Renal artery stenosis: comparative evaluation of gadolinium-enhanced MRA and DSA

Abstract

Purpose: This study was undertaken to evaluate the accuracy of contrast-enhanced magnetic resonance angiography (CE-MRA) in detecting renal artery stenosis using intra-arterial digital subtraction angiography (DSA) as the gold standard.

Materials and methods: Thirty-five consecutive patients with possible renovascular hypertension were prospectively studied; 26 of them underwent both MRA and DSA. In these 26 cases, two readers assessed the number of renal arteries, the presence of stenoses and their degree. Results were compared with DSA, and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and diagnostic accuracy of MRA were determined. Interobserver variability was also calculated.

Results: DSA showed 51 main renal arteries (one patient had a single kidney) and six accessory arteries (total number of arteries 57) in the 26 patients considered. Both MRA readers detected all of the 51 main renal arteries and only one accessory vessel. When the presence of stenosis was considered, the readers' results, respectively, were as follows: sensitivity 77% and 72%, specificity 69% and 69%, PPV 86% and 85%, NPV 55% and 50% and diagnostic accuracy 75% and 71%. When the detection of significant stenosis was considered, the results, respectively, were: sensitivity 83% and 83%, specificity 73% and 78%, PPV 60% and 65%, NPV 90% and 91%, and diagnostic accuracy 76% and 80%. Interobserver variation was good when considering stenosis detection (kappa=0.69) and excellent when considering detection of significant stenosis (kappa=0.85).

Conclusions: MRA results do not appear as positive as in the majority of papers in the literature. Multiple reasons can probably be invoked to explain this difference. The mean age of our patients, higher than in many other studies, should be noted and may have accounted for their possible poor cooperation. Moreover, all of the missed significant stenoses were distally located, and therefore, the failure to detect them might be related to the suboptimal spatial resolution of MRA. Nevertheless, MRA showed a high NPV for detecting significant stenoses, a finding of considerable clinical relevance in that it allows patients with normal MRA findings to be spared additional more invasive procedures.