Inter-observer variability in male pelvic-floor MRI measurements that might predict post-prostatectomy incontinence.

Abstract

Current studies are focusing on the relationship between anatomical variables in preoperative prostate MRI and the development of post-prostatectomy incontinence (PPI). Nevertheless, there is little evidence regarding the reliability of these measurements. The objective of this study was to analyze the concordance between urologists and radiologists for anatomical measurements that might be PPI predictors. Pelvic floor measurements with 3T-MRI were performed by two radiologists and two urologists independently and blindly. Interobserver agreement was determined using the intraclass correlation coefficient (ICC) and the Bland-Altman plot. The concordance was good-acceptable for most measurements, except for the levator ani and puborectalis muscle thickness (some ICC values < 0.20/p value > 0.05). The anatomical parameters with the highest degree of agreement were intravesical prostatic protrusion (IPP) and prostate volume (most of the ICC values > 0.60). The membranous urethral length (MUL) and the angle of the membranous urethra-prostate axis (aLUMP) presented ICC > 0.40. The obturator internus muscle thickness (OIT), urethral width and intraprostatic urethral length presented a fair-moderate degree of agreement (ICC > 0.20). Regarding the agreement between different specialists, the highest degree was obtained for the two radiologists and for urologist 1-radiologist 2 (moderate median agreement), while urologist 2 with each of the radiologists had a regular median agreement. MUL, IPP, prostate volume, aLUMP, OIT, urethral width and prostatic length show acceptable inter-observer concordance and they would be reliable as possible predictors of PPI. Levator ani and puborectalis muscle thickness show bad agreement. Interobserver agreement might not be greatly influenced by previous professional experience.