Peripheral vascular tree stenoses: Detection with subtracted and nonsubtracted MR angiography

PURPOSE: To compare subtracted and nonsubtracted gadolinium-enhanced magnetic resonance (MR) angiography and cardiac-synchronized time-of-flight MR angiography for help in detecting pelvic-region stenoses. MATERIALS AND METHODS: Twenty-eight patients with intermittent claudication underwent MR angiography with a 1.5-T system; two-dimensional cardiac-synchronized time-of-flight MR angiograms and three-dimensional MR angiograms (without and with gadolinium enhancement) were obtained. Subtracted images were obtained by subtracting unenhanced data from enhanced data of identical volumes, and maximum intensity projection images were constructed, which two observers independently evaluated in blinded fashion, with conventional angiographic results as the reference standard. RESULTS: Sensitivity and specificity for grading of hemodynamically significant stenoses (greater than or equal to 50% lumen reduction) on subtracted MR angiograms were 94% and 93%, respectively. Sensitivity of subtracted images was significantly higher compared with that of time-of-flight images (P < .05) but not with that of nonsubtracted images. Contrast-to-noise ratio on subtracted images was significantly higher compared with that on nonsubtracted images (P < .05) but not with that on time-of-flight images. There was good correlation between stenosis length measurements on gadolinium-enhanced MR angiograms and those on conventional angiograms. CONCLUSION: Subtracted MR angiography is superior to cardiac-synchronized time-of-flight MR angiography for imaging of iliac and upper femoral arteries and provides higher contrast-to-noise ratio, fewer artifacts, and easier image interpretability than nonsubtracted MR angiography.