Real-time vessel segmentation and tracking for ultrasound imaging applications

On the feasibility of ultrasound Doppler-based personalized hemodynamic modeling of the abdominal aorta

BACKGROUND

Personalized modeling is a promising tool to improve abdominal aortic aneurysm (AAA) rupture risk assessment. Computed tomography (CT) and quantitative flow (Q-flow) magnetic resonance imaging (MRI) are widely regarded as the gold standard for acquiring patient-specific geometry and velocity profiles, respectively. However, their frequent utilization is hindered by various drawbacks. Ultrasound is used extensively in current clinical practice and offers a safe, rapid and cost-effective method to acquire patient-specific geometries and velocity profiles. This study aims to extract and validate patient-specific velocity profiles from Doppler ultrasound and to examine the impact of the velocity profiles on computed hemodynamics.

METHODS

Pulsed-wave Doppler (PWD) and color Doppler (CD) data were successfully obtained for six volunteers and seven patients and employed to extract the flow pulse and velocity profile over the cross-section, respectively. The US flow pulses and velocity profiles as well as generic Womersley profiles were compared to the MRI velocities and flows. Additionally, CFD simulations were performed to examine the combined impact of the velocity profile and flow pulse.

RESULTS

Large discrepancies were found between the US and MRI velocity profiles over the cross-sections, with differences for US in the same range as for the Womersley profile. Differences in flow pulses revealed that US generally performs best in terms of maximum flow, forward flow and ratios between forward and backward flow, whereas it often overestimates the backward flow. Both spatial patterns and magnitude of the computed hemodynamics were considerably affected by the prescribed velocity boundary conditions. Larger errors and smaller differences between the US and generic CFD cases were observed for patients compared to volunteers.

CONCLUSION

These results show that it is feasible to acquire the patient-specific flow pulse from PWD data, provided that the PWD acquisition could be performed proximal to the aneurysm region, and resulted in a triphasic flow pattern. However, obtaining the patient-specific velocity profile over the cross-section using CD data is not reliable. For the volunteers, utilizing the US flow profile instead of the generic flow profile generally resulted in improved performance, whereas this was the case in more than half of the cases for the patients.

3D-Ultrasound Based Mechanical and Geometrical Analysis of Abdominal Aortic Aneurysms and Relationship to Growth

The heterogeneity of progression of abdominal aortic aneurysms (AAAs) is not well understood. This study investigates which geometrical and mechanical factors, determined using time-resolved 3D ultrasound (3D + t US), correlate with increased growth of the aneurysm. The AAA diameter, volume, wall curvature, distensibility, and compliance in the maximal diameter region were determined automatically from 3D + t echograms of 167 patients. Due to limitations in the field-of-view and visibility of aortic pulsation, measurements of the volume, compliance of a 60 mm long region and the distensibility were possible for 78, 67, and 122 patients, respectively. Validation of the geometrical parameters with CT showed high similarity, with a median similarity index of 0.92 and root-mean-square error (RMSE) of diameters of 3.5 mm. Investigation of Spearman correlation between parameters showed that the elasticity of the aneurysms decreases slightly with diameter (p = 0.034) and decreases significantly with mean arterial pressure (p < 0.0001). The growth of a AAA is significantly related to its diameter, volume, compliance, and surface curvature (p < 0.002). Investigation of a linear growth model showed that compliance is the best predictor for upcoming AAA growth (RMSE 1.70 mm/year). To conclude, mechanical and geometrical parameters of the maximally dilated region of AAAs can automatically and accurately be determined from 3D + t echograms. With this, a prediction can be made about the upcoming AAA growth. This is a step towards more patient-specific characterization of AAAs, leading to better predictability of the progression of the disease and, eventually, improved clinical decision making about the treatment of AAAs.

Processing Ultrasound Scans of the Inferior Vena Cava: Techniques and Applications

The inferior vena cava (IVC) is the largest vein in the body. It returns deoxygenated blood to the heart from the tissues placed under the diaphragm. The size and dynamics of the IVC depend on the blood volume and right atrial pressure, which are important indicators of a patient's hydration and reflect possible pathological conditions. Ultrasound (US) assessment of the IVC is a promising technique for evaluating these conditions, because it is fast, non-invasive, inexpensive, and without side effects. However, the standard M-mode approach for measuring IVC diameter is prone to errors due to the vein movements during respiration. B-mode US produces two-dimensional images that better capture the IVC shape and size. In this review, we discuss the pros and cons of current IVC segmentation techniques for B-mode longitudinal and transverse views. We also explored several scenarios where automated IVC segmentation could improve medical diagnosis and prognosis.

Automatic Segmentation and Mechanical Characterisation of the Intraluminal Thrombus and Arterial Wall of Abdominal Aortic Aneurysms Using Time Resolved 3D Ultrasound Images

OBJECTIVE

This study proposed a method for semi-automatic segmentation of abdominal aortic aneurysms (AAAs) and their intraluminal thrombus (ILT), based on time resolved 3D ultrasound (US), and validated results with computed tomography (CT). Mechanical properties of both wall and ILT were determined, and possible correlations with ILT size and blood pressure were investigated.

METHODS

A semi-automatic segmentation algorithm was developed combining a star-Kalman approach with a 3D snake algorithm. The segmented geometries of both lumen and inner vessel wall were validated with both manual US based segmentations and CT based segmentations. Finally, the lumen and vessel wall distensibility and ILT compressibility were estimated and correlated with ILT size and blood pressure.

RESULTS

For the vessel wall and lumen, the median Similarity Index (SI) was 92% (IQR 90, 94%) and 83% (IQR 75, 87%), respectively. The distensibility of the vessel wall could be determined in 37 of 40 cases and had a median value of 0.28 10-5 Pa-1 (IQR 0.18, 0.51 ×10-5). The median systolic to diastolic volume change of the ILT was determined successfully in 21 of 40 patients, and was -0.57% (IQR -1.1, 1.2%). The vessel and lumen distensibility showed a strong correlation with the systolic pressure (p < .010), rather than with the diastolic pressure. Lumen distensibility was strongly correlated with ILT thickness (p = .023). The performance of the semi-automatic segmentation algorithm was shown to be as good as the manual segmentations and highly dependent on the visibility of the ILT (limited contrast in seven patients and clutter in nine patients).

CONCLUSION

This study has shown promising results for mechanical characterisation of the vessel, and ILT, including a correlation between distensibility, ILT size, and blood pressure. For future work, the inclusion rate needs to be increased by improving the image contrast with novel US techniques.

