Diagnostic accuracy of prostate health index to identify aggressive prostate cancer. An Institutional validation study.

Introduction: New generations of tumor markers used to detect prostate cancer (PCa) should be able to discriminate men with aggressive PCa of those without PCa or nonaggressive tumors. The objective of this study has been to validate Prostate Health Index (PHI) as a marker of aggressive PCa in one academic institution.
Methods: PHI was assessed in 357 men scheduled to prostatic biopsy between June of 2013 and July 2014 in one academic institution. Thereafter a subset of 183 men younger than 75 years and total PSA (tPSA) between 3.0 and 10.0 ng/mL, scheduled to it first prostatic biopsy, was retrospectively selected for this study. Twelve cores TRUS guided biopsy, under local anaesthesia, was performed in all cases. Total PSA, free PSA (fPSA), and [-2] proPSA (p2PSA) and prostate volume were determined before the procedure and %fPSA, PSA density (PSAd) and PHI were calculated. Aggressive tumors were considered if any Gleason 4 pattern was found. PHI was compared to %fPSA and PSAd through their ROC curves. Thresholds to detect 90%, 95% of all tumors and 95% and 100% of aggressive tumors were estimated and rates of unnecessary avoided biopsies were calculated and compared.
Results: The rate of PCa detection was 37.2% (68) and the rate of aggressive tumors was 24.6% (45). The PHI area under the curve was higher than those of %fPSA and PSAd to detect any PCa (0.749 vs 0.606 and 0.668 respectively) or to detect only aggressive tumors (0.786 vs 0.677 and 0.708 respectively), however, significant differences were not found. The avoided biopsy rates to detect 95% of aggressive tumors were 20.2% for PHI, 14.8% for %fPSA, and 23.5% for PSAd. Even more, to detect all aggressive tumors these rates dropped to 4.9% for PHI, 9.3% for %fPSA, and 7.9% for PSAd.
Conclusions: PHI seems a good marker to PCa diagnosis. However, PHI was not superior to %fPSA and PSAd to identify at least 95% of aggressive tumors.
(Copyright © 2016 AEU. Publicado por Elsevier España, S.L.U. All rights reserved.)