Review of source images is necessary for the evaluation of gadolinium-enhanced MR angiography for renal artery stenosis

Ziele: The purpose of this study was to assess interoberserver variability and accuracy in the evaluation of renal artery stenosis (RAS) with gadolinium-enhanced MR angiography (MRA) and digital subtraction angiography (DSA) in patients with hypertension. The authors find that source images are more accurate than maximum intensity projection (MIP) for depicting renal artery stenosis. Methode: Two independent radiologists reviewed MRA and DSA from 38 patients with hypertension. Studies were postprocessed to display images in MIP and source images. DSA, performed in each patient, was the standard for comparison. For each main renal artery, an estimate of percentage stenosis was made for any stenosis that was detected by the two radiologists. To calculate sensitivity, specificity and accuracy, MRA studies and stenoses were categorized as normal, mild (1–39%), moderate (40–69%) or severe stenosis (³70%), or occluded. DSA stenosis estimates of 70% or greater were considered hemodynamically significant. Ergebnis: Analysis of variance demonstrated that MIP estimates of stenosis were greater than source image estimates for both readers. Differences in estimates for MIP versus DSA reached significance in one reader. The interobserver variance for MIP, source images and DSA was excellent (0.80< κ£ 0.90). The specificity of source images was high (97%) but less for MIP (87%); average accuracy was 92% for MIP and 98% for source images. Schlussfolgerung: In this study, source images are significantly more accurate than MIP images in one reader. A similar trend was observed in the second reader. The interobserver variability was excellent. When renal artery stenosis is a consideration, high accuracy can only be obtained when source images are examined.

Key words

Magnetic resonance - angiography - renal artery stenosis - MIP - Source Images