Local and global distensibility assessment of abdominal aortic aneurysms in vivo from probe tracked 2D ultrasound images

Rupture risk estimation of abdominal aortic aneurysm (AAA) patients is currently based on the maximum diameter of the AAA. Mechanical properties that characterize the mechanical state of the vessel may serve as a better rupture risk predictor. Non-electrocardiogram-gated (non-ECG-gated) freehand 2D ultrasound imaging is a fast approach from which a reconstructed volumetric image of the aorta can be obtained. From this 3D image, the geometry, volume, and maximum diameter can be obtained. The distortion caused by the pulsatility of the vessel during the acquisition is usually neglected, while it could provide additional quantitative parameters of the vessel wall. In this study, a framework was established to semi-automatically segment probe tracked images of healthy aortas (N = 10) and AAAs (N = 16), after which patient-specific geometries of the vessel at end diastole (ED), end systole (ES), and at the mean arterial pressure (MAP) state were automatically assessed using heart frequency detection and envelope detection. After registration AAA geometries were compared to the gold standard computed tomography (CT). Local mechanical properties, i.e., compliance, distensibility and circumferential strain, were computed from the assessed ED and ES geometries for healthy aortas and AAAs, and by using measured brachial pulse pressure values. Globally, volume, compliance, and distensibility were computed. Geometries were in good agreement with CT geometries, with a median similarity index and interquartile range of 0.91 [0.90-0.92] and mean Hausdorff distance and interquartile range of 4.7 [3.9-5.6] mm. As expected, distensibility (Healthy aortas: 80 ± 15·10^-3 kPa^-1; AAAs: 29 ± 9.6·10^-3 kPa^-1) and circumferential strain (Healthy aortas: 0.25 ± 0.03; AAAs: 0.15 ± 0.03) were larger in healthy vessels compared to AAAs. Circumferential strain values were in accordance with literature. Global healthy aorta distensibility was significantly different from AAAs, as was demonstrated with a Wilcoxon test (p-value = 2·10^-5). Improved image contrast and lateral resolution could help to further improve segmentation to improve mechanical characterization. The presented work has demonstrated how besides accurate geometrical assessment freehand 2D ultrasound imaging is a promising tool for additional mechanical property characterization of AAAs.

Learning-based initialization for correntropy-based level sets to segment atherosclerotic plaque in ultrasound images

Carotid artery atherosclerosis is a significant cause of stroke. Ultrasound imaging has been widely used in the diagnosis of atherosclerosis. Therefore, segmenting the atherosclerotic carotid plaque in an ultrasound image is an important task. Accurate plaque segmentation is helpful for the measurement of carotid plaque burden. This study proposes an automatic method for atherosclerotic plaque segmentation by using correntropy-based level sets (CLS) with learning-based initialization. We introduce the CLS model, containing the point-based local bias-field corrected image fitting method and correntropy-based distance measurement, to overcome the limitations of the ultrasound images. A supervised learning algorithm is employed to solve the automatic initialization problem of the variational methods. The proposed atherosclerotic plaque segmentation method is validated on 29 carotid ultrasound images, obtaining a Dice ratio of 90.6 ± 1.9% and an overlap index of 83.6 ± 3.2%. Moreover, by comparing the standard deviation of each evaluation index, it can be found that the proposed method is more robust for segmenting the atherosclerotic plaque. Our work shows that our proposed method can be more helpful than other variational models for measuring the carotid plaque burden.

Unsupervised domain adaptation method for segmenting cross-sectional CCA images

BACKGROUND AND OBJECTIVES

Automatic vessel segmentation in ultrasound is challenging due to the quality of the ultrasound images, which is affected by attenuation, high level of speckle noise and acoustic shadowing. Recently, deep convolutional neural networks are increasing in popularity due to their great performance on image segmentation problems, including vessel segmentation. Traditionally, large labeled datasets are required to train a network that achieves high performance, and is able to generalize well to different orientations, transducers and ultrasound scanners. However, these large datasets are rare, given that it is challenging and time-consuming to acquire and manually annotate in-vivo data.

METHODS

In this work, we present a model-based, unsupervised domain adaptation method that consists of two stages. In the first stage, the network is trained on simulated ultrasound images, which have an accurate ground truth. In the second stage, the network continues training on in-vivo data in an unsupervised way, therefore not requiring the data to be labelled. Rather than using an adversarial neural network, prior knowledge on the elliptical shape of the segmentation mask is used to detect unexpected outputs.

RESULTS

The segmentation performance was quantified using manually segmented images as ground truth. Due to the proposed domain adaptation method, the median Dice similarity coefficient increased from 0 to 0.951, outperforming a domain adversarial neural network (median Dice 0.922) and a state-of-the-art Star-Kalman algorithm that was specifically designed for this dataset (median Dice 0.942).

CONCLUSIONS

The results show that it is feasible to first train a neural network on simulated data, and then apply model-based domain adaptation to further improve segmentation performance by training on unlabeled in-vivo data. This overcomes the limitation of conventional deep learning approaches to require large amounts of manually labeled in-vivo data. Since the proposed domain adaptation method only requires prior knowledge on the shape of the segmentation mask, performance can be explored in various domains and applications in future research.

Development of an approach to extracting coronary arteries and detecting stenosis in invasive coronary angiograms

Purpose: In stable coronary artery disease (CAD), reduction in mortality and/or myocardial infarction with revascularization over medical therapy has not been reliably achieved. Coronary arteries are usually extracted to perform stenosis detection. As such, developing accurate segmentation of vascular structures and quantification of coronary arterial stenosis in invasive coronary angiograms (ICA) is necessary. Approach: A multi-input and multiscale (MIMS) U-Net with a two-stage recurrent training strategy was proposed for the automatic vessel segmentation. The proposed model generated a refined prediction map with the following two training stages: (i) stage I coarsely segmented the major coronary arteries from preprocessed single-channel ICAs and generated the probability map of arteries; and (ii) during the stage II, a three-channel image consisting of the original preprocessed image, a generated probability map, and an edge-enhanced image generated from the preprocessed image was fed to the proposed MIMS U-Net to produce the final segmentation result. After segmentation, an arterial stenosis detection algorithm was developed to extract vascular centerlines and calculate arterial diameters to evaluate stenotic level. Results: Experimental results demonstrated that the proposed method achieved an average Dice similarity coefficient of 0.8329, an average sensitivity of 0.8281, and an average specificity of 0.9979 in our dataset with 294 ICAs obtained from 73 patients. Moreover, our stenosis detection algorithm achieved a true positive rate of 0.6668 and a positive predictive value of 0.7043. Conclusions: Our proposed approach has great promise for clinical use and could help physicians improve diagnosis and therapeutic decisions for CAD.

Weakly- and Semisupervised Probabilistic Segmentation and Quantification of Reverberation Artifacts

Objective and Impact Statement. We propose a weakly- and semisupervised, probabilistic needle-and-reverberation-artifact segmentation algorithm to separate the desired tissue-based pixel values from the superimposed artifacts. Our method models the intensity decay of artifact intensities and is designed to minimize the human labeling error. Introduction. Ultrasound image quality has continually been improving. However, when needles or other metallic objects are operating inside the tissue, the resulting reverberation artifacts can severely corrupt the surrounding image quality. Such effects are challenging for existing computer vision algorithms for medical image analysis. Needle reverberation artifacts can be hard to identify at times and affect various pixel values to different degrees. The boundaries of such artifacts are ambiguous, leading to disagreement among human experts labeling the artifacts. Methods. Our learning-based framework consists of three parts. The first part is a probabilistic segmentation network to generate the soft labels based on the human labels. These soft labels are input into the second part which is the transform function, where the training labels for the third part are generated. The third part outputs the final masks which quantifies the reverberation artifacts. Results. We demonstrate the applicability of the approach and compare it against other segmentation algorithms. Our method is capable of both differentiating between the reverberations from artifact-free patches and modeling the intensity fall-off in the artifacts. Conclusion. Our method matches state-of-the-art artifact segmentation performance and sets a new standard in estimating the per-pixel contributions of artifact vs underlying anatomy, especially in the immediately adjacent regions between reverberation lines. Our algorithm is also able to improve the performance of downstream image analysis algorithms.

