Deep learning‐based carotid media‐adventitia and lumen‐intima boundary segmentation from three‐dimensional ultrasound images

Abstract

PurposeQuantification of carotid plaques has been shown to be important for assessing as well as monitoring the progression and regression of carotid atherosclerosis. Various metrics have been proposed and methods of measurements ranging from manual tracing to automated segmentations have also been investigated. Of those metrics, quantification of carotid plaques by measuring vessel‐wall‐volume (VWV) using the segmented media‐adventitia (MAB) and lumen‐intima (LIB) boundaries has been shown to be sensitive to temporal changes in carotid plaque burden. Thus, semi‐automatic MAB and LIB segmentation methods are required to help generate VWV measurements with high accuracy and less user interaction.MethodsIn this paper, we propose a semiautomatic segmentation method based on deep learning to segment the MAB and LIB from carotid three‐dimensional ultrasound (3DUS) images. For the MAB segmentation, we convert the segmentation problem to a pixel‐by‐pixel classification problem. A dynamic convolutional neural network (Dynamic CNN) is proposed to classify the patches generated by sliding a window along the norm line of the initial contour where the CNN model is fine‐tuned dynamically in each test task. The LIB is segmented by applying a region‐of‐interest of carotid images to a U‐Net model, which allows the network to be trained end‐to‐end for pixel‐wise classification.ResultsA total of 144 3DUS images were used in this development, and a threefold cross‐validation technique was used for evaluation of the proposed algorithm. The proposed algorithm‐generated accuracy was significantly higher than the previous methods but with less user interactions. Comparing the algorithm segmentation results with manual segmentations by an expert showed that the average Dice similarity coefficients (DSC) were 96.46 ± 2.22% and 92.84 ± 4.46% for the MAB and LIB, respectively, while only an average of 34 s (vs 1.13, 2.8 and 4.4 min in previous methods) was required to segment a 3DUS image. The interobserver experiment indicated that the DSC was 96.14 ± 1.87% between algorithm‐generated MAB contours of two observers' initialization.ConclusionsOur results showed that the proposed carotid plaque segmentation method obtains high accuracy and repeatability with less user interactions, suggesting that the method could be used in clinical practice to measure VWV and monitor the progression and regression of carotid plaques.

Deep morphology aided diagnosis network for segmentation of carotid artery vessel wall and diagnosis of carotid atherosclerosis on black‐blood vessel wall MRI

“…In addition, the CNNs can achieve high performance in medical image analysis by learning the hierarchical features from the raw image data, while a U‐shaped deep convolutional network (i.e., known as U‐Net) uses a skip connection between the downsampling and the upsampling paths and achieves significant performance in various medical image segmentation tasks. More recently, some CNNs‐based deep learning methods were proposed to segment the CCA on the ultrasound images. In particular, the encoder–decoder convolutional structure with the fusion of envelope and phase consistency data was utilized to segment the media‐outer film boundaries from the B‐mode 2D ultrasound images, while the dynamic CNN and the U‐Net were used to segment the media–adventitia and lumen–endometrial boundaries from the carotid 3D ultrasound images .…”

“…More recently, some CNNs‐based deep learning methods were proposed to segment the CCA on the ultrasound images. In particular, the encoder–decoder convolutional structure with the fusion of envelope and phase consistency data was utilized to segment the media‐outer film boundaries from the B‐mode 2D ultrasound images, while the dynamic CNN and the U‐Net were used to segment the media–adventitia and lumen–endometrial boundaries from the carotid 3D ultrasound images . However, these two methods only focus on the CCA and are unsuitable to the areas near the carotid bifurcation in ultrasound images, where the carotid bifurcation is the most common site of atherosclerosis progression.…”

“…In particular, the encoder-decoder convolutional structure with the fusion of envelope and phase consistency data was utilized to segment the media-outer film boundaries from the B-mode 2D ultrasound images, 39 while the dynamic CNN and the U-Net 38 were used to segment the media-adventitia and lumen-endometrial boundaries from the carotid 3D ultrasound images. 40 However, these two methods only focus on the CCA and are unsuitable to the areas near the carotid bifurcation in ultrasound images, where the carotid bifurcation is the most common site of atherosclerosis progression. Therefore, as far as we know, deep learning-based methods have not been studied for the automated segmentation of carotid vascular wall and the automated diagnosis of carotid atherosclerosis in MRI images.…”

Deep Learning for Carotid Plaque Segmentation using a Dilated U-Net Architecture

“…[12][13][14][15][16][17][18] Use of three-dimensional (3D) data sets provide significant improvements when compared to 2D data sets. 12,13,18 Qiu et al 13 developed a fully automatic segmentation method using CNN for 3D segmentation of the brain ventricle structure in embryonic mice. Their implementation consisted of a two-stage process where the first stage performed localization of the relevant structure within a bounding box, which was then supplied as an input to the second stage, which performed the segmentation.…”

Deep learning‐based digitization of prostate brachytherapy needles in ultrasound images

“…10 Deep learning for segmentation in ultrasound images has previously been studied by others. [11][12][13][14][15][16][17] Two recent studies using, in the field of prostate brachytherapy, deep learning has been utilized for needle digitization in prostate brachytherapy 18,19 trained the algorithm using patches instead of the whole image volume, and employed a weighted loss function between cross entropy and total variation for optimization. However, other metrics exist describing intersection over union, such as the Dice similarity coefficient.…”