Detection of left ventricular motion abnormality via information measures and bayesian filtering

Area of Science:

• Cardiology
• Biomedical Engineering
• Information Theory

Background:

• Heart wall motion analysis from functional images is challenging due to noise and segmentation inaccuracies.
• Accurate detection of left ventricular (LV) abnormalities requires incorporating prior knowledge and robust analytical methods.
• Distinguishing normal from abnormal heart motion is difficult due to subtle statistical similarities.

Purpose of the Study:

• To develop a novel information theoretic measure for detecting heart wall motion abnormalities.
• To enhance the accuracy of LV motion analysis using Shannon's differential entropy (SDE).
• To investigate alternative information theoretic criteria, including Rényi entropy and Fisher information.

Main Methods:

• Utilized the Kalman filter for estimating LV cavity points from noisy and incomplete functional imaging data.
• Developed a global measure based on Shannon's differential entropy (SDE) for abnormality detection.
• Constructed distributions of normalized radial distance estimates of the LV cavity for quantitative analysis.

Main Results:

• The proposed SDE criterion demonstrated significant improvement in detecting LV abnormalities compared to traditional features.
• Experimental analysis on MRI data from 30 subjects validated the effectiveness of the SDE method.
• The SDE approach outperformed mean radial displacement and mean radial velocity in LV cavity analysis.

Conclusions:

• Shannon's differential entropy offers a powerful tool for quantitative heart wall motion analysis and abnormality detection.
• The developed information theoretic measures provide a more accurate and robust approach to diagnosing cardiac motion disorders.
• This study highlights the potential of information theory in advancing cardiovascular imaging analysis.