3D Reconstruction of Carotid Artery from Ultrasound Images

3D reconstruction is an important area in computer vision, which can be applied to assist in medical diagnosis. Compared to observing 2D ultrasound images, 3D models are more suitable for diagnostic interpretation. In this paper, we describe an approach for 3D reconstruction of the carotid artery utilizing ultrasound images from the transverse and longitudinal views. We implement a human-computer interface to ensure the accuracy of the segmentation results by involving superpixels and ellipse fitting techniques. This approach is expected to achieve better accuracy to assist diagnostics in the future.

A CNN Segmentation-Based Approach to Object Detection and Tracking in Ultrasound Scans with Application to the Vagus Nerve Detection

Ultrasound scanning is essential in several medical diagnostic and therapeutic applications. It is used to visualize and analyze anatomical features and structures that influence treatment plans. However, it is both labor intensive, and its effectiveness is operator dependent. Real-time accurate and robust automatic detection and tracking of anatomical structures while scanning would significantly impact diagnostic and therapeutic procedures to be consistent and efficient. In this paper, we propose a deep learning framework to automatically detect and track a specific anatomical target structure in ultrasound scans. Our framework is designed to be accurate and robust across subjects and imaging devices, to operate in real-time, and to not require a large training set. It maintains a localization precision and recall higher than 90% when trained on training sets that are as small as 20% in size of the original training set. The framework backbone is a weakly trained segmentation neural network based on U-Net. We tested the framework on two different ultrasound datasets with the aim to detect and track the Vagus nerve, where it outperformed current state-of-the-art real-time object detection networks.Clinical Relevance-The proposed approach provides an accurate method to detect and localize target anatomical structures in real-time, assisting sonographers during ultrasound scanning sessions by reducing diagnostic and detection errors, and expediting the duration of scanning sessions.

Shape constraint function for artery tracking in ultrasound images

Ultrasound guided regional anesthesia (UGRA) has emerged as a powerful technique for pain management in the operating theatre. It uses ultrasound imaging to visualize anatomical structures, the needle insertion and the delivery of the anesthetic around the targeted nerve block. Detection of the nerves is a difficult task, however, due to the poor quality of the ultrasound images. Recent developments in pattern recognition and machine learning have heightened the need for computer aided systems in many applications. This type of system can improve UGRA practice. In many imaging situations nerves are not salient in images. Generally, practitioners rely on the arteries as key anatomical structures to confirm the positions of the nerves, making artery tracking an important aspect for UGRA procedure. However, artery tracking in a noisy environment is a challenging problem, due to the instability of the features. This paper proposes a new method for real-time artery tracking in ultrasound images. It is based on shape information to correct tracker location errors. A new objective function is proposed, which defines an artery as an elliptical shape, enabling its robust fitting in a noisy environment. This approach is incorporated in two well-known tracking algorithms, and shows a systematic improvement over the original trackers. Evaluations were performed on 71 videos of different axillary nerve blocks. The results obtained demonstrated the validity of the proposed method.

Ultrasound-Based Fluid-Structure Interaction Modeling of Abdominal Aortic Aneurysms Incorporating Pre-stress

Currently, the prediction of rupture risk in abdominal aortic aneurysms (AAAs) solely relies on maximum diameter. However, wall mechanics and hemodynamics have shown to provide better risk indicators. Patient-specific fluid-structure interaction (FSI) simulations based on a non-invasive image modality are required to establish a patient-specific risk indicator. In this study, a robust framework to execute FSI simulations based on time-resolved three-dimensional ultrasound (3D+t US) data was obtained and employed on a data set of 30 AAA patients. Furthermore, the effect of including a pre-stress estimation (PSE) to obtain the stresses present in the measured geometry was evaluated. The established workflow uses the patient-specific 3D+t US-based segmentation and brachial blood pressure as input to generate meshes and boundary conditions for the FSI simulations. The 3D+t US-based FSI framework was successfully employed on an extensive set of AAA patient data. Omitting the pre-stress results in increased displacements, decreased wall stresses, and deviating time-averaged wall shear stress and oscillatory shear index patterns. These results underline the importance of incorporating pre-stress in FSI simulations. After validation, the presented framework provides an important tool for personalized modeling and longitudinal studies on AAA growth and rupture risk.

Automatic extraction and stenosis evaluation of coronary arteries in invasive coronary angiograms

BACKGROUND

Coronary artery disease (CAD) is the leading cause of death in the United States (US) and a major contributor to healthcare cost. Accurate segmentation of coronary arteries and detection of stenosis from invasive coronary angiography (ICA) are crucial in clinical decision making.

PURPOSE

We aim to develop an automatic method to extract coronary arteries by deep learning and detect arterial stenosis from ICAs.

METHODS

In this study, a deep learning model which integrates a feature pyramid with a U-Net++ model was developed to automatically segment coronary arteries in ICAs. A compound loss function which contains Dice loss, dilated Dice loss, and L2 regularization was utilized to train the proposed segmentation model. Following the segmentation, an algorithm which extracts vascular centerlines, calculates the diameters, and measures the stenotic levels, was developed to detect arterial stenosis.

RESULTS AND CONCLUSIONS

In the dataset consisting of 314 ICAs obtained from 99 patients, the segmentation model achieved an average Dice score of 0.8899, a sensitivity of 0.8595, and a specificity of 0.9960. In addition, the stenosis detection algorithm achieved a true positive rate of 0.6840 and a positive predictive value of 0.6998 on all types of stenosis, which has great promise to advance to clinical uses and could provide auxiliary suggestions for CAD diagnosis and treatment.

Vessel segmentation for automatic registration of untracked laparoscopic ultrasound to CT of the liver

Purpose:

Registration of Laparoscopic Ultrasound (LUS) to a pre-operative scan such as Computed Tomography (CT) using blood vessel information has been proposed as a method to enable image-guidance for laparoscopic liver resection. Currently, there are solutions for this problem that can potentially enable clinical translation by bypassing the need for a manual initialisation and tracking information. However, no reliable framework for the segmentation of vessels in 2D untracked LUS images has been presented.

Methods:

We propose the use of 2D UNet for the segmentation of liver vessels in 2D LUS images. We integrate these results in a previously developed registration method, and show the feasibility of a fully automatic initialisation to the LUS to CT registration problem without a tracking device.

Results:

We validate our segmentation using LUS data from 6 patients. We test multiple models by placing patient datasets into different combinations of training, testing and hold-out, and obtain mean Dice scores ranging from 0.543 to 0.706. Using these segmentations, we obtain registration accuracies between 6.3 and 16.6 mm in 50% of cases.

