An Adaptive SCG-ECG Multimodal Gating Framework for Cardiac CTA

Area of Science:

• Biomedical Engineering
• Cardiovascular Imaging
• Artificial Intelligence in Medicine

Background:

• Cardiovascular disease (CVD) is a leading global cause of mortality.
• Current diagnostic methods like catheter coronary angiography (CCA) are invasive and costly.
• Cardiac computed tomography angiography (CTA) offers a less invasive alternative but is limited by motion artifacts due to poor cardiac motion capture.

Purpose of the Study:

• To develop and validate a novel multimodal approach for enhancing CTA image quality.
• To improve the prediction of cardiac quiescent periods for more accurate CTA acquisition.
• To reduce motion artifacts and enhance diagnostic capabilities for coronary artery disease (CAD).

Main Methods:

• A weighted fusion (WF) approach integrating seismocardiogram (SCG) and electrocardiogram (ECG) data was used.
• Artificial neural networks (ANNs), including a regression-based framework (r-ANN WF) and a classification-based framework (c-ANN WF), were developed.
• The proposed methods were compared against traditional ECG gating and ultrasound (US) data for accuracy and computational efficiency.

Main Results:

• The r-ANN WF approach demonstrated a 52.6% improvement in cardiac quiescence prediction accuracy compared to ECG-based methods, validated against US data.
• The average prediction time for the r-ANN WF method was 4.83 ms, indicating suitability for real-time applications.
• Reconstructed CTA images using both r-ANN WF and c-ANN WF achieved diagnostic quality comparable to US-based gating.

Conclusions:

• The proposed multimodal approach using SCG and ECG significantly enhances the prediction of cardiac quiescent periods for CTA.
• The r-ANN WF framework offers a computationally efficient and accurate method for improving CTA diagnostic quality.
• This innovative technique holds significant clinical potential for more accurate and efficient diagnosis and management of cardiovascular diseases.