This study investigates the effectiveness of a non-invasive polygenic risk score (PRS) for assessing prostate cancer (PCa) risk compared to benign outcomes. The research included 390 men who were prospectively enrolled to undergo a standard 3T multiparametric MRI followed by MRI/transrectal ultrasound (TRUS) fusion prostate biopsy (PBx). Saliva samples were collected from each participant prior to biopsy for genotyping, allowing for the calculation of a polygenic risk score based on 451 genetic variants linked to prostate cancer. This PRS was derived from a large multi-ancestry genome-wide association study (GWAS) involving over 150,000 PCa cases and nearly 800,000 controls. Each participant also underwent 12-14 core systematic biopsies and targeted biopsies for any PIRADSā„3 lesion observed in their MRI. Using multivariable logistic regression, the study analyzed the ability of PRS and other clinical indicators to predict prostate cancer presence.
The results show that PRS, along with factors like age, PSA density, and PIRADS score, were independent predictors of PCa, enhancing diagnostic accuracy. Specifically, adding PRS to pre-biopsy clinical predictors improved the area under the ROC curve from 0.74 to 0.78, with an increase in sensitivity from 69% to 74% and specificity from 67% to 69%. The cohort, with a median age of 66, varied across racial and ethnic backgrounds, and PIRADS scores ranged from benign findings to PIRADS 5. Among biopsy results, 36% were benign, while 19% had Gleason Grade (GG) 1, 20% had GG2, and smaller percentages were found with GG3 to GG5.
The findings highlight the utility of PRS as a non-invasive predictive tool, which, when combined with standard clinical data, can significantly improve PCa detection from saliva samples. This approach offers a promising method to stratify PCa risk pre-biopsy, potentially improving early diagnosis and patient outcomes.