Conclusions:

We demonstrate the first instance of deep learning (DL) for the segmentation of liver vessels in LUS. Our results show the feasibility of UNet in detecting multiple vessel instances in 2D LUS images, and potentially automating a LUS to CT registration pipeline.

Real-Time Multi-Modal Sensing and Feedback for Catheterization in Porcine Tissue

Objective: In this study, we introduce a multi-modal sensing and feedback framework aimed at assisting clinicians during endovascular surgeries and catheterization procedures. This framework utilizes state-of-the-art imaging and sensing sub-systems to produce a 3D visualization of an endovascular catheter and surrounding vasculature without the need for intra-operative X-rays. Methods: The catheterization experiments within this study are conducted inside a porcine limb undergoing motions. A hybrid position-force controller of a robotically-actuated ultrasound (US) transducer for uneven porcine tissue surfaces is introduced. The tissue, vasculature, and catheter are visualized by integrated real-time US images, 3D surface imaging, and Fiber Bragg Grating (FBG) sensors. Results: During externally-induced limb motions, the vasculature and catheter can be reliably reconstructed at mean accuracies of 1.9±0.3 mm and 0.82±0.21 mm, respectively. Conclusions: The conventional use of intra-operative X-ray imaging to visualize instruments and vasculature in the human body can be reduced by employing improved diagnostic technologies that do not operate via ionizing radiation or nephrotoxic contrast agents. Significance: The presented multi-modal framework enables the radiation-free and accurate reconstruction of significant tissues and instruments involved in catheterization procedures.

An Angle-Independent Cross-Sectional Doppler Method for Flow Estimation in the Common Carotid Artery

The Doppler ultrasound is the most common technique for noninvasive quantification of blood flow, which, in turn, is of major clinical importance for the assessment of the cardiovascular condition. In this article, a method is proposed in which the vessel is imaged in the short axis, which has the advantage of capturing the whole flow profile while measuring the vessel area simultaneously. This view is easier to obtain than the longitudinal image that is currently used in flow velocity estimation, reducing operator dependence. However, the Doppler angle in cross-sectional images is unknown since the vessel wall can no longer be used to estimate the flow direction. The proposed method to estimate the Doppler angle in these images is based on the elliptical intersection between a cylindrical vessel and the ultrasound plane. The parameters of this ellipse (major axis, minor axis, and rotation) are used to estimate the Doppler angle by solving a least-squares problem. Theoretical feasibility was shown in a geometrical model, after which the Doppler angle was estimated in simulated ultrasound images generated in Field II, yielding a mean error within 4°. In vitro, across 15 short-axis measurements with a wide variety of Doppler angles, errors in the flow estimates were below 10%, and in vivo, the average velocities in systole obtained from longitudinal ( v=69.1 cm/s) and cross-sectional ( v=66.5 cm/s) acquisitions were in agreement. Further research is required to validate these results on a larger population.

Automated 3D geometry segmentation of the healthy and diseased carotid artery in free-hand, probe tracked ultrasound images

PURPOSE

Rupture of an arterosclerotic plaque in the carotid artery is a major cause of stroke. Biomechanical analysis of plaques is under development aiming to aid the clinician in the assessment of plaque vulnerability. Patient-specific three-dimensional (3D) geometry assessment of the carotid artery, including the bifurcation, is required as input for these biomechanical models. This requires a high-resolution, 3D, noninvasive imaging modality such as ultrasound (US). In this study, a high-resolution two-dimensional (2D) linear array in combination with a magnetic probe tracking device and automatic segmentation method was used to assess the geometry of the carotid artery. The advantages of using this system over a 3D ultrasound probe are its higher resolution (spatial and temporal) and its larger field of view.

METHODS

A slow sweep (v = ± 5 mm/s) was made over the subject's neck so that the full geometry of the bifurcated geometry of the carotid artery is captured. An automated segmentation pipeline was developed. First, the Star-Kalman method was used to approximate the center and size of the vessels for every frame. Images were filtered with a Gaussian high-pass filter before conversion into the 2D monogenic signals, and multiscale asymmetry features were extracted from these data, enhancing low lateral wall-lumen contrast. These images, in combination with the initial ellipse contours, were used for an active deformable contour model to segment the vessel lumen. To segment the lumen-plaque boundary, Otsu's automatic thresholding method was used. Distension of the wall due to the change in blood pressure was removed using a filter approach. Finally, the contours were converted into a 3D hexahedral mesh for a patient-specific solid mechanics model of the complete arterial wall.

RESULTS

The method was tested on 19 healthy volunteers and on 3 patients. The results were compared to manual segmentation performed by three experienced observers. Results showed an average Hausdorff distance of 0.86 mm and an average similarity index of 0.91 for the common carotid artery (CCA) and 0.88 for the internal and external carotid artery. For the total algorithm, the success rate was 89%, in 4 out of 38 datasets the ICA and ECA were not sufficient visible in the US images. Accurate 3D hexahedral meshes were successfully generated from the segmented images .

CONCLUSIONS

With this method, a subject-specific biomechanical model can be constructed directly from a hand-held 2D US measurement, within 10 min, with a minimal user input. The performance of the proposed segmentation algorithm is comparable to or better than algorithms previously described in literature. Moreover, the algorithm is able to segment the CCA, ICA, and ECA including the carotid bifurcation in transverse B-mode images in both healthy and diseased arteries.

Deep Learning-Based Femoral Cartilage Automatic Segmentation in Ultrasound Imaging for Guidance in Robotic Knee Arthroscopy

Knee arthroscopy is a minimally invasive surgery used in the treatment of intra-articular knee pathology which may cause unintended damage to femoral cartilage. An ultrasound (US)-guided autonomous robotic platform for knee arthroscopy can be envisioned to minimise these risks and possibly to improve surgical outcomes. The first necessary tool for reliable guidance during robotic surgeries was an automatic segmentation algorithm to outline the regions at risk. In this work, we studied the feasibility of using a state-of-the-art deep neural network (UNet) to automatically segment femoral cartilage imaged with dynamic volumetric US (at the refresh rate of 1 Hz), under simulated surgical conditions. Six volunteers were scanned which resulted in the extraction of 18278 2-D US images from 35 dynamic 3-D US scans, and these were manually labelled. The UNet was evaluated using a five-fold cross-validation with an average of 15531 training and 3124 testing labelled images per fold. An intra-observer study was performed to assess intra-observer variability due to inherent US physical properties. To account for this variability, a novel metric concept named Dice coefficient with boundary uncertainty (DSC_UB) was proposed and used to test the algorithm. The algorithm performed comparably to an experienced orthopaedic surgeon, with DSC_UB of 0.87. The proposed UNet has the potential to localise femoral cartilage in robotic knee arthroscopy with clinical accuracy.

