Predicting Coronary Stenosis Progression Using Plaque Fatigue From IVUS-Based Thin-Slice Models: A Machine Learning...

Area of Science:

• Cardiovascular Medicine
• Biomedical Engineering
• Medical Imaging

Background:

• Coronary atherosclerosis restricts blood flow, increasing heart attack risk.
• Progression of atherosclerosis is linked to various risk factors, but fatigue is understudied.
• Understanding factors influencing stenosis progression is crucial for clinical risk assessment.

Purpose of the Study:

• To investigate the relationship between biomechanical fatigue and coronary stenosis progression.
• To assess if fatigue-related factors can predict atherosclerosis progression.
• To utilize intravascular ultrasound (IVUS) imaging and finite element models for this investigation.

Main Methods:

• Constructed IVUS-based thin-slice models from seven patients' data.
• Measured coronary biomechanics and stress/strain amplitudes as indicators of fatigue.
• Calculated change in lumen area (DLA) to quantify stenosis progression.
• Employed Random Forest (RF) method to identify predictive factors and assess classification accuracy.

Main Results:

• Significant correlations found between stenosis progression and maximum/average stress and average strain amplitudes (p < 0.05).
• RF model identified eight key factors, including fatigue, for predicting stenosis progression.
• RF model achieved 83.61% overall accuracy, 86.25% sensitivity, and 80.69% specificity in predicting progression.

Conclusions:

• Biomechanical fatigue is positively correlated with coronary stenosis progression.
• Fatigue-related factors enhance the prediction accuracy of atherosclerosis progression.
• Incorporating fatigue metrics may improve clinical risk stratification for cardiovascular events.