Ex-Vivo Validation of Cardiac Magnetic Resonance Velocity Mapping For Quantification of Aortic Regurgitation In a Porcine Model In The Presence Of a Transcatheter Heart Valve

摘要

Purpose: Paravalvular aortic regurgitation (PAR) following transcatheter aortic valve implantation (TAVI) is associated with impaired outcome. However, no data exist regarding maintenance of the accuracy of PAR measured by cardiac magnetic resonance (MR) in the presence of a transcatheter heart valve (THV). Therefore, we aimed to validate and determine the most appropriate location and the accuracy of MR velocity mapping for the quantification of PAR in presence of a THV. Methods: In an ex-vivo pig heart model, we studied the accuracy and repeatability of MR velocity mapping in the presence of a THV using a 1.5T MR system. All experiments were performed twice, at 5 different flow velocities (between 16.67 and 50 ml/sec) for a native valve, and two commonly used THVs: SAPIEN XT and CoreValve. For each THV, antegrade flow (AF) as well as retrograde flow (RF), were measured in the ascending aorta (Level A), in the valve (Level B), and in the left ventricular outflow tract (Level C) and compared to the True flow. Statistical analysis was performed by using Bland-Altman analysis and linear regression model. Results: There was a high level of accuracy and precision for the MR-derived flow volumes in both THVs and native valve. As expected, for both THVs, measurements through the THV frame were unreliable with errors of 36.7% to 76.6%. For Sapien XT, AF and RF results were comparable at level A and C. For AF, a systematic underestimation of 12.9% and 17.7%, and a clinically acceptable random bias of 6.9% and 7.1%, were observed at level A and C, for RF a systematic underestimation of 13.4% and 7.5%. For CoreValve, the AF was best estimated at level C with a systematic underestimation of 20.8% and a random bias of 10.5%. RF was best estimated at level A, with a systematic underestimation of 15.8% and a random bias of only 2.5%. For all valves and flow directions, the ICCs for intratest measurements were high with small intratest differences, indicating excellent repeatability. Conclusion: In this ex-vivo study, velocity encoded MR enabled accurate, precise and repeatable quantification of PAR after implantation of two commonly used THVs, when corrected for the systematic underestimation