Bimodal Automated Carotid Ultrasound Segmentation Using Geometrically Constrained Deep Neural Networks

For asymptomatic patients suffering from carotid stenosis, the assessment of plaque morphology is an important clinical task which allows monitoring of the risk of plaque rupture and future incidents of stroke. Ultrasound Imaging provides a safe and non-invasive modality for this, and the segmentation of media-adventitia boundaries and lumen-intima boundaries of the Carotid artery form an essential part in this monitoring process. In this paper, we propose a novel Deep Neural Network as a fully automated segmentation tool, and its application in delineating both the media-adventitia boundary and the lumen-intima boundary. We develop a new geometrically constrained objective function as part of the Network's Stochastic Gradient Descent optimisation, thus tuning it to the problem at hand. Furthermore, we also apply a bimodal fusion of amplitude and phase congruency data proposed by us in previous work, as an input to the network, as the latter provides an intensity-invariant data source to the network. We finally report the segmentation performance of the network on transverse sections of the carotid. Tests are carried out on an augmented dataset of 81,000 images, and the results are compared to other studies by reporting the DICE coefficient of similarity, modified Hausdorff Distance, sensitivity and specificity. Our proposed modification is shown to yield improved results on the standard network over this larger dataset, with the advantage of it being fully automated. We conclude that Deep Neural Networks provide a reliable trained manner in which carotid ultrasound images may be automatically segmented, using amplitude data and intensity invariant phase congruency maps as a data source.

Siam-U-Net: encoder-decoder siamese network for knee cartilage tracking in ultrasound images

The tracking of the knee femoral condyle cartilage during ultrasound-guided minimally invasive procedures is important to avoid damaging this structure during such interventions. In this study, we propose a new deep learning method to track, accurately and efficiently, the femoral condyle cartilage in ultrasound sequences, which were acquired under several clinical conditions, mimicking realistic surgical setups. Our solution, that we name Siam-U-Net, requires minimal user initialization and combines a deep learning segmentation method with a siamese framework for tracking the cartilage in temporal and spatio-temporal sequences of 2D ultrasound images. Through extensive performance validation given by the Dice Similarity Coefficient, we demonstrate that our algorithm is able to track the femoral condyle cartilage with an accuracy which is comparable to experienced surgeons. It is additionally shown that the proposed method outperforms state-of-the-art segmentation models and trackers in the localization of the cartilage. We claim that the proposed solution has the potential for ultrasound guidance in minimally invasive knee procedures.

Far-field acoustic subwavelength imaging and edge detection based on spatial filtering and wave vector conversion

The resolution of acoustic imaging suffers from diffraction limit due to the loss of evanescent field that carries subwavelength information. Most of the current methods for overcoming the diffraction limit in acoustics still operate in the near-field of the object. Here we demonstrate the design and experimental realization of an acoustic far-field subwavelength imaging system. Our system is based on wave vector filtering and conversion with a transmitter at the near-field and a spatially symmetrical receiver at the far-field. By tuning geometric parameters of the transmitting/receiving pair, different spatial frequency bands can be separated and projected to the far-field. Furthermore, far-field imaging and edge detection of subwavelength objects are experimentally demonstrated. The proposed system brings new possibilities for far-field subwavelength wave manipulation, which can be further applied to medical imaging, nondestructive testing, and acoustic communication.

Plasmonic effects and subwavelength scattering arrays are used in the optical domain to access subwavelength resolution imaging in the far field. Here, the authors develop an analogous strategy for far-field, subwavelength imaging at acoustic wavelengths and demonstrate edge detection of acoustic scattering objects.

Semi-Automatic Algorithms for Estimation and Tracking of AP-Diameter of the IVC in Ultrasound Images

Acutely ill patients presenting with conditions such as sepsis, trauma, and congestive heart failure require judicious resuscitation in order to achieve and maintain optimal circulating blood volume. Increasingly, emergency and critical care physicians are using portable ultrasound to approximate the temporal changes of the anterior–posterior (AP)-diameter of the inferior vena cava (IVC) in order to guide fluid administration or removal. This paper proposes semi-automatic active ellipse and rectangle algorithms capable of improved and quantified measurement of the AP-diameter. The proposed algorithms are compared to manual measurement and a previously published active circle model. Results demonstrate that the rectangle model outperforms both active circle and ellipse irrespective of IVC shape and closely approximates tedious expert assessment.

Real-Time Blood Pressure Estimation From Force-Measured Ultrasound

OBJECTIVE

Our objective is to create a blood pressure measurement device, which may provide a way to easily acquire frequent measurements. Common techniques to measure blood pressure include an arterial catheter, an oscillometric pressure cuff, or an auscultatory pressure cuff.

METHODS

The approach takes as input ultrasound images of an artery and contact force between the ultrasound array and subject. A subject may perform the self-measurements. Image and force data is analyzed for its quality and used to provide guidance or reject poor measurements. Tissue motions, due to probe contact forces and pulsing blood pressure, are estimated from the ultrasound image. Tissues elasticities and blood pressure are found by optimally fitting the observed tissue motion versus applied forces to a table of predicted motion-pre-generated with a finite element tissue deformation model. The output of the optimization is an estimate of systolic and diastolic blood pressure, arterial stiffness, and surrounding tissue stiffness.

RESULTS

The real-time implementation of the algorithm was validated on a cohort of 21 single-visit volunteers and on four volunteers self-monitored longitudinally. The systolic and diastolic pressures were compared to oscillometric cuff readings. Regression and Bland-Altman analyses were performed.

CONCLUSION

Systolic pressure and diastolic pressure can be estimated in real-time and by the subject using this novel non-invasive ultrasound-based method (systolic accuracy/precision: -5.2 mmHg/10.7 mmHg; diastolic accuracy/precision: -3.9/8.0 mmHg).

SIGNIFICANCE

The method occupies a middle ground between the arterial catheter and cuff-based techniques. It has the potential to give calibration-free results.

Ultrasound Image Segmentation: A Deeply Supervised Network With Attention to Boundaries

OBJECTIVE

Segmentation of anatomical structures in ultrasound images requires vast radiological knowledge and experience. Moreover, the manual segmentation often results in subjective variations, therefore, an automatic segmentation is desirable. We aim to develop a fully convolutional neural network (FCNN) with attentional deep supervision for the automatic and accurate segmentation of the ultrasound images.

METHOD

FCNN/CNNs are used to infer high-level context using low-level image features. In this paper, a sub-problem specific deep supervision of the FCNN is performed. The attention of fine resolution layers is steered to learn object boundary definitions using auxiliary losses, whereas coarse resolution layers are trained to discriminate object regions from the background. Furthermore, a customized scheme for downweighting the auxiliary losses and a trainable fusion layer are introduced. This produces an accurate segmentation and helps in dealing with the broken boundaries, usually found in the ultrasound images.

RESULTS

The proposed network is first tested for blood vessel segmentation in liver images. It results in F1 score, mean intersection over union, and dice index of 0.83, 0.83, and 0.79, respectively. The best values observed among the existing approaches are produced by U-net as 0.74, 0.81, and 0.75, respectively. The proposed network also results in dice index value of 0.91 in the lumen segmentation experiments on MICCAI 2011 IVUS challenge dataset, which is near to the provided reference value of 0.93. Furthermore, the improvements similar to vessel segmentation experiments are also observed in the experiment performed to segment lesions.

CONCLUSION

Deep supervision of the network based on the input-output characteristics of the layers results in improvement in overall segmentation accuracy.

