Cardiac Image Reconstruction via Nonlinear Motion Correction Based on Partial Angle Reconstructed Images

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Radiology

Background:

• X-ray Computed Tomography (CT) enables fast imaging, but cardiac imaging is challenged by motion artifacts due to insufficient gantry rotation speed relative to heart motion.
• Motion Estimation (ME) and Motion Compensation (MC) are standard techniques to mitigate motion artifacts in cardiac CT.
• Developing advanced ME/MC algorithms is crucial for acquiring high-quality, artifact-free cardiac images.

Purpose of the Study:

• To propose a novel ME/MC algorithm for estimating a nonlinear heart motion model from sinograms with less than 360° rotation.
• To refine motion vector field (MVF) estimation for improved accuracy in nonlinear heart motion modeling.
• To achieve time-resolved cardiac imaging with reduced motion artifacts.

Main Methods:

• An initial 4-D MVF is estimated using conjugate partial angle reconstructed images, assuming nonrigid but linear heart motion.
• The MVF is refined to a nonlinear model by maximizing the information potential of a motion-compensated image.
• Motion compensation is integrated into the image reconstruction process using the determined nonlinear MVF.

Main Results:

• The proposed ME/MC algorithm successfully estimated a nonlinear heart motion model.
• Significant reduction in motion artifacts was observed across numerical phantoms, physical cardiac phantoms, and animal datasets.
• The algorithm demonstrated a noticeable improvement in overall image quality for cardiac imaging.

Conclusions:

• The developed ME/MC algorithm effectively reduces motion artifacts in cardiac CT imaging.
• This approach enables the acquisition of clearer, time-resolved heart images.
• The algorithm shows promise for enhancing diagnostic accuracy in cardiac CT applications.