/3,641)	0.01
*OR (95%CI)**	1 (Ref.)	1.33 (1.07-1.65)	1.42 (1.15-1.77)	1.38 (1.11-1.71)	3.73×10^-3
*HR (95%CI)**	1 (Ref.)	1.29 (1.06-1.56)	1.33 (1.10-1.61)	1.40 (1.16-1.69)	7.52×10^-4
5-year survival probability (95%CI, %)	96.3 (95.7 - 97.0)	95.9 (95.2 - 96.6)	94.9 (94.1 - 95.6)	94.8 (94.0 - 95.6)	3.10×10^-3
10-year survival probability (95%CI, %)	93.5 (92.6 - 94.5)	92.2 (91.2 - 93.3)	91.7 (90.6 - 92.8)	91.64 (90.3 - 92.6)	8.44×10^-⁵
20-year survival probability (95%CI, %)	87.4 (84.0 - 90.8)	84.5 (81.4 - 87.8)	84.2 (81.0 - 87.5)	83.7 (80.4 - 87.1)	<2.20×10^-¹⁶
Restricted mean survival time (95%CI, %)	0 (Ref.)	-0.01 (-0.26 ~ 0.24)	-0.09 (-0.34 ~ 0.15)	-0.37 (-0.61 ~ -0.14)	1.88×10^-3

Abbreviation: PCa, Prostate cancer; vs., versus; OR, odds ratio; CI, confidence interval; Ref, reference.

* OR and HR is calculated by logistic and cox model adjusted for age at onset, charlson comorbidity index score and 10 principal components.

Table 4. Association of germline pathogenic mutations in guideline-recommended genes with lethal prostate cancer risk in the UK Biobank PCa patients with WES results.

Genes	No. (%) of Carrier			Cancer-specific mortality
Genes	Lethal PCa (n=392)	Non-lethal PCa (n=5,507)	PCa patients (N=5,899)	HR* (95% CI)	P value
*BRCA2*	16 (4.1)	31 (0.56)	47 (0.80)	3.91 (2.34 - 6.51)	1.79×10^-7
*PALB2*	3 (0.77)	3 (0.05)	6 (0.15)	4.24 (1.34 - 13.34)	0.01
*BRCA1*	1 (0.26)	9 (0.18)	10 (0.17)	1.95 (0.27 - 14.05)	0.51
*ATM*	3 (0.77)	43 (0.78)	46 (0.78)	1.12 (0.36 - 3.51)	0.84
*CHEK2*	8 (2.0)	89 (1.62)	97 (1.64)	1.36 (0.67 - 2.75)	0.39
*HOXB13*	5 (1.3)	63 (1.14)	68 (1.15)	1.17 (0.48 - 2.83)	0.73
*MSH2*	0	0	0	N/A	N/A
*MSH6*	2 (0.51)	22 (0.40)	24 (0.41)	1.47 (0.37 - 5.94)	0.59
*MLH1*	1 (0.26)	0	1 (0.02)	N/A	N/A
*PMS2*	1 (0.26)	11 (0.20)	12 (0.20)	1.11 (0.16 - 7.94)	0.92
*MMR (4 genes)*	4 (1.0)	33 (0.60)	37 (0.63)	1.64 (0.61 - 4.39)	0.33
*DDR (9 genes)*	35 (8.93)	208 (3.78)	243 (4.12)	2.24 (1.58 - 3.18)	6.18×10^-7
*BRCA2/PALB2*	19 (4.8)	37 (0.67)	56 (0.95)	4.00 (2.50 - 6.40)	7.93×10^-9

Abbreviation: WES, Whole-exome sequencing; OR, Odds ratio; CI, Confidence interval; MMR, Mismatch repair genes (include MSH2, MSH6, MLH1 and PMS2); DDR, DNA damage repair genes (include BRCA1/2, ATM, CHEK2, MSH2, MSH6, MLH1, PMS2, PALB2); N/A, not applicable. * Association results were adjusted for age at diagnosis, Charlson Comorbidity Index score and 10 principal components.

SupplementaryMaterials.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Genetic Risk Assessment of Lethal Prostate Cancer Using Polygenic Risk Score and Hereditary Cancer Susceptibility genes

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Introduction

Materials And Methods

Study Population

Genotyping and Whole Exome Sequencing

PRS calculation

Statistical Analyses

Results

Discussion

Abbreviations

Declarations

References

Tables

Supplementary Files

Status:

Version 1