SIGNIFICANCE

Sub-problem specific deep supervision for ultrasound image segmentation is the main contribution of this paper. Currently the network is trained and tested for fixed size inputs. It requires image resizing and limits the performance in small size images.

Highlighting nerves and blood vessels for ultrasound-guided axillary nerve block procedures using neural networks

Ultrasound images acquired during axillary nerve block procedures can be difficult to interpret. Highlighting the important structures, such as nerves and blood vessels, may be useful for the training of inexperienced users. A deep convolutional neural network is used to identify the musculocutaneous, median, ulnar, and radial nerves, as well as the blood vessels in ultrasound images. A dataset of 49 subjects is collected and used for training and evaluation of the neural network. Several image augmentations, such as rotation, elastic deformation, shadows, and horizontal flipping, are tested. The neural network is evaluated using cross validation. The results showed that the blood vessels were the easiest to detect with a precision and recall above 0.8. Among the nerves, the median and ulnar nerves were the easiest to detect with an F -score of 0.73 and 0.62, respectively. The radial nerve was the hardest to detect with an F -score of 0.39. Image augmentations proved effective, increasing F -score by as much as 0.13. A Wilcoxon signed-rank test showed that the improvement from rotation, shadow, and elastic deformation augmentations were significant and the combination of all augmentations gave the best result. The results are promising; however, there is more work to be done, as the precision and recall are still too low. A larger dataset is most likely needed to improve accuracy, in combination with anatomical and temporal models.

Noninvasive Blood Pressure Estimation Using Ultrasound and Simple Finite Element Models

OBJECTIVE

Many commercially available arterial blood pressure measurement devices suffer from a range of weaknesses. For example, common weaknesses include being inaccurate, invasive, and ad hoc; many also require explicit user calibration or cut off blood flow to a limb. A novel algorithmic approach is presented to accurately estimate systolic and diastolic blood pressure in a way that does not require any explicit user calibration, is noninvasive, and does not cut off blood flow.

METHODS

The approach uses ultrasound images of the arterial wall and corresponding contact force data to obtain blood pressure estimates. To acquire data, an ultrasound probe was placed on the patient's carotid artery and the contact force was increased from 1.5 to 12 N. The artery was then algorithmically segmented from the recorded DICOM B-Mode data. The segmentation data and the contact force were used as input into the Levenberg-Marquardt optimization method to solve for the parameters, including blood pressure, of a simple finite element model of the carotid artery.

RESULTS

The algorithm was validated on 24 healthy volunteers. Algorithm arterial blood pressure predictions were compared to oscillometric blood pressure cuff readings. Regression and Bland-Altman analyses were performed on the data.

CONCLUSION

Both systolic pressure and diastolic pressure can be estimated using this novel noninvasive ultrasound-based method (systolic accuracy/precision: $-$ 2.36 mmHg/10.21 mmHg; diastolic accuracy/precision: $-$ 0.32/8.23 mmHg).

SIGNIFICANCE

The method occupies a clinical middle ground between the arterial catheter and cuff-based techniques. It has the potential to give accurate results for patients with hypertension and atherosclerosis.

Arterial blood pressure estimation using ultrasound: Clinical results on healthy volunteers and a medicated hypertensive volunteer

This study presents a non-occlusive and non-invasive ultrasound-based technique to measure blood pressure. Most popular clinically-used arterial blood pressure measurement techniques suffer from important weaknesses including being inaccurate, invasive, or occlusive. In the proposed technique, an ultrasound probe is placed on the patient's carotid artery and the contact force between the probe and the tissue is slowly increased while ultrasound images and contact force data are recorded. From this data, the artery is segmented and the segmentation data is sent into an optimization procedure; after post-processing, blood pressure is displayed to the user. This technique was applied to 24 healthy single-visit volunteers, one multi-visit healthy volunteer, and one multi-visit medicated hypertensive volunteer. Compared to the oscillometric cuff, the accuracy and precision of the algorithm-reported systolic pressure is -2.4 ± 10.2 mmHg, and for diastolic pressure is -0.3 ± 8.2 mmHg. This method has the potential to occupy a clinical middle-ground between the arterial catheter and the oscillometric cuff.

Adaptive median binary patterns for fully automatic nerves tracking in ultrasound images

BACKGROUND AND OBJECTIVE

In the last decade, Ultrasound-Guided Regional Anesthesia (UGRA) gained importance in surgical procedures and pain management, due to its ability to perform target delivery of local anesthetics under direct sonographic visualization. However, practicing UGRA can be challenging, since it requires high skilled and experienced operator. Among the difficult task that the operator can face, is the tracking of the nerve structure in ultrasound images. Tracking task in US images is very challenging due to the noise and other artifacts.

METHODS

In this paper, we introduce a new and robust tracking technique by using Adaptive Median Binary Pattern(AMBP) as texture feature for tracking algorithms (particle filter, mean-shift and Kanade-Lucas-Tomasi(KLT)). Moreover, we propose to incorporate Kalman filter as prediction and correction steps for the tracking algorithms, in order to enhance the accuracy, computational cost and handle target disappearance.

RESULTS

The proposed method have been applied on real data and evaluated in different situations. The obtained results show that tracking with AMBP features outperforms other descriptors and achieved best performance with 95% accuracy.

CONCLUSIONS

This paper presents the first fully automatic nerve tracking method in Ultrasound images. AMBP features outperforms other descriptors in all situations such as noisy and filtered images.

Estimation and tracking of AP-diameter of the inferior vena cava in ultrasound images using a novel active circle algorithm

Medical research suggests that the anterior-posterior (AP)-diameter of the inferior vena cava (IVC) and its associated temporal variation as imaged by bedside ultrasound is useful in guiding fluid resuscitation of the critically-ill patient. Unfortunately, indistinct edges and gaps in vessel walls are frequently present which impede accurate estimation of the IVC AP-diameter for both human operators and segmentation algorithms. The majority of research involving use of the IVC to guide fluid resuscitation involves manual measurement of the maximum and minimum AP-diameter as it varies over time. This effort proposes using a time-varying circle fitted inside the typically ellipsoid IVC as an efficient, consistent and novel approach to tracking and approximating the AP-diameter even in the context of poor image quality. In this active-circle algorithm, a novel evolution functional is proposed and shown to be a useful tool for ultrasound image processing. The proposed algorithm is compared with an expert manual measurement, and state-of-the-art relevant algorithms. It is shown that the algorithm outperforms other techniques and performs very close to manual measurement.

Echogenicity based approach to detect, segment and track the common carotid artery in 2D ultrasound images

Automatic detection and segmentation of the common carotid artery in transverse ultrasound (US) images of the thyroid gland play a vital role in the success of US guided intervention procedures. We propose in this paper a novel method to accurately detect, segment and track the carotid in 2D and 2D+t US images of the thyroid gland using concepts based on tissue echogenicity and ultrasound image formation. We first segment the hypoechoic anatomical regions of interest using local phase and energy in the input image. We then make use of a Hessian based blob like analysis to detect the carotid within the segmented hypoechoic regions. The carotid artery is segmented by making use of least squares ellipse fit for the edge points around the detected carotid candidate. Experiments performed on a multivendor dataset of 41 images show that the proposed algorithm can segment the carotid artery with high sensitivity (99.6 ±m 0.2%) and specificity (92.9 ±m 0.1%). Further experiments on a public database containing 971 images of the carotid artery showed that the proposed algorithm can achieve a detection accuracy of 95.2% with a 2% increase in performance when compared to the state-of-the-art method.

