Skip to main content
Premium Trial:

Request an Annual Quote

Prostate Cancer Polygenic Score Does Not Improve High-Grade Disease Prediction Over Clinical Factors

NEW YORK – A prostate cancer polygenic risk score comprised of 269 germline variants does not bolster clinicians' ability to predict aggressive disease in individuals undergoing a prior prostate biopsy, according to a new study by researchers at Vanderbilt University Medical Center.

"[W]e found that the PRS neither performed better than the clinical predictor, nor did it enhance the performance of the predictor, when predicting high-grade cancers," senior and corresponding author Jonathan Mosley, a researcher at VUMC, noted in an email. He added that this suggests the PRS has little clinical utility for guiding clinical decisions about diagnostic work-ups for patients.

As they reported in a research letter appearing in JAMA Internal Medicine on Monday, the researchers looked at genetic and clinical data for 655 participants from Vanderbilt's DNA biobank, known as BioVU, focusing on individuals of European or African ancestry between the ages of 40 and 80 years who had received a prostate biopsy in the past. The cohort included 341 individuals diagnosed with prostate cancer and 176 cancer cases of grade 2 or higher.

"[T]he important clinical question is, 'Can these PRS predict clinically significant cancer and discriminate these cancers from indolent cancers?'" Mosley explained.

When the team compared the prostate cancer PRS to an existing Prostate Biopsy Collaborative Group (PBCG) risk calculator — a tool used to gauge individuals' overall cancer risk and their risk of high-grade cancer — it found that the polygenic score was linked to a slight improvement in overall cancer prediction when evaluated alongside the PBCG risk calculator tool. Even so, the PRS did not enhance prediction of high-grade cancer.

The score "did not improve risk prediction of aggressive prostate cancer compared with a contemporary clinical risk predictor," Mosley and his colleagues wrote, noting that while the PRS "improved model discrimination for all cancers, improvement was less than has been observed for other validated prostate cancer biomarker predictors such as the Prostate Health Index."

The authors cautioned that their retrospective analysis did not evaluate prostate cancer predictions based on the 269-variant PRS alone and noted that more accurate risk stratification tools are yet to be developed, particularly in individuals from African American and other non-European ancestry groups.

In an accompanying commentary article, Robert Klein, a researcher at the Icahn School of Medicine at Mount Sinai, pointed to the importance of taking not only PRS performance but also population transferability into account when developing such risk scores.

In addition, he suggested that the search for a prostate cancer PRS that specifically identifies individuals at risk of aggressive or potentially deadly forms of the disease will likely require larger studies of individuals with high-grade prostate cancer patients followed over time, rather than analyses that include genetic risk variants found in patients with indolent or low-grade prostate cancer.

"Given the experience with [prostate specific antigen] testing, where a marker for any prostate cancer resulted in detection of many indolent cancers that did not need treatment, care will need to be taken that any PRS that makes it to the clinic can discriminate between indolent and potentially lethal prostate cancer," Klein wrote.