Aortic Stenosis assessment with a 3-directional phase contrast magnetic resonance technique. Comparison to transthoracic echocardiography

Juliana Serafim da Silveira, Matthew Smyke, Ning Jin, Rizwan Ahmad, Lua Jafari, Debbie Scandling, Jennifer A. Dickerson, Subha V. Raman, Orlando P. SimonettiJournal of Cardiovascular Magnetic Resonance 2015 · DOI

Abstract

Traditional unidirectional phase-contrast MRI (1DPC-MRI) often underestimates blood flow velocity in Aortic Stenosis (AS) due to misalignment with eccentric jets. This study demonstrates that a 3-directional phase-contrast MRI (3DPC-MRI) technique, which captures velocity vectors in three dimensions within a single breath-hold, significantly improves diagnostic accuracy and correlation with transthoracic echocardiography (TTE).

Introduction

Aortic Stenosis (AS) is a narrowing of the aortic valve that requires precise staging for clinical management. Transthoracic Doppler-echocardiography (TTE) is the current gold standard for measuring peak blood velocity and pressure gradients. While MRI is a powerful alternative, 1DPC-MRI only measures velocity perpendicular to the imaging plane. If the plane is not perfectly aligned with the blood flow—common in patients with eccentric jets—1DPC-MRI consistently underestimates disease severity. This study investigates whether a rapid 3DPC-MRI technique can overcome these limitations.

Methods

The study involved 13 patients (median age 65) with mild to severe AS. Three diagnostic approaches were compared: TTE, 1DPC-MRI, and the experimental 3DPC-MRI. The 3DPC-MRI sequence was optimized to capture velocity in the x, y, and z directions within a single 10-14 second breath-hold.

Comparison of MRI sequence parameters for 1DPC and 3DPC techniques.

Quantitative analysis was performed using custom MATLAB software to identify peak velocities and calculate trans-valvular gradients from the acquired MRI data.

Key Results and Insights

The 3DPC-MRI technique showed superior agreement with TTE across all primary hemodynamic parameters. The Intraclass Correlation Coefficient (ICC) for peak velocity improved from 0.82 (1D) to 0.88 (3D). Most notably, the correlation for the mean trans-valvular gradient jumped from 0.42 to 0.72, indicating that 3D encoding captures the overall hemodynamic profile much more reliably.

Bland-Altman plots comparing TTE with 1DPC-MRI (A-C) and 3DPC-MRI (D-F), showing reduced bias for the 3D technique.

Bland-Altman analysis confirmed that 3DPC-MRI (plots D-F) exhibited less overall bias compared to 1DPC-MRI (plots A-C). By measuring all three velocity vectors, the technique effectively "future-proofs" the scan against human error in slice positioning, as the true magnitude of blood flow can be calculated regardless of plane tilt.

Conclusion

3DPC-MRI provides a more accurate and robust assessment of Aortic Stenosis than traditional 1D methods. Its ability to be performed within a standard breath-hold makes it a clinically viable tool, particularly for patients with complex flow patterns or poor acoustic windows for ultrasound. While this proof-of-concept study used a small cohort, the results strongly support further exploration of 3DPC-MRI as a standard diagnostic tool.

Reference: da Silveira, J. S., et al. (2015). Aortic Stenosis assessment with a 3-directional phase contrast magnetic resonance technique. Comparison to transthoracic echocardiography. Journal of Cardiovascular Magnetic Resonance, 17(Suppl 1):P381.

Summary updated

Abstract

3DPC-MRI parameters (D-F) are shown in Figure 1. Note less overall bias for the 3DPC-MRI technique.
BibTeX
@article{silveira2015aortic,
  author = {Juliana Serafim da Silveira and Matthew Smyke and Ning Jin and Rizwan Ahmad and Lua Jafari and Debbie Scandling and Jennifer A. Dickerson and Subha V. Raman and Orlando P. Simonetti},
  title = {{Aortic Stenosis assessment with a 3-directional phase contrast magnetic resonance technique. Comparison to transthoracic echocardiography}},
  journal = {Journal of Cardiovascular Magnetic Resonance},
  year = {2015},
  doi = {10.1186/1532-429x-17-s1-p381},
}
Figure