Automated Diagnosis and Assessment of Cardiac Structural Alteration in Hypertension Ultrasound Images

Area of Science:

• Cardiology
• Medical Imaging
• Computer-Aided Diagnosis

Background:

• Hypertension (HTN) is a significant risk factor for cardiovascular diseases, contributing to substantial mortality from heart disease and stroke.
• Prevalence of HTN is high in India and projected to increase, emphasizing the need for effective blood pressure control.
• Early detection of HTN-induced myocardial structural changes is crucial for managing cardiovascular risk.

Purpose of the Study:

• To develop and evaluate a multi-resolution analysis-based computer-aided diagnosis (CAD) system.
• To detect early left ventricular heart muscle changes associated with hypertension using ultrasound imaging.
• To assess the efficacy of LSDA and decision tree classification with contourlet and shearlet transforms for HTN detection.

Main Methods:

• Feature extraction from ultrasound imagery of the heart.
• Dimensionality reduction using locality sensitive discriminant analysis (LSDA).
• Classification using a decision tree model with contourlet and shearlet transform features, optimized to two key features.

Main Results:

• The developed CAD system effectively detected early HTN-induced left ventricular structural alterations.
• The combination of LSDA and decision tree classifier with specific transform features achieved high accuracy.
• The model demonstrated the potential for accurate diagnosis using a minimal set of features.

Conclusions:

• A multi-resolution analysis-based CAD system utilizing ultrasound imaging is effective for early detection of HTN-related cardiac changes.
• The system, employing LSDA and decision tree classification, offers a promising noninvasive tool for evaluating cardiac structural alterations in hypertensive patients.
• This standalone tool can be valuable for clinical applications in hospitals and polyclinics for managing hypertension.