High-resolution MR imaging of human atherosclerotic femoral arteries in vivo: validation with intravascular ultrasound.

PURPOSE

To establish a magnetic resonance (MR) imaging protocol for noninvasive in-vivo analysis of atherosclerotic femoral artery segments in humans and to compare the results to those of intravascular ultrasonography (IVUS).

MATERIALS AND METHODS

In seven patients with peripheral arterial occlusive disease, 20 femoral arterial segments per person were examined by high-resolution (HR) MR imaging and IVUS. Comparison was possible in 123 of 140 segments. MR imaging was performed at 1.5 T with use of a three-dimensional (3D) time-of-flight sequence with an in-plane resolution of 0.78 x 0.49 mm(2). 3D contrast-enhanced MR angiography was used for exact positioning of the HR MR imaging slices. IVUS (3.5 F, 40 MHz) was performed with use of a motorized pullback system. Parameters analyzed included cross-sectional lumen area (LA), vessel area (VA), and extent of vessel wall calcification.

RESULTS

Agreement between IVUS and HR MR imaging was analyzed with use of the Bland-Altman method. The paired LA measurements were in close agreement: the Bland-Altman mean bias in LA was -0.4 mm with a precision of +/-5.1 mm (P =.062). As a result of dorsal echo extinction in IVUS, VA measurements were feasible in only 74 of 140 segments. VA measurements were moderately correlated (r = 0.74; P <.0001), and a 25% overestimation by HR MR imaging compared to IVUS was observed. Intra- and interobserver comparisons for LA and VA measured with HR MR imaging did not show significant differences. Vessel wall calcifications were classified with a sensitivity of 91%, a specificity of 93%, and an accuracy of 93%.

CONCLUSIONS

The MR imaging protocol introduced in the present study permits precise assessment of LA and extent of calcification in peripheral arterial occlusive disease in vivo. HR MR imaging shows high concordance with IVUS and may have the potential for noninvasive therapy monitoring.