Assessment of a Polygenic Risk Score in Screening for Prostate Cancer

Abstract

Background: The incidence of prostate cancer is increasing. Screening with an assay of prostate-specific antigen (PSA) has a high rate for false positive results. Genomewide association studies have identified common germline variants in persons with prostate cancer, which can be used to calculate a polygenic risk score associated with risk of prostate cancer.

Methods: We recruited persons 55 to 69 years of age from primary care centers in the United Kingdom. Using germline DNA extracted from saliva, we derived polygenic risk scores from 130 variants known to be associated with an increased risk of prostate cancer. Participants with a polygenic risk score in the 90th percentile or higher were invited to undergo prostate cancer screening with multiparametric magnetic resonance imaging (MRI) and transperineal biopsy, irrespective of PSA level.

Results: Among 40,292 persons invited to participate, 8953 (22.2%) expressed interest in participating and 6393 had their polygenic risk score calculated; 745 (11.7%) had a polygenic risk score in the 90th percentile or higher and were invited to undergo screening. Of these 745 participants, 468 (62.8%) underwent MRI and prostate biopsy; prostate cancer was detected in 187 participants (40.0%). The median age at diagnosis was 64 years (range, 57 to 73). Of the 187 participants with cancer, 103 (55.1%) had prostate cancer classified as intermediate or higher risk according to the 2024 National Comprehensive Cancer Network (NCCN) criteria, so treatment was indicated; cancer would not have been detected in 74 (71.8%) of these participants according to the prostate cancer diagnostic pathway currently used in the United Kingdom (high PSA level and positive MRI results). In addition, 40 of the participants with cancer (21.4%) had disease classified as unfavorable intermediate risk or as high or very high risk according to NCCN criteria.

Conclusions: In a prostate cancer screening program involving participants in the top decile of risk as determined by a polygenic risk score, the percentage found to have clinically significant disease was higher than the percentage that would have been identified with the use of PSA or MRI. (Funded by the European Research Council Seventh Framework Program and others; BARCODE1 ClinicalTrials.gov number, NCT03857477.).

Figure 1

is a consort diagram that summarises the participant pathway through the BARCODE1 study from expression of interest through to biopsy outcome. PRS = Polygenic Risk Score, QC = Quality Control; PSA = prostate-specific antigen, MRI = Magnetic Resonance Imaging, PI-RADS = Prostate Imaging Reporting and Data System. Cancers are classified by the NCCN (2023) Risk Groups (very low, low, intermediate favourable, intermediate unfavourable, high or very high risk).

Figure 2

The top bar shows all PrCa detected using PRS alone. The cancers are divided into 3 groups: Gleason Score ≤6 shown in green, Gleason Score 3+4 with ≤10% Gleason pattern 4 present shown in orange and Gleason Score 3+4 with >10% Gleason pattern 4 shown in red. The second bar shows the cancers detected if a PSA threshold of >3.0ug/l were used to stratify. The third bar shows the cancers detected if an MRI PI-RADS threshold of ≥3 were used to stratify. The fourth bar shows the cancers detected if a PSA threshold of >3.0ug/L and MRI PI-RADS score of ≥3 were used to stratify. The final bar shows the cancers detected if a PSA threshold of >3.0ug/L, PSA density of >0.12ng/ml/cc and MRI PI-RADS score of ≥3 were used to stratify. For the groups that include MRI, the total number of cancers evaluated were 186 of the overall 187, as one patient was unable to undergo MRI due to claustrophobia.

Figure 3

A threshold of 3.5%-4% ten-year absolute risk has been suggested as generating the greatest number of quality-adjusted life-years from risk-based screening