Prostate Cancer Testing Via Polygenic Risk Score Outperforms Standard Approaches

A polygenic risk score (PRS) using germline DNA extracted from saliva identified a higher proportion of individuals with clinically significant prostate cancer warranting treatment compared with prostate-specific antigen (PSA) testing or MRI, according to data from the observational BARCODE1 screening study (NCT03857477) published in the New England Journal of Medicine.¹

BARCODE1 findings revealed that among 745 participants had a PRS in the 90th percentile or higher, and 468 (62.8%) underwent MRI and prostate biopsy. Of that subset, 187 (40.0%) were subsequently diagnosed with prostate cancer. Most of these patients (55.1%) had prostate cancer classified as intermediate or higher risk according to 2024 NCCN criteria. The study authors noted that cancer would not have been detected in 71.8% of these participants according to the prostate cancer diagnostic pathway currently used in the United Kingdom (UK), which is composed of a high PSA level and positive MRI results.

“The current tool to detect prostate cancer risk is a blood test [that] measures levels of PSA,” Rosalind Eeles, FMedSci, PhD, FRCP, FRCR, a professor and leader of the Oncogenetics Group at The Institute of Cancer Research in London, England, an honorary consultant in clinical oncology and oncogenetics at the Royal Marsden NHS Foundation Trust, and the lead study author of BARCODE1, said in a statement to OncLive. “The PSA test [can] falsely indicate prostate [levels at a high rate] and often detects cancers [that] grow so slowly they are unlikely to ever be life-threatening. The UK’s National Screening Committee recently reviewed the evidence and found that PSA testing in the general population would do more harm than good. Researchers are therefore searching for new ways to identify prostate cancer risk, including through genetics.”

What Was the Design of BARCODE1?

BARCODE1 was a prospectively designed, single-arm study that enrolled men of European ancestry aged 55 to 69 years across 69 centers in the UK between March and July 2019. Key eligibility criteria included having no personal history of prostate cancer or being currently under investigation for suspicion of the disease, no prostate biopsy within 12 months, and no contraindications to MRI or biopsy. Individuals who were interested in BARCODE1 completed a health screening questionnaire and provided a postal saliva sample for genetic analysis.

“My team calculated the risk of prostate cancer from DNA variants extracted from saliva, called a PRS,” Eeles explained. “The score is based on 130 genetic variations in the DNA code that are linked to prostate cancer, and it was developed by studying the DNA of hundreds of thousands of men.”

The DNA panel used prostate cancer–risk single nucleotide polymorphisms that were validated for use only in individuals of European ancestry. Study participants with a PRS in the 90th percentile or above were referred to a center for genetic-risk counselling and were then offered PSA testing, multiparametric MRI, and transperineal biopsy. Cancers were deemed to be clinically significant if patients had a Gleason score of at least 7.

In total, 6393 participants were genotyped; among the 745 individuals who had a PRS in the 90th percentile or above, 468 (62.8%) accepted MRI and biopsy, 177 withdrew due to personal choice, and 95 were withdrawn by the study team. The mean age at enrollment was 61.2 years, and 20.9% reported a family history of prostate cancer. The median age at diagnosis was 64 years (range, 57-73).

Other Key Findings From BARCODE1

Additional findings from BARCODE1 showed that the 187 patients identified by PRS had low/very low– (44.9%), intermediate favorable– (33.7%), intermediate unfavorable– (15.0%), or high/very high– (6.4%) risk cancers. Comparatively, patients with cancers identified by a PSA threshold of over 3.0 µg/L (n = 69) had cancers in these risk categories at respective rates of 24.6%, 34.8%, 27.5%, and 13.0%. Patients who had prostate cancer diagnosed by MRI alone (n = 61) had cancers in these risk categories at respective rates of 26.2%, 27.9%, 29.5%, and 16.4%. Patients whose disease was diagnosed by PSA and MRI (n = 30) had cancers in these risk categories at respective rates of 6.7%, 20.0%, 46.7%, and 26.7%.

Individuals who had low/very low–, intermediate favorable–, intermediate unfavorable–, or high/very high–risk cancers per PSA testing only had cancers missed at respective rates of 56.8%, 33.1%, 7.6%, and 2.5%. Those who were diagnosed by MRI only in these risk categories had missed cancers at rates of 54.4%, 36.8%, 8.0%, and 0.8%, respectively. Individuals who were diagnosed by PSA plus MRI in these risk categories had cancers missed at corresponding rates of 52.6%, 36.5%, 9.0%, and 1.9%.

“In our study, for the men with the highest genetic risk, the test falsely identified fewer people with prostate cancer [compared with] the PSA test, picked up [individuals] with cancer who would have been missed by the PSA test alone, and picked up a higher proportion of the aggressive cancers than the PSA test,” Eeles explained. “The test also accurately identified men with prostate cancer [whose disease was] missed by an MRI scan.”

Eeles and her coauthors estimated that 20.8% (range, 9.7%-33.9% for ages 55-74) of patients in the 90th percentile by PRS and with screen-detected cancer would be overdiagnosed, meaning that their screen-detected prostate cancer would take longer than their remaining lifetime to progress to clinical cancer. Using only a PSA threshold of over 3.0 µg/L, 17.2% (range, 7.0%-21.0%) of patients would have been overdiagnosed and 15.6% (range, 4.0%-25.0%) would have been overdiagnosed using PI-RADS 2.1 criteria.

“Since the [BARCODE1] study began, more risk variants for men of Asian and African ancestry have been identified, and we are trialing a saliva test for this population, through the national TRANSFORM trial, to assess the most cost-effective and accurate way to screen men for prostate cancer,” Eeles said.

TRANSFORM is recruiting men aged 50 to 74 years, or aged 45 to 74 years for groups known to have poorer outcomes, such as Black men.² The study's goal is to identify the safest, most accurate, and most cost-effective method to screen men for prostate cancer. Initially 16,000 men will be included in stage 1 of the study, which will compare new testing techniques with the current National Health System diagnostic pathway. The approaches that are found to be the most effective will then be tested in up to 300,000 men.² The study was initiated in November 2025, and participants will be followed for at least 10 years.^2,3

“We hope that this simple, cheap spit test could turn the tide for prostate cancer,” Eeles said. “Building on decades of research into the genetic markers of prostate cancer, our study shows that the theory does work in practice – we can identify men at risk of aggressive cancers who need further tests and spare the men who are at lower risk from unnecessary treatments.”

References

1. McHugh JK, Bancroft EK, Saunders E, et al; BARCODE1 Steering Committee and Collaborators. Assessment of a polygenic risk score in screening for prostate cancer. N Engl J Med. 2025;392(14):1406-1417. doi:10.1056/NEJMoa2407934
2. First men invited to take part in the most ambitious prostate cancer trial in decades. News release. The Institute of Cancer Research. November 21, 2025. Accessed December 22, 2025. https://www.icr.ac.uk/about-us/icr-news/detail/first-men-invited-to-take-part-in-most-ambitious-prostate-cancer-trial-in-decades
3. Clarke H. TRANSFORM trial to evaluate screening methods in prostate cancer. Urology Times. May 2, 2024. Accessed December 24, 2025. https://www.urologytimes.com/view/transform-trial-to-evaluate-screening-methods-in-prostate-cancer