Texture analysis of acute myocardial infarction with CT: First experience study

Area of Science:

• Cardiovascular Imaging
• Radiology
• Medical Image Analysis

Background:

• Distinguishing healthy from infarcted myocardium is crucial for diagnosing acute myocardial infarction (MI).
• Traditional methods may lack objective, quantitative assessment of myocardial tissue characteristics.

Purpose of the Study:

• To evaluate the feasibility and accuracy of texture analysis in computed tomography (CT) for differentiating healthy from infarcted myocardium.
• To identify objective, quantitative image features for improved diagnostic capabilities.

Main Methods:

• Contrast-enhanced cardiac CT was performed on patients with acute MI and controls.
• Texture analysis was applied to left ventricular short-axis CT images reconstructed at various slice thicknesses (1mm, 2mm, 5mm).
• Features were extracted using histogram, gray-level co-occurrence matrix (GLCM), and gray-level run-length matrix (GLRLM) analyses.

Main Results:

• Texture analysis demonstrated significant differences between healthy and infarcted myocardium across various features and slice thicknesses.
• A multivariate model incorporating kurtosis and short run high gray-level emphasis (SRHGE) achieved an area under the curve (AUC) of 0.90 for MI detection.
• Optimal performance for kurtosis and SRHGE was observed at a 5mm slice thickness, predicting acute MI with 95% sensitivity.

Conclusions:

• Texture analysis is a feasible method for objectively distinguishing healthy from acutely infarcted myocardium using cardiac CT.
• Quantitative texture features, particularly kurtosis and SRHGE at 5mm slice thickness, show high accuracy for acute MI detection.
• This approach offers a promising tool for non-invasive assessment of myocardial infarction.