Figure 3 - uploaded by Aaron Fenster
Content may be subject to copyright.
3 A 3-D B-mode ultrasound image of carotid arteries obtained with the free-hand scanning approach using a magnetic position and orientation measurement. The 3-D image has been " sliced " to reveal a plaque at the entrance of the internal carotid artery. The plaque is heterogenous, with calcified regions casting a shadow.
Source publication
Determining the severity of carotid atherosclerotic stenosis has been an important step in establishing patient management pathways and identifying patients who can benefit from carotid endarterectomy versus those who should be treated using life- style and pharmaceutical interventions. Recently a number of research groups have developed phenotypes...
Citations
... 42 A total of 224 3DUS volumes from 56 subjects acquired at baseline and a follow-up session for both carotid arteries were available for training, validating, and testing the proposed algorithm. Highresolution 3DUS images were obtained by translating an ultrasound transducer (L12-5, Philips, Bothel, WA, USA) mounted on a mechanical assembly shown in figure 3.1 in Fenster et al. 43 along the neck of the subjects for approximately 4 cm. The 2D ultrasound frames from the ultrasound machine (ATL HDI 5000, Philips, Bothel, WA, USA) were captured by a frame grabber every 0.2 mm and reconstructed into a 3D image. ...
Purpose
Vessel wall volume (VWV) and localized vessel‐wall‐plus‐plaque thickness (VWT) measured from three‐dimensional (3D) ultrasound (US) carotid images are sensitive to anti‐atherosclerotic effects of medical/dietary treatments. VWV and VWT measurements require the lumen‐intima (LIB) and media‐adventitia boundaries (MAB) at the common and internal carotid arteries (CCA and ICA). However, most existing segmentation techniques were capable of segmenting the CCA only. An approach capable of segmenting the MAB and LIB from the CCA and ICA was required to accelerate VWV and VWT quantification.
Methods
Segmentation for CCA and ICA was performed independently using the proposed two‐channel U‐Net, which was driven by a novel loss function known as the adaptive triple Dice loss (ADTL) function. The training set was augmented by interpolating manual segmentation along the longitudinal direction, thereby taking continuity of the artery into account. A test‐time augmentation (TTA) approach was applied, in which segmentation was performed three times based on the input axial images and its flipped versions; the final segmentation was generated by pixel‐wise majority voting.
Results
Experiments involving 224 3DUS volumes produce a Dice similarity coefficient (DSC) of 95.1% ± 4.1% and 91.6% ± 6.6% for the MAB and LIB, in the CCA, respectively, and 94.2% ± 3.3% and 89.0% ± 8.1% for the MAB and LIB, in the ICA, respectively. TTA and ATDL independently contributed to a statistically significant improvement to all boundaries except the LIB in ICA.
Conclusions
The proposed two‐channel U‐Net with ADTL and TTA can segment the CCA and ICA accurately and efficiently from the 3DUS volume. Our approach has the potential to accelerate the transition of 3DUS measurements of carotid atherosclerosis to clinical research.
... Recently, research in this area has instead focused more on 3D ultrasound imaging of the plaque. In fact, 3D carotid artery ultrasound imaging helps the visualization, measurement (volume) and a more comprehensive characterization of the plaque [180]. Moreover, it also provides the ability to monitor both plaque progression and regression in addition to identifying vulnerable plaques [181][182][183]. ...
B-mode ultrasound imaging is used extensively in medicine. Hence, there is a need to have efficient segmentation tools to aid in computer-aided diagnosis, image-guided interventions, and therapy. This paper presents a comprehensive review on automated localization and segmentation techniques for B-mode ultrasound images. The paper first describes the general characteristics of B-mode ultrasound images. Then insight on the localization and segmentation of tissues is provided, both in the case in which the organ/tissue localization provides the final segmentation and in the case in which a two-step segmentation process is needed, due to the desired boundaries being too fine to locate from within the entire ultrasound frame. Subsequenly, examples of some main techniques found in literature are shown, including but not limited to shape priors, superpixel and classification, local pixel statistics, active contours, edge-tracking, dynamic programming, and data mining. Ten selected applications (abdomen/kidney, breast, cardiology, thyroid, liver, vascular, musculoskeletal, obstetrics, gynecology, prostate) are then investigated in depth, and the performances of a few specific applications are compared. In conclusion, future perspectives for B-mode based segmentation, such as the integration of RF information, the employment of higher frequency probes when possible, the focus on completely automatic algorithms, and the increase in available data are discussed.
... Because thrombosis generation and subsequent cerebral emboli have a high risk leading to death, carotid atherosclerosis is becoming a significant issue for researches recently [2]. Measurement accuracy and geometric understanding of common carotid artery (CCA) play an important role in carotid atherosclerosis assessment and management [3], which requires precise segmentation. Ultrasound (US) has been employed as a routine examination for inexpensive and noninvasive clinical diagnosis of atherosclerosis (the hardening of the arteries) [4,5]. ...
