PD11-02: Evaluation of variety and performance between PSA assays and PHI in a clinical cohort of men undergoing prostate biopsy for suspected prostate cancer

Basil Kaufmann*, Paloma Pellegrino, Noémie Lautenbach, Katharina Spanaus, Arnold von Eckardstein, Thomas Hermanns, Cédric Poyet, Zurich, Switzerland

Introduction: Prostate specific antigen (PSA) testing is of paramount importance as a diagnostic tool for detecting and monitoring patients with prostate cancer. Less is known about how different PSA tests perform in the same men and which additional tests (i.e., %fPSA and/or PHI) might help distinguish which men harbours significant prostate cancer and which do not. The aim of this study was to test different PSA assays and PHI in the same men undergoing prostate biopsy for suspected prostate cancer.

Methods: Clinical data from a consecutive group of men who underwent mpMRI/fusion and transperineal template prostate biopsy were prospectively gathered. Five different commercially available PSA assays (Roche, Diasorin, Brahms, Abott, Beckmann) and PHI (Beckmann) were tested in an additional blood sample taken before prostate biopsy. The results of all six assays were compared directly and in relation to clinical data and biopsy outcome. To study the different assay performance in predicting significant prostate cancer (defined as ISUP grading group =2), the area under the curve (ROC test) was used.

Results: Seventy-five men were available for final analysis. A total of 29 (38.7%) were diagnosed with significant PCa and in 46 (61.3%) the biopsy showed insignificant PCa (defined as ISUP 1) or no cancer. We found that some PSA assays consistently showed relevant higher or lower values across all PSA levels compared to our internal control assay from Roche. Compared to Roche, Brahms showed consistently higher values and Beckmann and Abott showed uniformly lower values, whereas Diasorin showed similar values in most blood samples tested. In contrast to total PSA, %FPSA showed less variety between the assays when tested in the same blood sample. The different assays including PHI, were then tested for their performance in detecting significant PCa using the AUC method. The average AUC value for the five PSA assays was 0.603 (ranging from 0.596 to 6.12), whereas PHI showed by far the highest AUC value with 0.808.

Conclusions: We found relevant and constant differences in PSA levels between different tested PSA providers within the same blood sample. Thus, all clinical studies reporting PSA data should report the provider used. Furthermore, clinicians should know the specifics of the PSA test used in daily clinical practice for accurate and better interpretation of the test result. Finally, PHI outperformed all tested PSA assays in its ability to predict significant PCa before prostate biopsy and therefore could add value in intial PCa risk assessment.

Source of Funding: This study received no financial support.