Automatic phase determination for retrospectively gated cardiac CT.

Area of Science:

• Medical Imaging
• Cardiovascular Imaging
• Computed Tomography (CT)

Background:

• Advancements in CT technology and algorithms allow high-resolution 3D heart visualization.
• Limited temporal resolution in CT hinders artifact-free cardiac imaging at arbitrary phases.
• Identifying quasistationary cardiac phases is crucial for optimal image quality but challenging due to variability.

Purpose of the Study:

• To introduce a simple, efficient, and automatic image-based technique for identifying stable cardiac phases.
• To overcome the limitations of electrocardiogram (ECG) in accurately representing heart motion.
• To enable patient-specific, optimized high-resolution CT reconstructions at phases of minimal motion.

Main Methods:

• Utilized low-resolution four-dimensional (4D) CT datasets.
• Calculated object similarity between consecutive cardiac phases to derive stable phases.
• Determined patient-specific object motion for optimized reconstruction phases.

Main Results:

• The image-based technique successfully identified stable cardiac phases automatically and in a patient-specific manner.
• Demonstrated the ability to perform optimized high-resolution reconstructions during phases of minimal heart motion.
• Results validated through simulation studies and analysis of three real patient datasets.

Conclusions:

• The proposed image-based method provides a robust solution for identifying optimal cardiac phases for CT imaging.
• This technique enhances cardiac CT image quality by minimizing motion artifacts.
• It offers a patient-specific approach to improve diagnostic accuracy in cardiovascular imaging.