Carotid atherosclerosis is a major reason of stroke, a leading cause of death and disability. In this paper, a segmentation method based on Active Shape Model (ASM) is developed and evaluated to outline common carotid artery (CCA) for carotid atherosclerosis computer-aided evaluation and diagnosis. The proposed method is used to segment both media-adventitia-boundary (MAB) and lumen-intima-boundary (LIB) on transverse views slices from three-dimensional ultrasound (3D US) images. The data set consists of sixty-eight, 17 × 2 × 2, 3D US volume data acquired from the left and right carotid arteries of seventeen patients (eight treated with 80 mg atorvastatin and nine with placebo), who had carotid stenosis of 60% or more, at baseline and after three months of treatment. Manually outlined boundaries by expert are adopted as the ground truth for evaluation. For the MAB and LIB segmentations, respectively, the algorithm yielded Dice Similarity Coefficient (DSC) of 94.4% ± 3.2% and 92.8% ± 3.3%, mean absolute distances (MAD) of 0.26 ± 0.18 mm and 0.33 ± 0.21 mm, and maximum absolute distances (MAXD) of 0.75 ± 0.46 mm and 0.84 ± 0.39 mm. It took 4.3 ± 0.5 mins to segment single 3D US images, while it took 11.7 ± 1.2 mins for manual segmentation. The method would promote the translation of carotid 3D US to clinical care for the monitoring of the atherosclerotic disease progression and regression.
... Whereas up to now the detection of the severity of carotid atherosclerotic stenosis has been considered crucial for establishing patient management pathways and selecting patients who might benefit from carotid endarterectomy instead of drug intervention and lifestyle adaption only, other phenotypes than carotid artery stenosis developed by usage of noninvasive imaging have evolved recently for recognizing patients at high risk for ischemic stroke [1] . Monitoring the progression or regression of carotid plaques with three-dimensional ultrasound imaging by identifying vulnerable and high-risk plaques prone to rupture has emerged as an important research tool for antiatherosclerotic treatment and patient management [2] . ...
... Conventional ultrasound measurements, such as the 1-D measurement of the intima-media thickness and the 2-D measurement of plaque surface, have proven useful for monitoring carotid atherosclerotic burden [3] . Conventional 2-D US has also been applied for relating plaque morphology and component parts with the risk of stroke [1] , although the results have not always been consistent. While some studies have reported a high precision for visualizing components such as intraplaque hemorrhage [4,5] , others have found only moderate sensitivity and specificity [6] . ...
... Despite its usefulness for noninvasive plaque imaging of the carotid arteries, risk stratification and genetic research [10,11] , evaluation of response to antiatherosclerotic and antihypertensive therapy [12,14] and evaluation of the influence of suspected risk factors [15,16] , the modality of three-dimensional ultrasound has not yet been established as a routine diagnostic tool on ultrasound machines. One possible explanation for this may be the fact that images of the carotid arteries comprise a scanning length of at least 4 cm, so that real-time 3-D systems cannot be used satisfactorily [1] . Besides these real-time 3-D probes, various other 3-D ultrasound systems have been presented recently. ...
Background and Purpose: 3D ultrasonic measurement of carotid atherosclerotic lesions has emerged as an important tool for research and patient management. We sought to evaluate the accuracy and the reliability of a new easy-to-use magnetically tracked freehand 3D ultrasonic device with quick reconstruction time, allowing for anatomical orientation “at a glance” for calibration and distance measurements, with unique usability as a combined tool for image acquisition, registration and measurement within 3-D ultrasonic (US) volumes, which is furthermore connectable to every conventional US-machine. To enable a complete evaluation of the whole system, three different key points had to be addressed: the reproducibility of the calibration procedure, the absolute accuracy of the whole system and a direct comparison to CT and MRI imaging modalities.
Materials and Methods: For validating our calibration method, a set of 6 calibrations was performed; each consisting of 6 records of a pyramid phantom taken from different positions and angles. To evaluate the accuracy of the whole 3D-ultrasound system, the point reconstruction accuracy and the distance accuracy were determined in a point phantom made out of a single metal wire vertically attached to the bottom of a plastic tub filled with water. For distance measurements, a precisely manufactured plastic tube phantom was scanned and the length between fixed landmarks on the tube was measured. In a final step, 3D US records acquired with Curefab CS were compared to CT and MRI scans; for this purpose all ultrasonic data was manually registered to the CT/MRI data.
Results: Concerning calibration precision the tested Curefab CS system performs state of the art compared to reviews of recent freehand 3D-Ultrasound calibration methods. The point reconstruction measure for evaluation of system accuracy retrieved a mean point accuracy of 1.52 mm in contrast to values ranging from 1.67 to 3.63 mm. Mean total error of distance measurements was 0.9% with standard deviation 0.56% in our study, compared with values reaching from about 1% up to 2.3% in other studies on this subject. All quantitative measurement results are listed in a summarized form in Table 1. Besides quantitative evaluation, 3D-ultrasound records acquired with the Curefab CS system were also compared to CT and MRI scans of patients (see Figure 3 and Figure 4). The alignment of both image modalities showed promising results for future development of diagnostic tools using all image data.
Discussion and Conclusion: Our study demonstrates that 3D measurements with Curefab CS are feasible with satisfactory reliability and accuracy. From the results gathered in our study we conclude that 3D-imaging with Curefab CS might start off the possibility of accurate visualization, volume measurement, carotid plaque characterization and identification of vulnerable plaques in the very near future.
