Background: Early prostate cancer antigen (EPCA), a nuclear matrix protein, has been recently suggested as a novel biomarker in malignant lesions of the prostate. This study was to determine whether preoperative serum EPCA levels predicted the presence of incidental prostate cancer (IPCa) in patients undergoing TURP for BPH.
Methods: Serum EPCA levels were measured by ELISA in 449 consecutive patients with symptomatic BPH treated with TURP and 112 healthy men. Predictive performance of serum EPCA levels for IPCa were evaluated.
Results: With a cutoff of 10ng/ml, serum EPCA protein had a 100% specificity for the healthy men and a 98% specificity and a 100% sensitivity in separating men with IPCa from those without. Serum EPCA levels in patients with IPCa were significantly higher than in those without and in healthy controls (17.63±2.42ng/ml vs. 5.58±1.61 ng/ml and 4.95±1.43 ng/ml, all P<0.001), whereas an indwelling transurethral catheter presence and 5α-reductase inhibitor therapy had no effect on EPCA levels (P=0.144 and P=0.238, respectively). The area under ROC curves (AUC) showed that serum EPCA level had the best predictive accuracy of all IPCa (AUC: 0.952, 95% CI: 0.912-0.981, P<0.001). Univariate and multivariate Cox regression analyses further demonstrated the independently predictive performance by preoperative serum EPCA (Hazards Ratio: 4.23, 95% CI: 3.62-6.46, P<0.001).
Conclusions: This study firstly shows that EPCA might be used as a highly sensitive and specific serum biomarker to predict IPCa presence and to help reduce the unnecessary biopsies taken before TURP in patients with BPH.
© 2010 Wiley-Liss, Inc.