Recognizing artery segments on carotid ultrasonography using embedding concatenation of deep image and vision-language models
View abstract on PubMed
Summary
This summary is machine-generated.An AI model accurately classifies carotid artery segments in ultrasound images, improving stroke risk assessment. This automated approach enhances the consistency and precision of carotid artery examinations for better patient outcomes.
Area Of Science
- Medical Imaging
- Artificial Intelligence
- Cardiovascular Disease
Background
- Large artery atherosclerosis is a key predictor of ischemic strokes.
- Carotid artery ultrasound is a noninvasive, first-line diagnostic tool.
- Accurate quantification of carotid artery segments is crucial for risk assessment.
Purpose Of The Study
- To develop an automated classification model for common carotid artery (CCA), carotid bulb, internal carotid artery (ICA), and external carotid artery (ECA).
- To enhance the accuracy and consistency of carotid artery examinations using deep learning.
- To improve the quantification of carotid artery structures in B-mode ultrasound.
Main Methods
- Collected 2943 B-mode ultrasound images from 288 patients.
- Extracted features using DenseNet201, vision transformer, and echo contrastive language-image pre-training models.
- Combined features in a support vector machine classifier for anatomical structure interpretation.
Main Results
- Achieved an overall classification accuracy of 82.3% after ten-fold cross-validation.
- Demonstrated significantly improved performance compared to individual feature sets (p < 0.001).
- The model effectively interprets anatomical structures in B-mode ultrasound images.
Conclusions
- The proposed automated model enhances the accuracy and consistency of carotid artery examinations.
- This AI-driven approach aids in better prediction and prevention of ischemic strokes.
- The developed model offers a valuable tool for cardiovascular disease assessment.
Related Concept Videos
The coronary arteries, originating from the ascending aorta, bifurcate from two sinuses located within the ascending aorta. Positioned just above the aortic semilunar valve, these sinuses house essential aortic baroreceptors and chemoreceptors, crucial for maintaining cardiac function. The left coronary artery and the right coronary artery branch off from the left posterior and anterior aortic sinuses, respectively.
Encircling the heart, the coronary arteries form a ring-like structure before...
The human body's intricate network of arteries ensures that every organ system receives the necessary oxygen and nutrients for optimal function. The arterial network in the head and neck region is particularly complex, providing vital blood flow to the brain, eyes, and other critical structures. Prominent arteries in this region include the internal carotid arteries and the vertebral arteries.
The internal carotid arteries supply blood to the anterior portion of the cerebrum. They enter the...

