AbstractIn 2013, thiosulfate in urine has been proposed as promising prostate cancer (PCa) biomarker. However, a missing comparison with other proven PCa markers suggested a re-evaluation study. Therefore, together with the authors from the initial study, the diagnostic accuracy of thiosulfate was compared with that of urinary prostate cancer associated 3 (PCA3), serum prostate health index (Phi), and percent free prostate-specific antigen (%fPSA). Thiosulfate was further measured in a multicenter approach to exclude center-related biases.Thiosulfate, calculated as ratio of thiosulfate to urinary creatinine (TS/Crea ratio), was measured in two cohorts in a total of 269 patients. In the retrospective study (n=160) PCA3, Phi, PSA, and %fPSA were compared with the TS/Crea ratio between patients with and without PCa according to the prostate needle biopsy results. The second prospective cohort included 109 patients from four centers.The median TS/Crea ratio was not statistically different between the patients with and without PCa. The receiver-operating characteristics showed that the TS/Crea ratio was unable to discriminate between patients with and without PCa in contrast to %fPSA, Phi, and PCA3. In all four centers, the low median TS/Crea ratios (<1 mmol/mol) in both patient cohorts were confirmed and thiosulfate was again not able to distinguish between them (p-values, 0.13–0.90).This study could not confirm the previously observed high median TS/Crea ratio in PCa patients in comparison to non-PCa patients. Thiosulfate subsequently failed as PCa biomarker while PCA3 and Phi showed the expected diagnostic improvement.