Gadobenate Dimeglumine and Gadofosveset Trisodium for MR Angiography of the Renal Arteries: Multicenter Intraindividual Crossover Comparison

Abstract

This prospective multicenter intraindividual crossover study was designed to compare gadobenate dimeglumine and gadofosveset trisodium at approved doses with respect to the image quality and diagnostic performance of contrast-enhanced MR angiography (CE-MRA) in the detection of clinically relevant renal artery stenosis. Thirty-nine subjects (17 men, 22 women; age range, 30-86 years; mean 62 +/- 13.3 [SD] years) with known or suspected renovascular disease underwent a first CE-MRA examination with 0.1 mmol/kg gadobenate dimeglumine and a second examination 3-12 days later with 0.03 mmol/kg gadofosveset. Identical T1-weighted spoiled gradient-refocused echo coronal first-pass images were acquired for 38 of the 39 patients. For 15 of the 38 patients, additional sagittal or axial images or both were acquired with gadofosveset during the steady-state phase. Thirty-four patients underwent digital subtraction angiography, which was the reference standard. Three independent blinded readers assessed source images and maximum-intensity-projection reconstructions to detect clinically relevant (> 50%) renal artery stenosis. Diagnostic performance (sensitivity, specificity, accuracy, positive and negative predictive values) was evaluated with the McNemar and Wald tests. Matched-pair determinations of diagnostic preference were evaluated with Wilcoxon's signed rank test. Reader agreement was determined with kappa analysis, and safety was assessed. Comparison of first-pass images revealed superior sensitivity (75.7-86.5% vs 68.4-76.3%), specificity (92.1-98.6% vs 90.5-93.9%), accuracy (88.9-96.2% vs 85.9-90.3%), positive predictive value (70.0-94.1% vs 65.0-76.3%), and negative predictive value (94.0-96.6% vs 91.7-93.9%) with gadobenate dimeglumine. Significant superiority was noted for specificity (p < or = 0.02), accuracy (p < or = 0.005), and positive predictive value (p < or = 0.018). Steady-state images showed no benefit of gadofosveset. Reader agreement was good to excellent (gadobenate dimeglumine, kappa = 0.855; gadofosveset, kappa = 0.776). Reader preference was for gadobenate dimeglumine in 11, 17, and 13 patients and for gadofosveset in five, four, and five patients. No safety concerns were noted. Better diagnostic performance and reader preference were found for gadobenate dimeglumine than gadofosveset in first-pass renal CE-MRA.