Predictive power of the ESUR scoring system for prostate cancer diagnosis verified with targeted MR-guided in-bore biopsy

Purpose: This study evaluates the diagnostic value of the ESUR scoring system (PI-RADS) regarding prostate cancer detection using MR-guided in-bore biopsies (IB-GB) as the reference standard. Methods: 566 lesions in 235 consecutive patients (65.7 +/- 7.9 years, PSA 9.9 +/- 8.5 ng/m1) with a multiparametric (mp)-MRI (T2WI, DWI, DCE) of the prostate at 3 T were scored using the PI-RADS scoring system. PI-RADS single (PSsingle), summed (PSsum), and overall (PSoverall) scores were determined. All lesions were histologically verified by IB-GB. Results: Lesions with a PSsum below 9 contained no prostate cancer (PCa) with Gleason score (GS) > 4 + 3 = 7. A PS. of 13-15 (PSoverall V) resulted in 87.8% (n=108) in PCa and in 42.3% (n=52) in GS > 4 + 3 = 7. Transition zone (TZ) lesions with a PS. of 13-15 (PSoveraii V) resulted in 76.3% (n=36) in PCa and in 26.3% (n=10) in GS >= 4 + 3 = 7, whereas for peripheral zone (PZ) lesions cancer detection rate at this score was 92.9% (n=79) and 49.4% (n=42) for GS >= 4 + 3 = 7. Using a threshold of PSsum >= 10, sensitivity was 86.0%, and negative predictive value (NPV) was 86.2%. For higher grade PCa sensitivity was 98.6%, and NPV was 99.5%. Conclusion: A PSsum below 9 excluded a higher grade PCa, whereas lesions with a PSsum >= 13 (PSoverall V) represented in 88% PCa, and in 42% higher grade PCa. The PSsum or PSoverall demonstrated a better diagnostic value for PZ lesions with higher detection rates for higher grade PCa compared to TZ lesions. (C) 2014 Elsevier Ireland Ltd. All rights reserved.