Diagnostic Classification of Patients with Dilated Cardiomyopathy Using Ventricular Strain Analysis Algorithm

Area of Science:

• Cardiology
• Medical Imaging
• Machine Learning

Background:

• Dilated cardiomyopathy (DCM) is a significant cause of heart failure and sudden cardiac death.
• Cardiac MRI is crucial for diagnosing heart conditions but manual DCM parameter measurement is time-consuming.
• Accurate and efficient diagnosis of DCM is critical for timely patient management and treatment.

Purpose of the Study:

• To develop an automated system for diagnosing dilated cardiomyopathy (DCM) using cardiac MRI.
• To enable automatic segmentation of cardiac ventricles and extraction of relevant parameters.
• To improve the speed and accuracy of DCM diagnosis compared to manual methods.

Main Methods:

• Utilized parasternal short-axis cardiac MRI sequences for automatic left and right ventricle segmentation.
• Extracted key cardiac parameters during end-diastole and end-systole phases.
• Employed machine learning classifiers, including Random Forest, for DCM prediction based on extracted parameters.

Main Results:

• The proposed automated system effectively detects and diagnoses DCM.
• Random Forest classifier achieved the highest accuracy in distinguishing between normal individuals and DCM patients.
• The method demonstrates comparable results to complex techniques with minimal sample input.

Conclusions:

• The developed automated system offers an efficient and accurate approach for DCM diagnosis via cardiac MRI.
• This method has the potential to significantly aid cardiologists in diagnosing DCM.
• The system's capabilities highlight advantages in speed and precision for cardiac disease diagnosis.