A new method for objective evaluation of cardiac parametric images

Area of Science:

• Cardiology
• Medical Imaging Analysis
• Biomedical Engineering

Background:

• Assessing regional heart wall motion is crucial for diagnosing cardiac conditions.
• Current methods for analyzing cardiac parametric images can be subjective or limited in scope.
• There is a need for objective and automated tools to quantify cardiac function.

Purpose of the Study:

• To introduce a new method for analyzing cardiac parametric images.
• To represent local heart motion in a polar coordinate plane for integrated analysis.
• To develop an automated classification system for regional wall motion abnormalities.

Main Methods:

• Utilized first harmonic Fourier filtering to generate maximum volume, local ejection fraction, and phase images.
• Developed a complex ejection fraction parametric image represented by a 2D distribution in the complex plane.
• Defined four indexes (hypokinesia, asynergy, mean phase shift, mean ejection fraction) by segmenting the complex plane.
• Compared automated classification results with visual classification by three observers in 219 studies.

Main Results:

• The new method integrates information from three parametric images using a polar coordinate representation.
• Four defined indexes quantify the degree and extent of regional wall motion abnormalities.
• Automated classification of cardiac studies (normal, hypokinetic, akinetic, dyskinetic) showed good agreement with visual assessment via ROC curves.

Conclusions:

• The proposed method offers an efficient and automated approach to classifying cardiac studies.
• The polar coordinate representation and derived indexes provide a robust analysis of local heart motion.
• This technique has the potential to improve the objectivity and consistency of cardiac image interpretation.