Similarity-driven multi-dimensional binning algorithm (SIMBA) for free-running motion-suppressed whole-heart MRA.

Area of Science:

• Cardiovascular Imaging
• Medical Physics
• Biomedical Engineering

Background:

• Whole-heart MRA typically requires complex planning or computationally intensive reconstructions to handle cardiac and respiratory motion.
• Existing free-running MRA techniques often rely on gating or complex self-gating methods, limiting speed and applicability.

Purpose of the Study:

• To develop and evaluate a fast, data-driven reconstruction algorithm for whole-heart MRA compatible with ungated free-running sequences.
• To assess the performance of the proposed similarity-driven multi-dimensional binning algorithm (SIMBA) against conventional methods.

Main Methods:

• Proposed SIMBA, a clustering algorithm for k-space data to identify motion-consistent subsets for MRA reconstruction.
• Reconstructed free-running 3D radial datasets from healthy volunteers and pediatric cardiac patients using SIMBA, standard regridding ('All Data'), and a free-running framework (FRF).
• Compared image quality based on blood-myocardium sharpness, contrast ratio, and visualization of coronary artery ostia.

Main Results:

• SIMBA and FRF reconstructions showed significantly higher blood-myocardium sharpness than 'All Data' (P < .05).
• SIMBA provided superior blood-myocardium interface sharpness in volunteers (P < .001) and higher contrast ratio than both 'All Data' and FRF (P < .05).
• Visualization of coronary artery ostia was significantly improved with SIMBA and FRF compared to 'All Data' (P < .001).

Conclusions:

• SIMBA enables fast, data-driven reconstruction of whole-heart MRA with improved image quality.
• The technique offers image quality comparable to more time-consuming methods like FRF, while being simpler and faster.
• SIMBA presents a promising advancement for efficient and high-quality whole-heart MRA acquisition.