RESUMO
PURPOSE: To evaluate the diagnostic and prognostic performance of Secreted Protein Acidic and Rich in Cysteine (SPARC) in detecting urinary bladder cancer (UBC). METHODS: The Integrated Study on Bladder Cancer (n = 571; mean age:69.4 ± 12.2 years) evaluates cross-sectionally SPARC diagnostic performance in primary (n = 264) and recurrent (n = 307) UBC. SPARC prognostic performance is evaluated in a nested cohort (n = 250) prospectively followed for 80 months to detect tumor relapse, recurrence and/or progression. Baseline urine samples are analyzed blindly using a commercially available SPARC ELISA assay, characterized for its analytical performance according to clinical test regulatory requirements (R&D Manufactures Inc.). RESULTS: While higher mean SPARC levels are detected in primary (p = 0.008) and recurrent (p < 0.0001) UBC, the assay has limited diagnostic performance (AUC:0.593; 95% CI:0.524-0.663). SPARC positive patients undergoing disease monitoring are more likely to develop tumor relapse (age and gender Adj. HR:1.52; 95% CI:1.04-2.22) and progression (Adj. HR:1.83; 95% CI:1.02-3.27). However, prognostic performance is affected by hematuria. CONCLUSIONS: SPARC diagnostic performance for detecting UBC appears insufficient for clinical implementation. In patients undergoing disease monitoring, SPARC is a promising prognostic marker for tumor relapse and/or progression, but is affected by hematuria.