Accuracy of Phase-Contrast Velocity Mapping Proximal and Distal to Stent Artifact During Cardiac Magnetic Resonance Imaging.

Year of Publication


Number of Pages


Date Published

2018 Jun 15

ISSN Number



<p>Little data are available on the accuracy of phase-contrast magnetic resonance imaging (PC-MRI) velocity mapping in the vicinity of intravascular metal stents other than nitinol stents. Therefore, we sought to determine this accuracy using in vitro experiments. An in vitro flow phantom was used with 3 stent types: (1) 316L stainless steel, (2) nitinol self-expanding, and (3) platinum-iridium. Steady and pulsatile flow was delivered with a magnetic resonance imaging-compatible pump (CardioFlow 5000, Shelley Medical, London, Ontario, Canada). Flows were measured using a transit time flow meter (ME13PXN, Transonic, Inc, Ithaca, New York). Mean flows ranged from 0.5 to 7 L/min. For each condition, 5 PC-MRI acquisitions were made: within the stent, immediately adjacent to both edges of the stent artifact, and 1 cm upstream and downstream of the artifact. Mean PC-MRI flows were calculated by segmenting the tube lumen using clinical software (ARGUS, Siemens, Inc, Erlangen, Germany). PC-MRI and flow meter flows were compared by location and stent type using linear regression, Bland-Altman, and intraclass correlation (ICC). PC-MRI flows within the stent artifact were inaccurate for all stents studied, generally underestimating flow meter-measured flow. Agreement between PC-MRI and flow meter-measured flows was excellent for all stent types, both immediately adjacent to and 1 cm away from the edge of the stent artifact. Agreement was highest for the platinum-iridium stent (R = 0.999, ICC = 0.999) and lowest for the nitinol stent (R = 0.993, ICC = 0.987). In conclusion, PC-MRI flows are highly accurate just upstream and downstream of a variety of clinically used stents, supporting its use to directly measure flows in stented vessels.</p>



Alternate Title

Am. J. Cardiol.




Subscription is not available for this page.