A novel approach in the assessment of paravalvular regurgitation after transcatheter aortic valve replacement using cardiac magnetic resonance imaging

Abstract

Listen

Translate article

Abstract Funding Acknowledgements Type of funding sources: None. Background Paravalvular regurgitation (PVR) is a common complication after transcatheter aortic valve replacement (TAVR) that poses an increased risk of rehospitalization for heart failure and mortality. Two-dimensional (2D) cardiac magnetic resonance (CMR) quantitative flow assessment is the gold standard for evaluation of PVR after TAVR. Two novel techniques in flow assessment are 2D multi-velocity encoding (Venc) and four-dimensional (4D) flow. The 2D multi-venc sequence provides the advantage of integrating three different Venc values in a rapid, single breath-hold acquisition, specifically providing accurate measurements for both high and low velocities. 4D-flow allows a free-breathing acquisition with unlimited plane selection, and provides a comprehensive visualization of the blood flow. Although the use of these novel techniques is expanding, these have not yet been assessed for their correlation with standard 2D-flow. Purpose To compare the standard 2D-flow acquisition with both 2D multi-venc and 4D-flow acquisitions for the assessment of PVR using CMR-derived regurgitant fraction (CMR-RF). Methods In this substudy of the APPOSE trial (NCT04281771), 19 patients underwent CMR one month after TAVR using 2D multi-venc and 4D-flow, in addition to the standard 2D-flow acquisition. Medis Suite MR (Medis Medical Imaging, Leiden, the Netherlands) was used for analysis of the flow data. Scatterplots and Bland-Altman plots were used to assess correlation and visualize agreement between the techniques. Results Mean age was 79.8 ± 4.9 years, and 42.9% of the patients were men. Mean CMR-RF was 11.7 ± 10.0% using the standard 2D-flow sequence, 10.6 ± 7.0% using the 2D multi-venc sequence, and 9.6 ± 7.3% using the 4D-flow acquisition. Both the CMR-RF assessed with 2D multi-venc and 4D-flow showed a high correlation with the CMR-RF assessed with standard 2D-flow (r = 0.88, p < 0.001 and r = 0.74, p < 0.001, respectively). Bland-Altman plots revealed no significant bias between the mean values of CMR-RF (2D multi-venc: 1.31%; 4D-flow: 0.27%). Conclusion Both the 2D multi-venc and 4D-flow sequence produce accurate quantification of PVR after TAVR. Given the advantages of these approaches in terms of acquisition speed and plane selection, use of these techniques should be encouraged.