Blood test shows predictive accuracy for high-grade prostate cancer

EV-Fingerprint, a blood-based biomarker test, can accurately predict patients who have grade group 3 or greater prostate cancer and may reduce unnecessary biopsies, according to data published in Cancer Medicine.¹

Investigators found that the test was able to predict which patients had grade group 3 or greater prostate cancer with 95% accuracy.

According to a news release on the test, “The technology measures levels of prostate cancer biomarkers in a patient’s blood sample, combines that data with their clinical information, then uses machine learning to generate a risk score that predicts the presence of clinically significant prostate cancer.”²

In total, the study of the test included 415 men who were referred for biopsies on the basis of high PSA levels from June 2014 to January 2017 in Alberta, Canada. Of those, 157 (38%) men had a negative biopsy, 258 (62%) received a diagnosis of any prostate cancer, and 73 (18%) were diagnosed with grade group 3 or greater prostate cancer.

Investigators found that the test was able to predict which patients had grade group 3 or greater prostate cancer with 95% accuracy (0.81 area under the curve [AUC]). The negative predictive value of the test was 97%.

Overall, 144 (35%) patients included in the study could have avoided the recommended biopsy when using a probability cutoff of at least 7.85%. Grade group 3 or greater prostate cancer would have gone undetected in 4 (5%) men. At a risk threshold of at least 5%, 35 (7%) patients could have avoided an unnecessary biopsy, and no grade group 3 or greater prostate cancer would have been missed.

“The ClarityDX Prostate test will reduce the number of unnecessary prostate biopsies, which are invasive, uncomfortable, and carry some risk,” stated John D. Lewis, PhD, in a news release.² Lewis is the Bird Dogs Chair in translational oncology at the University of Alberta, and the CEO of Nanostics Inc, a University of Alberta spinoff company.

In comparison with the online risk calculator PCPTRC 2.0 alone, the EV-Fingerprint test demonstrated a greater specificity in predicting grade group 1 or greater disease, with 20% specificity for EV-Fingerprint compared with 15% with PCPTRC 2.0 alone. Both tests showed an AUC of 0.69 and similar negative predictive values.

In predicting grade group 2 or greater prostate cancer, EV-Fingerprint demonstrated a greater AUC but a lower specificity and negative predictive value compared with PCPTRC 2.0 alone.

Further, prediction of grade group 3 or greater prostate cancer with EV-Fingerprint demonstrated significantly superior AUC (0.81 vs 0.73), specificity (41% vs 24%), and negative predictive value (97% vs 95%) compared with PCPTRC 2.0 alone.

The authors note that a larger prospective study is being conducted for clinical validation of the test in a cohort of men from Canada and the US.

According to the news release, Nanostics is also currently seeking Health Canada and US FDA approval of the test.²

References

1. Fairey A, Paproski RJ, Pink D, et al. Clinical analysis of EV-Fingerprint to predict grade group 3 and above prostate cancer and avoid prostate biopsy. Cancer Med. 2023;12(15):15797-15808. doi:10.1002/cam4.6216

2. Blood test for prostate cancer could help avoid unnecessary biopsies. News release. University of Alberta. October 5, 2023. Accessed October 15, 2023. https://www.ualberta.ca/folio/2023/10/blood-test-for-prostate-cancer-could-help-avoid-unnecessary-biopsies.html