Martin Sanda, MD
An assay based on a prostate-specific antigen (PSA) isoform performed significantly better than did conventional PSA measures for detecting prostate cancer and identifying histopathologically aggressive cancers, according to a multicenter cohort study.
The Prostate Health Index (PHI) (Beckman Coulter), recently approved by the FDA to evaluate the probability of a prostate cancer diagnosis, not only demonstrated superior accuracy versus PSA and percent-free PSA (fPSA) for predicting the presence of prostate cancer, but also outperformed PSA testing for prediction of aggressive prostate cancer.
At a 90% threshold of sensitivity for prostate cancer detection, the PHI had a specificity of 31.1% compared with 19.8% and 10.8% for fPSA and PSA, respectively. Corresponding values for histopathologically aggressive prostate cancer were 30.1%, 21.7%, and 9.6% for PHI, fPSA, and PSA.
“The ability to identify aggressive cancers more accurately and reduce the number of unnecessary biopsies of indolent cancers by use of the PHI ultimately will help make it possible to retain the benefits of prostate cancer screening while reducing possible harms of overdetection and overtreatment,” said principal investigator Martin Sanda, MD, chair of Urology at Emory University in Atlanta, Georgia, during a presentation at the 2013 Annual Meeting of the American Urological Association (AUA).
Although PSA testing has revolutionized detection and treatment of prostate cancer, the test has well-recognized limitations. The presence of PSA does not necessarily mean a man has cancer, and circulating levels of the protein can change in response to a variety of circumstances, including physical activity, sexual activity, benign prostatic hyperplasia, and chronic prostatitis.
Among promising biomarker alternatives to PSA, the PSA isoform [-2] proPSA (p2PSA) has been evaluated extensively, as reflected in the FDA’s recent approval of the PHI. The p2PSA-based assay is derived from p2PSA divided by fPSA, multiplied by the square root of total PSA.
During a poster session at the AUA meeting, Sanda presented results from a multicenter, case-control, prospective cohort study. Investigators enrolled nearly 900 men ≥50 whose digital rectal exams were negative for cancer, and who had not yet undergone prostate biopsy. The primary analysis comprised 658 men with total PSA levels of 4-10 ng/mL—334 with biopsy-proven prostate cancer and 324 without cancer.
Blood samples from each participant were evaluated by the PHI, PSA, and fPSA assays. Investigators determined sensitivity and specificity of each assay by means of receiver operating characteristic area under the curve (AUC), and the assays’ accuracy for prostate cancer detection were compared, using the PHI as the reference.
The PHI resulted in an AUC of 0.708 compared with 0.648 for fPSA (P
=.009) and 0.516 for PSA (P
<.001). A comparison of the assays’ ability to identify aggressive prostate cancer yielded an AUC of 0.698 for the PHI, 0.654 for fPSA (P
=.091), and 0.549 for PSA (P
The study population included 160 men with “clinically significant” prostate cancer, as defined by Epstein criteria. The remaining patients had either clinically insignificant or benign disease.
Sanda reported that the PHI value correlated significantly with the probability of cancer and with clinically significant cancer. A score of 0-26.9 was associated with a cancer probability of 9.8% and a 3.9% probability of clinically significant cancer. At the top of the value range, a PHI of ≥55 was associated with a 50.1% probability of cancer and a 28.9% probability of clinically significant cancer.
The PHI cutoffs retained their predictive accuracy for aggressive prostate cancer whether aggression was defined by Epstein criteria (P
=.009), biopsy Gleason score ≥3+4 (P<.001), or biopsy Gleason score ≥4+3 (P
At the 90% sensitivity cut-point for PHI (PHI=27), 120 of 639 (18.8%) men (104 biopsy-negative for cancer and 16 with clinically insignificant cancer) could have been spared from undergoing prostate biopsy or overdiagnosis of nonaggressive disease, the authors wrote.
“Using the PHI to measure p2PSA can improve the specificity of prostate cancer detection,” Sanda concluded. “The PHI can facilitate detecting histopathologically aggressive, clinically significant prostate cancer, whether by Epstein criteria or Gleason score. The PHI can help reduce overdetection of clinically indolent prostate cancer.”
Sanda M, Wei J, Broyles D, et al. Prostate health index for reducing overdetection of indolent prostate cancer and unnecessary biopsy while improving detection of aggressive cancers. Presented at: the Annual Meeting of the American Urological Association; May 4-8, 2013; San Diego, CA. Abstract 2055.