Inter-reader agreement of the ESUR score for prostate MRI using in-bore MRI-guided biopsies as the reference standard.

Abstract

The recent European Society of Urogenital Radiology (ESUR) guidelines for evaluation and reporting of prostate multiparametric magnetic resonance imaging (mp-MRI) include the Prostate Imaging Reporting and Data System (PI-RADS). The aim of this study was to investigate the inter-reader agreement of this scoring system. One hundred and sixty-four lesions in 67 consecutive patients with elevated prostate-specific antigen and previously negative trans-rectal ultrasound (TRUS)-guided biopsy were scored retrospectively by three blinded readers using PI-RADS. Mp-MRI was performed at 3 T using T2-weighted, diffusion-weighted and dynamic contrast-enhanced imagings (T2WI, DWI, DCE-MRI). Histology of all lesions was obtained by in-bore MRI-guided biopsy. Cohen's kappa statistics were calculated for all readers. Inter-reader agreement for all lesions was good to moderate (T2WI, κ = 0.55; DWI, κ = 0.64; DCE-MRI, κ = 0.65). For tumour lesions it was good (T2WI, κ = 0.66; DWI, κ = 0.80; DCE-MRI, κ = 0.63) and for benign lesions moderate to good (T2WI, κ = 0.46; DWI, κ = 0.52; DCE-MRI, κ = 0.67). Using an overall PI-RADS score with a threshold of ≥10, we achieved a sensitivity of 85.7 %, and negative predictive value of 90.1 % for biopsied lesions. PI-RADS score shows good to moderate inter-reader agreement and enables standardised evaluation of prostate mp-MRI, with high sensitivity and negative predictive value. • The European Society of Urogenital Radiology recently published guidelines for prostate MRI. • We have evaluated inter-reader agreement of ESUR scoring for multiparametric prostate MRI. • PI-RADS shows good to moderate inter-reader agreement and is clinically applicable. • PI-RADS achieves in our series high sensitivity and negative predictive value for biopsied lesions. • PI-RADS can be used as standardised scoring system in prostate cancer detection.