Morphologic and functional magnetic resonance imaging of renal artery stenosis: a multireader tricenter study

Abstract

The effect of combined morphologic and functional magnetic resonance (MR) imaging on the interobserver and intermodality variability for the grading of renal artery stenosis is assessed. In a randomized, blinded tricenter analysis, seven readers evaluated 43 renal arteries on x-ray digital subtraction angiography (DSA), 3D-Gadolinium MR angiography (3D-Gd-MRA), cine phase-contrast flow measurement (PC-flow), and a combined analysis of the last two. Interobserver variability was assessed for the grading of renal artery stenosis as well as regional vessel visibility. Intermodality variability for stenosis grading was analyzed in cases in which the readers agreed on the degree of stenosis in DSA. DSA had a substantial interobserver variability for the grading of stenosis (mean kappa kappa 0.64). 3D-Gd-MRA revealed a slightly improved interobserver variability but incorrectly graded 6 of 34 stenoses on a two-point scale (<50%, > or =50%). The combined approach of 3D-Gd-MRA and PC-flow revealed the best (P = 0.0003) interobserver variability (median kappa = 0.75) and almost perfect intermodality agreement with DSA (97% of cases). These findings were confirmed in a prospective analysis of 97 renal arteries. The vessel visibility of the renal artery ostium was significantly better in 3D-Gd-MRA than in DSA, whereas the visibility of the hilar and intrarenal vessels was significantly worse (P = 0.0001). A combined morphologic and functional MR examination significantly reduces interobserver variability and offers reliable and reproducible grading of renal artery stenosis based on stenosis morphology and hemodynamic changes. It can be considered a safe and noninvasive alternative for diagnostic DSA in cases that do not require assessment of intrarenal vessels.