Multiphase magnetic resonance angiography of the abdominal and pelvic arteries: results of a bicenter multireader analysis.

RATIONALE AND

OBJECTIVES

The objective is to assess the diagnostic accuracy and interobserver variability of multiphase 3D gadolinium-enhanced magnetic resonance angiography (3D-Gd-MRA) for assessment of abdominal and pelvic vascular disease.

METHODS

In 41 patients from two different institutions multiphase 3D-Gd-MRA of the aorta and pelvis was performed using an identical scanning protocol. In a single breath-hold three to four consecutive phases were acquired. Stenoses in the renal arteries, and aorta and pelvic arteries were independently evaluated by three readers and compared with digital subtraction angiography. Interobserver variability was compared by means of a kappa statistic.

RESULTS

Accuracy for stenosis grading consistently ranged between 80% and 90% for all three readers in all vessel segments studied. Good interobserver agreement was found with kappa values exceeding 0.75. Vessel segments with delayed fill-in could be reliably detected on the multiple successive MRA phases. Overall, MRA was rated slightly superior to Digital Subtraction Angiography in terms of interobserver variability, diagnostic confidence and image quality.

CONCLUSIONS

Multiphase MRA is a highly robust technique with reproducible accuracy for different observers and different institutions. It can therefore be recommended for screening of atherosclerotic abdominal and pelvic disease.