Genomic Prostate Score may help refine risk assessment in localized prostate cancer

Findings from the retrospective GPS-ProMPT study have been published in European Urology Oncology, suggesting that the 17-gene Genomic Prostate Score (GPS) assay may identify patients with localized or locally advanced prostate cancer whose tumor biology is more favorable than conventional clinical risk classification would suggest.¹

The investigators reported consistently favorable 6-year oncologic outcomes among patients with low genomic risk despite intermediate- or high-risk National Comprehensive Cancer Network (NCCN) disease. The analysis, which included 409 patients with at least 6 years of follow-up, found that nearly half of clinically higher-risk patients had a low GPS result. These "discordant low-risk" patients experienced outcomes comparable to those of patients with both low clinical and genomic risk, supporting prospective evaluation of genomics-guided treatment de-escalation.

“The ProMPT findings demonstrate that patients with favorable GPS mdx scores achieved excellent long-term outcomes—including 92% freedom-from-failure at 6 years in active surveillance and up to 100% in surgery and radiotherapy cohorts—even among those with adverse clinical risk factors,” said lead author Nikita Sushentsev, MD, PhD, of Cambridge University Hospitals NHS Foundation of Trust, in a news release on the results.² “This suggests GPS mdx can identify favorable tumor biology that traditional risk stratification alone would miss, with meaningful implications for how we counsel patients facing treatment decisions.”

About the study

The study evaluated archived diagnostic biopsy specimens from 409 men with valid GPS results. Of these, 357 (87%) had clinically localized disease managed with active surveillance (AS; n = 103), radical prostatectomy (RP; n = 155), or radiotherapy plus 3 to 6 months of androgen deprivation therapy (ADT; n = 99). An additional 52 patients (13%) with locally advanced disease received radiotherapy with at least 18 months of ADT. Median follow-up among patients without an event was 6 years or longer across all treatment cohorts.

Across treatment settings, higher GPS values were associated with significantly worse oncologic outcomes. A GPS score of at least 25 was associated with a significantly increased risk of AS failure (HR, 4.9; 95% CI, 1.3 to 18.8; P = .03). In patients undergoing RP or radiotherapy, a GPS score of at least 40 identified all metastatic events and high-risk biochemical recurrence events (P < .001 and P = .02, respectively). In the exploratory cohort with locally advanced disease, a GPS score of at least 55 was associated with a significantly higher risk of treatment failure (HR, 7.1; 95% CI, 1.8 to 28.4; P = .03).

Investigators then assessed whether patients with favorable genomic profiles but higher clinical risk might represent candidates for future management de-escalation. Of the overall study population, 58% had GPS scores below prespecified thresholds. These patients were categorized as either "concordant low-risk," in whom genomic and NCCN clinical risk agreed, or "discordant low-risk," in whom a low GPS score contrasted with higher NCCN risk classification.

The discordant low-risk phenotype was common across treatment groups. Among patients managed with AS who had NCCN risk group 2 or greater, 46% had low GPS scores. Similar discordance was observed in 57% of RP patients with NCCN risk group 3 or greater, 36% of radiotherapy patients with NCCN risk group 4 or greater, and 48% of patients with locally advanced disease classified as NCCN risk group 5.

Long-term outcomes in these discordant patients remained favorable. Six-year event-free survival was 88% among patients receiving AS, 100% among those treated with RP or radiotherapy, and 95% in the locally advanced cohort. All patients classified as concordant low-risk remained event-free through 6 years across all treatment groups.

Interpretations and future directions

The GPS assay, commercially available as GPS mdx, measures expression of 17 genes involved in multiple biologic pathways associated with prostate cancer aggressiveness. The results may be used to estimate the risk of adverse pathology, high-grade disease, non-organ confined disease, metastasis within 10 years, prostate cancer death within 10 years and disease progression beyond conventional clinical factors. The assay is intended to complement established clinical risk models.

The current analysis focused specifically on whether genomic risk might identify patients whose favorable tumor biology could justify less intensive treatment or surveillance despite adverse clinical features. However, the authors emphasized that the retrospective design, relatively small number of outcome events, and exploratory nature of the de-escalation analysis preclude changes in current management.

Instead, the investigators suggest that the findings provide a rationale for future trials of genomics-guided de-escalation strategies. The publication serves as a scientific foundation for the ongoing GPS-ProtecT study, which is evaluating GPS mdx within the United Kingdom ProtecT trial cohort. That effort aims to generate prospective randomized evidence regarding the clinical utility of genomic classification in men undergoing active surveillance for localized prostate cancer.

“The acceptance of Professor Freddie Hamdy’s GPS-ProMPT study in European Urology Oncology is an important milestone in our collaboration with the University of Oxford. These findings provide the scientific foundation for the ongoing GPS-ProtecT study, which has the potential to establish GPS mdx as the first genomic classifier supported by randomized clinical trial evidence in localized prostate cancer,” explained Michael K. McGarrity, CEO of mdxhealth, in the news release.² “This publication supports our advancement of GPS with regard to guideline development, payer coverage, and physician adoption.”

REFERENCES

1. Sushentsev N, Colling R, Teague R, et al. Genomic Prostate Score for Identifying Biologically Low-risk Disease in Patients with Prostate Cancer Undergoing Active Surveillance, Surgery, or Radiotherapy. Eur Urol Oncol. 2026:S2588-9311(26)00176-8. doi:10.1016/j.euo.2026.06.008

2. Mdxhealth announces peer-reviewed publication of Oxford GPS-ProMPT study in European Urology Oncology, establishing scientific foundation for the landmark GPS-ProtecT randomized controlled trial. News release. mdxhealth. June 29, 2026. Accessed July 14, 2026. https://www.globenewswire.com/news-release/2026/06/29/3318913/0/en/mdxhealth-announces-peer-reviewed-publication-of-oxford-gps-prompt-study-in-european-urology-oncology-establishing-scientific-foundation-for-the-landmark-gps-protect-randomized-con.html