Analysis of Carotid Artery Transverse Sections in Long Ultrasound Video Sequences

Examination of the common carotid artery (CCA) based on an ultrasound video sequence is an effective method for detecting cardiovascular diseases. Here, we propose a video processing method for the automated geometric analysis of CCA transverse sections. By explicitly compensating the parasitic phenomena of global movement and feature drift, our method enables a reliable and accurate estimation of the movement of the arterial wall based on ultrasound sequences of arbitrary length and in situations where state-of-the-art methods fail or are very inaccurate. The method uses a modified Viola-Jones detector and the Hough transform to localize the artery in the image. Then it identifies dominant scatterers, also known as interest points (IPs), whose positions are tracked by means of the pyramidal Lucas-Kanade method. Robustness to global movement and feature drift is achieved by a detection of global movement and subsequent IP re-initialization, as well as an adaptive removal and addition of IPs. The performance of the proposed method is evaluated using simulated and real ultrasound video sequences. Using the Harris detector for IP detection, we obtained an overall root-mean-square error, averaged over all the simulated sequences, of 2.16 ± 1.18 px. The computational complexity of our method is compatible with real-time operation; the runtime is about 30-70 ms/frame for sequences with a spatial resolution of up to 490 × 490 px. We expect that in future clinical practice, our method will be instrumental for non-invasive early-stage diagnosis of atherosclerosis and other cardiovascular diseases.

An integrated method for atherosclerotic carotid plaque segmentation in ultrasound image

BACKGROUND AND OBJECTIVE

Carotid artery atherosclerosis is an important cause of stroke. Ultrasound imaging has been widely used in the diagnosis of atherosclerosis. Therefore, segmenting atherosclerotic carotid plaque in ultrasound image is an important task. Accurate plaque segmentation is helpful for the measurement of carotid plaque burden. In this paper, we propose and evaluate a novel learning-based integrated framework for plaque segmentation.

METHODS

In our study, four different classification algorithms, along with the auto-context iterative algorithm, were employed to effectively integrate features from ultrasound images and later also the iteratively estimated and refined probability maps together for pixel-wise classification. The four classification algorithms were support vector machine with linear kernel, support vector machine with radial basis function kernel, AdaBoost and random forest. The plaque segmentation was implemented in the generated probability map. The performance of the four different learning-based plaque segmentation methods was tested on 29 B-mode ultrasound images. The evaluation indices for our proposed methods were consisted of sensitivity, specificity, Dice similarity coefficient, overlap index, error of area, absolute error of area, point-to-point distance, and Hausdorff point-to-point distance, along with the area under the ROC curve.

RESULTS

The segmentation method integrated the random forest and an auto-context model obtained the best results (sensitivity 80.4 ± 8.4%, specificity 96.5 ± 2.0%, Dice similarity coefficient 81.0 ± 4.1%, overlap index 68.3 ± 5.8%, error of area -1.02 ± 18.3%, absolute error of area 14.7 ± 10.9%, point-to-point distance 0.34 ± 0.10 mm, Hausdorff point-to-point distance 1.75 ± 1.02 mm, and area under the ROC curve 0.897), which were almost the best, compared with that from the existed methods.

CONCLUSIONS

Our proposed learning-based integrated framework investigated in this study could be useful for atherosclerotic carotid plaque segmentation, which will be helpful for the measurement of carotid plaque burden.

Automatic identification of blood vessel cross-section for central venous catheter placement using a cascading classifier

This paper presents a system that is able to automatically identify, segment, and track the cross-section of the internal jugular vein (IJV) and the common carotid artery (CCA) in an ultrasound image feed during a central venous catheter (CVC) placement procedure. The goal is to provide assistance to the practitioner in order to decrease the probability of complications stemming from inadvertent punctures of the CCA during the procedure. In the system, a modified Star algorithm is implemented to segment and track the blood vessel throughout an ultrasound video feed. A novel algorithm based on a cascading classifier is used to identify the location of the IJV and the CCA for two main tasks: (1) selecting the initial seed point at the start of tracking and (2) validating the segmentation results at each subsequent frame. The classifier uses shape features (vessel area, ellipse fitting error, vessel depth, vessel eccentricity) and pixel-based features (pixel intensity and the histogram of oriented gradients descriptor) to differentiate between vessel and non-vessel structures and also differentiate between the IJV and the CCA. Evaluated on a database of 800 ultrasound images containing the cross-section of both vessels, the cascading classifier was able to identify the IJV and the CCA in 92.25% and 85.13% of the images respectively without any initialization from the user at a maximum processing rate of 40.65 frames per second. This allows identification to be conducted in real-time with existing ultrasound machines.

Freehand three-dimensional ultrasound imaging of carotid artery using motion tracking technology

Ultrasound imaging has been extensively used for determining the severity of carotid atherosclerotic stenosis. In particular, the morphological characterization of carotid plaques can be performed for risk stratification of patients. However, using 2D ultrasound imaging for detecting morphological changes in plaques has several limitations. Due to the scan was performed on a single longitudinal cross-section, the selected 2D image is difficult to represent the entire morphology and volume of plaque and vessel lumen. In addition, the precise positions of 2D ultrasound images highly depend on the radiologists' experience, it makes the serial long-term exams of anti-atherosclerotic therapies are difficult to relocate the same corresponding planes by using 2D B-mode images. This has led to the recent development of three-dimensional (3D) ultrasound imaging, which offers improved visualization and quantification of complex morphologies of carotid plaques. In the present study, a freehand 3D ultrasound imaging technique based on optical motion tracking technology is proposed. Unlike other optical tracking systems, the marker is a small rigid body that is attached to the ultrasound probe and is tracked by eight high-performance digital cameras. The probe positions in 3D space coordinates are then calibrated at spatial and temporal resolutions of 10μm and 0.01s, respectively. The image segmentation procedure involves Otsu's and the active contour model algorithms and accurately detects the contours of the carotid arteries. The proposed imaging technique was verified using normal artery and atherosclerotic stenosis phantoms. Human experiments involving freehand scanning of the carotid artery of a volunteer were also performed. The results indicated that compared with manual segmentation, the lowest percentage errors of the proposed segmentation procedure were 7.8% and 9.1% for the external and internal carotid arteries, respectively. Finally, the effect of handshaking was calibrated using the optical tracking system for reconstructing a 3D image.

