Non-enhanced MR angiography of renal arteries: comparison with contrast-enhanced MR angiography

Abstract

Background The main causes of renal artery stenosis (RAS) are atherosclerosis and fibromuscular dysplasia. Despite contrast-enhanced magnetic resonance angiography (CE-MRA) being a safe and reliable method for diagnosis of RAS especially in young individuals, recently it has been possible to adopt innovative technologies that do not require paramagnetic contrast agents. Purpose To assess the accuracy of steady-state free-precession (SSFP) non-contrast-enhanced magnetic resonance angiography (NC-MRA) by using a 1.5 T MR scanner for the detection of renal artery stenosis, in comparison with breath-hold CE-MRA as the reference standard. Material and Methods Sixty-three patients (33 men, 30 women) with suspected renovascular hypertension (RVHT) were examined by a 1.5T MR scanner; NC-MRA with an electrocardiography (ECG)-gated SSFP sequence was performed in 58.7% (37/63) of patients; in 41.3% (26/63) of patients a respiratory trigger was used in addition to cardiac gating. CE-MRA, with a three-dimensional gradient echo (3D-GRE) T1-weighted sequence, was performed in all patients within the same session. Maximum intensity projection (MIP) image quality, number of renal arteries, and the presence of stenosis were assessed by two observers (independently for NC-MRA and together for CE-MRA). The agreement between NC-MRA and CE-MRA as well as the inter-observer reproducibility were calculated with Bland-Altman plots. Results MIP image quality was considered better for NC-MRA. NC-MRA identified 143 of 144 (99.3%) arteries detected by CE-MRA (an accessory artery was not identified). Fourteen stenoses were detected by CE-MRA (11 atherosclerotic, 3 dysplastic) with four of 14 (28.5%) significant stenosis. Bland-Altman plot demonstrated an excellent concordance between NC-MRA and CE-MRA; particularly, the reader A evaluated correctly all investigated arteries, while over-estimation of two stenoses occurred for reader B. Regarding NC-MRA, inter-observer agreement was excellent. Conclusion NC-MRA is a valid alternative to CE-MRA for the assessment of renal arteries.