Urine spermine and multivariable Spermine Risk Score predict high-grade prostate cancer.

2021 
Background To investigate the role of urine spermine and Spermine Risk Score in prediction of high-grade prostate cancer (HGPCa, ISUP grade group ≥2). Methods Nine hundred and five consecutive men with elevated PSA were prospectively recruited from two hospitals. Core analyses focused on consecutive men with PSA 4-20 ng/mL (n = 600). Pre-biopsy urine without prior prostatic massage was analyzed for spermine level with ultra-high performance liquid chromatography with triple quadrupole mass spectrometer (UPLC-MS/MS). The proportions of PCa and HGPCa were compared across different spermine ranges. Logistic regressions were used to form different models, and their performances were compared using area under curve (AUC) and decision curve analysis (DCA). Results PCa and HGPCa were diagnosed in 30.8% (185/600) and 17.2% (103/600) men, respectively, and were significantly associated with lower urine spermine levels. Between the lowest and highest quartiles of spermine results, a threefold increase in PCa risk (49.3% vs. 16.7%) and 3.5-fold increase in ISUP grade group ≥2 PCa risk (31.3% vs. 8.7%) were observed. Multivariate analysis showed PSA, prostate volume (PV), digital rectal examination (DRE), and spermine, which were independent predictors for PCa and HGPCa, and a Spermine Risk Score with these factors achieved the highest AUC of 0.78 for PCa and 0.82 for HGPCa. At 90% sensitivity for HGPCa, 36.7% biopsies and 24.4% ISUP grade group 1 diagnoses could have been avoided, with a negative predictive value of 95.4%. DCA revealed net clinical benefit of the Spermine Risk Score. Internal validation with bootstrapping showed good discrimination and calibration. Conclusion Urine spermine and Spermine Risk Score identified men at higher risk of HGPCa and reduced unnecessary biopsies.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    17
    References
    3
    Citations
    NaN
    KQI
    []