An Improved Algorithm for Coronary Bypass Anastomosis Segmentation in Epicardial Ultrasound Sequences

Epicardial ultrasound (EUS) can be used for intra-operative quality assessment of coronary artery bypass anastomoses. To quantify the anastomotic quality from EUS images, the area of anastomotic structures has to be extracted from EUS sequences. Currently, this is done manually as no objective methods are available. We used an automatic anastomosis segmentation algorithm to extract the area of anastomotic structures from in vivo EUS sequences obtained from 16 porcine anastomoses. The algorithm consists of four major components: vessel detection, vessel segmentation, segmentation quality control and inter-frame contour alignment. The segmentation accuracy was assessed using m-fold cross-validation based on 830 manual segmentations of the anastomotic structures. A Dice coefficient of 0.879 (±0.073) and an absolute area difference of 16.95% (±17.94) were obtained. The proposed segmentation algorithm has potential to automatically extract the area of anastomotic structures.

Real-Time Automatic Artery Segmentation, Reconstruction and Registration for Ultrasound-Guided Regional Anaesthesia of the Femoral Nerve

The goal is to create an assistant for ultrasound- guided femoral nerve block. By segmenting and visualizing the important structures such as the femoral artery, we hope to improve the success of these procedures. This article is the first step towards this goal and presents novel real-time methods for identifying and reconstructing the femoral artery, and registering a model of the surrounding anatomy to the ultrasound images. The femoral artery is modelled as an ellipse. The artery is first detected by a novel algorithm which initializes the artery tracking. This algorithm is completely automatic and requires no user interaction. Artery tracking is achieved with a Kalman filter. The 3D artery is reconstructed in real-time with a novel algorithm and a tracked ultrasound probe. A mesh model of the surrounding anatomy was created from a CT dataset. Registration of this model is achieved by landmark registration using the centerpoints from the artery tracking and the femoral artery centerline of the model. The artery detection method was able to automatically detect the femoral artery and initialize the tracking in all 48 ultrasound sequences. The tracking algorithm achieved an average dice similarity coefficient of 0.91, absolute distance of 0.33 mm, and Hausdorff distance 1.05 mm. The mean registration error was 2.7 mm, while the average maximum error was 12.4 mm. The average runtime was measured to be 38, 8, 46 and 0.2 milliseconds for the artery detection, tracking, reconstruction and registration methods respectively.

Locally rigid, vessel-based registration for laparoscopic liver surgery

PURPOSE

Laparoscopic liver resection has significant advantages over open surgery due to less patient trauma and faster recovery times, yet is difficult for most lesions due to the restricted field of view and lack of haptic feedback. Image guidance provides a potential solution but is challenging in a soft deforming organ such as the liver. In this paper, we therefore propose a laparoscopic ultrasound (LUS) image guidance system and study the feasibility of a locally rigid registration for laparoscopic liver surgery.

METHODS

We developed a real-time segmentation method to extract vessel centre points from calibrated, freehand, electromagnetically tracked, 2D LUS images. Using landmark-based initial registration and an optional iterative closest point (ICP) point-to-line registration, a vessel centre-line model extracted from preoperative computed tomography (CT) is registered to the ultrasound data during surgery.

RESULTS

Using the locally rigid ICP method, the RMS residual error when registering to a phantom was 0.7 mm, and the mean target registration error (TRE) for two in vivo porcine studies was 3.58 and 2.99 mm, respectively. Using the locally rigid landmark-based registration method gave a mean TRE of 4.23 mm using vessel centre lines derived from CT scans taken with pneumoperitoneum and 6.57 mm without pneumoperitoneum.

CONCLUSION

In this paper we propose a practical image-guided surgery system based on locally rigid registration of a CT-derived model to vascular structures located with LUS. In a physical phantom and during porcine laparoscopic liver resection, we demonstrate accuracy of target location commensurate with surgical requirements. We conclude that locally rigid registration could be sufficient for practically useful image guidance in the near future.

Methods for 2-D and 3-D Endobronchial Ultrasound Image Segmentation

Endobronchial ultrasound (EBUS) is now commonly used for cancer-staging bronchoscopy. Unfortunately, EBUS is challenging to use and interpreting EBUS video sequences is difficult. Other ultrasound imaging domains, hampered by related difficulties, have benefited from computer-based image-segmentation methods. Yet, so far, no such methods have been proposed for EBUS. We propose image-segmentation methods for 2-D EBUS frames and 3-D EBUS sequences. Our 2-D method adapts the fast-marching level-set process, anisotropic diffusion, and region growing to the problem of segmenting 2-D EBUS frames. Our 3-D method builds upon the 2-D method while also incorporating the geodesic level-set process for segmenting EBUS sequences. Tests with lung-cancer patient data showed that the methods ran fully automatically for nearly 80% of test cases. For the remaining cases, the only user-interaction required was the selection of a seed point. When compared to ground-truth segmentations, the 2-D method achieved an overall Dice index = 90.0% ±4.9%, while the 3-D method achieved an overall Dice index = 83.9 ± 6.0%. In addition, the computation time (2-D, 0.070 s/frame; 3-D, 0.088 s/frame) was two orders of magnitude faster than interactive contour definition. Finally, we demonstrate the potential of the methods for EBUS localization in a multimodal image-guided bronchoscopy system.

Acoustic metamaterial for subwavelength edge detection

Metamaterials have demonstrated the possibility to produce super-resolved images by restoring propagative and evanescent waves. However, for efficient information transfer, for example, in compressed sensing, it is often desirable to visualize only the fast spatial variations of the wave field (carried by evanescent waves), as the one created by edges or small details. Image processing edge detection algorithms perform such operation, but they add time and complexity to the imaging process. Here we present an acoustic metamaterial that transmits only components of the acoustic field that are approximately equal to or smaller than the operating wavelength. The metamaterial converts evanescent waves into propagative waves exciting trapped resonances, and it uses periodicity to attenuate the propagative components. This approach achieves resolutions ∼5 times smaller than the operating wavelength and makes it possible to visualize independently edges aligned along different directions.

Super-resolution imaging is based on the restoration of evanescent and propagative waves. Here, the authors present an acoustic metamaterial that transmits only components of the acoustic field equal to or smaller than the operating wavelength, which can be used to provide sharp images of the edge of an object.

Deep venous thrombosis identification from analysis of ultrasound data

PURPOSE

The purpose of this research was to determine whether combined ultrasound- and sensor-based compressibility and augmented blood flow measures yielded better results for DVT detection than for the individual measures alone.

METHODS

Twenty-six limbs from 19 patients were scanned using a sensorized ultrasound DVT screening system, and compressibility and flow measures were obtained at 125 locations. Results from conventional compression ultrasound examination were used as gold standard, with seven vessels (four patients) positive for DVT. A classification approach was used to combine the individual DVT measures per vessel and generate an optimal feature for every possible combination of individual measures. Sensitivity and specificity were calculated for the individual measures and for all combined measures, as was a usefulness criteria [Formula: see text] for measuring class separability.

RESULTS

Seven optimal combined features were found with 100% sensitivity and 100% specificity, with the best combined feature having a [Formula: see text] value over two orders of magnitude greater than the best individual DVT measure.

CONCLUSIONS

The proposed approach for DVT detection combines different aspects of thrombus detection in a novel way generating a quantifiable measure and outperforms any of the individual measures when used independently. All of the combined measures included the flow measure as well as the slope compressibility measure, which uses the magnitude of the force applied by the ultrasound probe, suggesting that these measurements provide important information when characterizing DVT.