VWV-SSL: Carotid vessel-wall-volume segmentation via sequence structural similarity and augmentation
Insights
A new self-supervised learning (SSL) method, VWV-SSL, improves 3D carotid ultrasound segmentation for vessel wall volume measurement. This approach requires fewer labeled images, making it efficient for monitoring atherosclerosis progression.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Cardiovascular Research
Background:
- Vessel wall volume (VWV) quantifies carotid atherosclerosis progression using 3D ultrasound.
- Accurate VWV measurement necessitates segmenting media-adventitia (MAB) and lumen intima boundaries (LIB).
- Deep learning requires extensive annotated data, posing a challenge for 3D ultrasound segmentation.
Purpose of the Study:
- To develop a novel self-supervised learning (SSL) algorithm, VWV-SSL, for 3D carotid ultrasound (3DUS) image segmentation.
- To improve VWV measurement accuracy by leveraging sequence structural similarity and feature consistency in 3DUS images.
- To reduce the reliance on large annotated datasets for training deep learning models in carotid atherosclerosis assessment.
Main Methods:
- Proposed VWV-SSL algorithm utilizing sequence structural similarity and strong-weak augmented feature consistency for self-supervised training.
- Applied VWV-SSL to the 3D U-Net architecture for segmenting MAB and LIB in 3DUS images.
- Evaluated performance on 1158 3D US datasets from 250 subjects, comparing with baseline and state-of-the-art SSL methods using limited labeled data.
Main Results:
- VWV-SSL demonstrated significant improvements in segmentation performance compared to baseline networks when trained on small labeled datasets (15, 45, 75 subjects).
- The proposed method outperformed existing state-of-the-art SSL algorithms in segmentation accuracy.
- VWV-SSL effectively enhanced the feature learning capabilities of the 3D U-Net for vessel segmentation.
Conclusions:
- VWV-SSL offers a promising solution for accurate 3D carotid ultrasound segmentation with reduced annotation requirements.
- The method facilitates improved performance of 3D U-Net models trained on limited labeled data for VWV measurement.
- VWV-SSL has potential for clinical application in monitoring carotid atherosclerosis progression and regression.
Abstract:
Vessel wall volume (VWV) is a critical three dimensional ultrasound metric used to assess the progression and regression of carotid atherosclerosis. Ac curate measurement of VWV requires the segmentation of the media-adventitia boundary (MAB) and the lumen intima boundary (LIB) of the carotid arteries. Although deep learning methods can automatically segment the MAB and LIB and quantify VWV, they rely heavily on a large dataset with annotated images for training, which is time consuming and labor-intensive. Self-supervised learning (SSL) provides a possible solution to this challenge. However, existing SSL methods do not consider the similarities in the image sequences of 3D ultrasound. This paper proposes a novel SSL algorithm, named VWV-SSL, for 3D carotid ultrasound (3DUS) image segmentation to generate VWV measurement. VWV-SSL utilizes the sequence structural similarity and strong-weak augmented feature consistency of carotid ultrasound images to conduct the self-supervised task, which enables the networks to better learn the feature presentations of the vessel in the self-supervised task training. We applied VWV-SSL on the widely used 3D U-Net and evaluated it on 1158 3D US (579 of the common carotid artery and 579 of the bifurcation) from250subjects.Comparedtobaselinenetworks,our SSL method showed a significant improvement in segmentation performance when trained on a small number of labeled images (n = 15, 45 and 75 subjects). Moreover, the performance of VWV-SSL was superior to that of state-of-art SSL algorithms. These results indicate that our method can improve the performance of 3D U-Net when trained on a small number of labeled images, suggesting that VWV SSL could be applied in clinical practice to monitor the progression of atherosclerosis.
Related Concept Videos
01:10Mechanism of Angiogenesis
01:13Structure of Cardiac Muscles
Compared to skeletal muscles, cardiac muscle cells are small and mostly have a single nucleus. Additionally, they are usually...
01:16Heart Valves
The AV valves prevent the backflow of blood from the ventricles to the atria during ventricular contraction. These valves function with the assistance of the chordae tendineae and papillary muscles. When the ventricles are relaxed, the chordae tendineae are slack, allowing blood to flow from the atria into the...
01:27Regulation of Stroke Volume
Preload refers to the degree of stretch on the heart before it contracts. It's analogous to the stretching of a rubber band; the more it's stretched, the more forcefully it snaps back. This concept is encapsulated in the Frank-Starling law of the...
01:15Structure of Blood Vessels
01:20Anatomy of Blood Vessels
Arteries
Arteries circulate oxygenated blood from the heart, except the pulmonary artery, which transports deoxygenated blood to the lungs. Large arteries, such as the aorta, have...