Highly accelerated phase-contrast MRI-based multi-directional flow imaging for peak velocity estimation in aortic stenosis patients.

Juliana Serafim da Silveira, Adam Rich, Yingmin Liu, Matthew Smyke, Ning Jin, Debbie Scandling, Jennifer A. Dickerson, Carlos Eduardo Rochitte, Subha V. Raman, Lee C. Potter, Rizwan Ahmad, Orlando P. SimonettiJournal of Cardiovascular Magnetic Resonance 2016 · DOI

Introduction

Aortic stenosis (AS), characterized by the narrowing of the aortic valve, is the most prevalent valvular heart disease and continues to rise in incidence. Accurate grading of AS severity is critical for clinical management, primarily relying on the measurement of peak blood flow velocity ($V_{peak}$) through the valve. Currently, Transthoracic Echocardiography (TTE) serves as the diagnostic gold standard. However, TTE is often limited by poor acoustic windows in certain patients and significant misalignment errors. If the ultrasound beam is not perfectly parallel to the blood flow—particularly in cases of eccentric jets—TTE tends to underestimate the true peak velocity.

The Limitations of Standard MRI

Cardiovascular Magnetic Resonance (CMR) has emerged as a powerful alternative, but standard unidirectional phase-contrast MRI (PC-MRI) faces similar alignment challenges. It only measures flow perpendicular to the imaging plane. While multi-directional PC-MRI can capture the full 3D velocity vector and eliminate alignment errors, the acquisition time is traditionally too long for clinical use, often requiring breath-holds that exceed patient capabilities.

Methods: Accelerating Multi-Directional Flow Imaging

To address the speed-accuracy trade-off, researchers applied the ReVEAL (Reduced-encoding Visually Enhanced ALgorithm) data processing method. ReVEAL significantly accelerates multi-directional PC-MRI by exploiting spatiotemporal sparsity—the inherent redundancy in medical image sequences over space and time. Furthermore, it leverages the mathematical relationship between encoded flow images and compensated images used for phase error correction.

Imaging Protocol

The study prospectively enrolled 14 patients (median age 68) with varying degrees of AS. Scanning was performed on a 1.5T Siemens Avanto system using a 4-point encoding scheme to capture all three velocity components ($V_x, V_y, V_z$). By employing a variable-density sampling pattern, the team achieved a net acceleration rate of 8. This allowed for the acquisition of an 8mm slice in a clinically feasible 10-second breath-hold.

The peak velocity was calculated as the magnitude of the reconstructed velocity vector: $$V_{peak} = \sqrt{V_x^2 + V_y^2 + V_z^2}$$

Key Results and Visualization

The 8x acceleration successfully maintained image integrity while drastically reducing scan time. As shown in Figure 1, the method produced high-quality magnitude, phase, and $V_{peak}$ maps.

Figure 1: Magnitude, multi-directional phase, and peak velocity ($V_{peak}$) maps obtained using the 8x accelerated ReVEAL method.

The study found a strong correlation between ReVEAL-based MRI measurements and clinical TTE reports, with an $R^2$ value of 0.75. Notably, TTE tended to slightly underestimate $V_{peak}$ compared to the MRI results, supporting the hypothesis that multi-directional MRI is less sensitive to jet eccentricity and alignment errors.

Discussion and Clinical Impact

The ability to capture full 3D velocity vectors within a single breath-hold represents a significant step toward making multi-directional flow imaging a routine clinical tool. By overcoming the "angle problem" inherent in TTE and unidirectional MRI, this approach offers a more robust method for grading stenosis, especially in patients with complex or eccentric flow jets.

Limitations

While promising, the study was limited by a small sample size (n=14) and a significant temporal gap (average 40 days) between TTE and MRI scans. Future research with larger cohorts and tighter scan windows is necessary to fully validate the diagnostic accuracy of this accelerated technique.

Conclusion

Highly accelerated multi-directional PC-MRI using the ReVEAL algorithm provides a feasible and potentially more accurate alternative to TTE for measuring peak velocity in aortic stenosis. This technology bridges the gap between high-fidelity research imaging and the practical constraints of the clinical environment.

Summary updated

Abstract

Background Aortic stenosis (AS) is the most common valvular disease, and its prevalence is on the rise. Transthoracic echocardiography (TTE) is the current gold standard for diagnosis and grading of AS. However, TTE suffers from inadequate acoustic windows, and misalignment errors. While CMR has emerged as a robust tool for numerous applications, flow analysis by unidirectional phase-contrast MRI (PC-MRI) is known to underestimate velocity if the imaging plane is not set perpendicular to flow direction. Selecting the proper orientation can be challenging as the jet direction may vary with respect to the valve orifice. Thus, multi-directional flow imaging is likely to improve the accuracy of peak velocity (Vpeak) measurements. However, multi-directional acquisition can be prohibitively long, limiting its clinical utility. The purpose of this study is to apply a recently proposed data processing method called ReVEAL [1] to significantly accelerate multi-directional PC-MRI. ReVEAL exploits spatiotemporal sparsity and leverages the relationship between encoded and compensated images to enable highly accelerated PC-MRI.
BibTeX
@article{silveira2016highly,
  author = {Juliana Serafim da Silveira and Adam Rich and Yingmin Liu and Matthew Smyke and Ning Jin and Debbie Scandling and Jennifer A. Dickerson and Carlos Eduardo Rochitte and Subha V. Raman and Lee C. Potter and Rizwan Ahmad and Orlando P. Simonetti},
  title = {{Highly accelerated phase-contrast MRI-based multi-directional flow imaging for peak velocity estimation in aortic stenosis patients.}},
  journal = {Journal of Cardiovascular Magnetic Resonance},
  year = {2016},
  doi = {10.1186/1532-429x-18-s1-w22},
}
Figure