Cardiac Anomaly Detection from Cine MRI Images using Physiological Features and Random Forest Classifier.

Area of Science:

• Medical Imaging
• Cardiovascular Diseases
• Artificial Intelligence

Background:

• Current cardiovascular disease (CVD) diagnosis using cine MRI faces limitations in accuracy for borderline cases.
• Manual visualization of 3D/4D MR data is time-consuming and requires expert interpretation.

Purpose of the Study:

• To develop an end-to-end automated computer-aided diagnosis (CAD) system for cardiovascular diseases.
• To improve the accuracy and efficiency of cardiac diagnosis using cine MRI.

Main Methods:

• Automated segmentation of critical heart substructures from cine MRI data.
• Calculation of novel domain-specific physiological features from segmented regions.
• Classification of cardiac anomalies using a random forest classifier.

Main Results:

• Achieved very high accuracy on the Automated Cardiac Diagnosis challenge (ACDC) dataset.
• Demonstrated generalizability with over 90% accuracy on the multi-center M&Ms-2 dataset.

Conclusions:

• The developed end-to-end automated CAD system offers a significant improvement over conventional methods for diagnosing CVD.
• The system shows strong potential for accurate, efficient, and generalizable cardiac diagnosis across diverse datasets.