Automatic cardiac MRI segmentation using a biventricular deformable medial model.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Computational Anatomy

Background:

• Accurate segmentation of the myocardium (heart muscle) in cardiac MRI is crucial for diagnosing and monitoring cardiovascular diseases.
• Existing segmentation methods often struggle with variations in heart shape, thickness, and image quality.

Purpose of the Study:

• To develop and evaluate a novel automatic segmentation method for the myocardium in short-axis cardiac MRI.
• To incorporate an explicit representation of myocardial thickness and a Markov prior to constrain segmentation.

Main Methods:

• Utilized deformable medial models with an explicit thickness representation.
• Adapted Active Shape Modeling techniques, including global PCA shape priors, statistical appearance models, and local search.
• Constrained segmentation using a Markov prior on myocardial thickness.

Main Results:

• Achieved an average boundary displacement error of under 1.4 mm for both left and right ventricles.
• Demonstrated favorable comparison with existing published segmentation methods.
• Validated performance on a heterogeneous adult MRI dataset.

Conclusions:

• The proposed deformable medial model approach offers a robust and accurate method for automatic myocardial segmentation in cardiac MRI.
• Explicitly modeling myocardial thickness improves segmentation accuracy and consistency.
• This technique shows significant potential for clinical application in cardiovascular imaging analysis.