A geometric model of the beating heart

Area of Science:

• Cardiovascular Imaging
• Medical Image Analysis
• Computational Anatomy

Background:

• Automated diagnostic and interventional procedures in cardiology require robust anatomical models.
• Existing models may lack comprehensive coverage of cardiac structures and anatomical variability.

Purpose of the Study:

• To present a comprehensive statistical model of the human heart.
• To provide a priori information for automated cardiac procedures.
• To incorporate anatomical landmarks and coronary arteries into the model.

Main Methods:

• Adapted an end-diastolic phase model to 27 clinical multi-slice computed tomography (CT) images.
• Calculated a mean cardiac motion model from 11 multi-phase CT image sets.
• Performed experiments to validate model-based predictions on unseen cardiac images.

Main Results:

• Determined cardiac chambers and vessels with an average accuracy of 1.1 mm using anatomical landmarks.
• Estimated coronary artery centerlines with an average accuracy of 5.2 mm after surface adaptation.
• Utilized the mean motion model to estimate cardiac phase from unknown CT images.

Conclusions:

• The presented mean shape model of the human heart enhances automated image analysis.
• It provides essential a priori anatomical information for robust and rapid cardiac procedures.
• The model improves the performance of automated diagnostic and interventional methods.