Cardiac Image Registration: Rotational Error Correction and Gated Stabilization for Cardiac Motion

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Cardiovascular Research

Background:

• Cardiac and respiratory motion, plus patient rotation, introduce errors in cardiac image registration.
• These dynamic movements complicate accurate alignment across different imaging modalities.
• Phantom, patient, and animal models were employed to investigate and rectify these registration inaccuracies.

Purpose of the Study:

• To identify and correct rotational errors in cardiac imaging.
• To assess and improve the accuracy of electrocardiogram (ECG)-gated image registration in patients.
• To evaluate a tracking algorithm for correcting respiratory motion artifacts.

Main Methods:

• Rotational errors were corrected using various phantom orientations.
• ECG-gated fluoroscopy and CT images were aligned in 9 patients, with accuracy evaluated during atrial fibrillation (AF) and sinus rhythm.
• A respiration tracking algorithm was developed and compared against an automated method by four observers.

Main Results:

• The a posteriori ECG gating method achieved high accuracy (97-98%) during both AF and sinus rhythm, outperforming the a priori method (38%).
• The respiration tracking algorithm demonstrated good or fair performance 96% of the time, with no significant difference between observer and automated assessments.
• Corrected target registration errors were minimal in phantom (1.75±1.03 mm) and animal models (0-0.5 mm).

Conclusions:

• Cardiac image registration errors stemming from dynamic motion and rotation are identifiable and correctable.
• Image stabilization for cardiac procedures can be effectively achieved through the integration of ECG gating and respiratory motion compensation.