Segmentation of the heart and great vessels in CT images using a model-based adaptation framework

HVSMR-2.0: A 3D cardiovascular MR dataset for whole-heart segmentation in congenital heart disease

Patients with congenital heart disease often have cardiac anatomy that deviates significantly from normal, frequently requiring multiple heart surgeries. Image segmentation from a preoperative cardiovascular magnetic resonance (CMR) scan would enable creation of patient-specific 3D surface models of the heart, which have potential to improve surgical planning, enable surgical simulation, and allow automatic computation of quantitative metrics of heart function. However, there is no publicly available CMR dataset for whole-heart segmentation in patients with congenital heart disease. Here, we release the HVSMR-2.0 dataset, comprising 60 CMR scans alongside manual segmentation masks of the 4 cardiac chambers and 4 great vessels. The images showcase a wide range of heart defects and prior surgical interventions. The dataset also includes masks of required and optional extents of the great vessels, enabling fairer comparisons across algorithms. Detailed diagnoses for each subject are also provided. By releasing HVSMR-2.0, we aim to encourage development of robust segmentation algorithms and clinically relevant tools for congenital heart disease.

Shape-constrained deformable brain segmentation: Methods and quantitative validation

MRI-guided neuro interventions require rapid, accurate, and reproducible segmentation of anatomical brain structures for identification of targets during surgical procedures and post-surgical evaluation of intervention efficiency. Segmentation algorithms must be validated and cleared for clinical use. This work introduces a methodology for shape-constrained deformable brain segmentation, describes the quantitative validation used for its clinical clearance, and presents a comparison with manual expert segmentation and FreeSurfer, an open source software for neuroimaging data analysis. ClearPoint Maestro is software for fully-automatic brain segmentation from T1-weighted MRI that combines a shape-constrained deformable brain model with voxel-wise tissue segmentation within the cerebral hemispheres and the cerebellum. The performance of the segmentation was validated in terms of accuracy and reproducibility. Segmentation accuracy was evaluated with respect to training data and independently traced ground truth. Segmentation reproducibility was quantified and compared with manual expert segmentation and FreeSurfer. Quantitative reproducibility analysis indicates superior performance compared to both manual expert segmentation and FreeSurfer. The shape-constrained methodology results in accurate and highly reproducible segmentation. Inherent point based-correspondence provides consistent target identification ideal for MRI-guided neuro interventions.

Generation of synthetic aortic valve stenosis geometries for in silico trials

In silico trials are a promising way to increase the efficiency of the development, and the time to market of cardiovascular implantable devices. The development of transcatheter aortic valve implantation (TAVI) devices, could benefit from in silico trials to overcome frequently occurring complications such as paravalvular leakage and conduction problems. To be able to perform in silico TAVI trials virtual cohorts of TAVI patients are required. In a virtual cohort, individual patients are represented by computer models that usually require patient-specific aortic valve geometries. This study aimed to develop a virtual cohort generator that generates anatomically plausible, synthetic aortic valve stenosis geometries for in silico TAVI trials and allows for the selection of specific anatomical features that influence the occurrence of complications. To build the generator, a combination of non-parametrical statistical shape modeling and sampling from a copula distribution was used. The developed virtual cohort generator successfully generated synthetic aortic valve stenosis geometries that are comparable with a real cohort, and therefore, are considered as being anatomically plausible. Furthermore, we were able to select specific anatomical features with a sensitivity of around 90%. The virtual cohort generator has the potential to be used by TAVI manufacturers to test their devices. Future work will involve including calcifications to the synthetic geometries, and applying high-fidelity fluid-structure-interaction models to perform in silico trials.

Automated detection, segmentation and measurement of major vessels and the trachea in CT pulmonary angiography

Mediastinal structure measurements are important for the radiologist's review of computed tomography pulmonary angiography (CTPA) examinations. In the reporting process, radiologists make measurements of diameters, volumes, and organ densities for image quality assessment and risk stratification. However, manual measurement of these features is time consuming. Here, we sought to develop a time-saving automated algorithm that can accurately detect, segment and measure mediastinal structures in routine clinical CTPA examinations. In this study, 700 CTPA examinations collected and annotated. Of these, a training set of 180 examinations were used to develop a fully automated deterministic algorithm. On the test set of 520 examinations, two radiologists validated the detection and segmentation performance quantitatively, and ground truth was annotated to validate the measurement performance. External validation was performed in 47 CTPAs from two independent datasets. The system had 86-100% detection and segmentation accuracy in the different tasks. The automatic measurements correlated well to those of the radiologist (Pearson's r 0.68-0.99). Taken together, the fully automated algorithm accurately detected, segmented, and measured mediastinal structures in routine CTPA examinations having an adequate representation of common artifacts and medical conditions.

MWG-UNet: Hybrid Deep Learning Framework for Lung Fields and Heart Segmentation in Chest X-ray Images

Deep learning technology has achieved breakthrough research results in the fields of medical computer vision and image processing. Generative adversarial networks (GANs) have demonstrated a capacity for image generation and expression ability. This paper proposes a new method called MWG-UNet (multiple tasking Wasserstein generative adversarial network U-shape network) as a lung field and heart segmentation model, which takes advantages of the attention mechanism to enhance the segmentation accuracy of the generator so as to improve the performance. In particular, the Dice similarity, precision, and F1 score of the proposed method outperform other models, reaching 95.28%, 96.41%, and 95.90%, respectively, and the specificity surpasses the sub-optimal models by 0.28%, 0.90%, 0.24%, and 0.90%. However, the value of the IoU is inferior to the optimal model by 0.69%. The results show the proposed method has considerable ability in lung field segmentation. Our multi-organ segmentation results for the heart achieve Dice similarity and IoU values of 71.16% and 74.56%. The segmentation results on lung fields achieve Dice similarity and IoU values of 85.18% and 81.36%.

Segment aorta and localize landmarks simultaneously on noncontrast CT using a multitask learning framework for patients without severe vascular disease

BACKGROUND

Non-contrast chest CT is widely used for lung cancer screening, and its images carry potential information of the thoracic aorta. The morphological assessment of the thoracic aorta may have potential value in the presymptomatic detection of thoracic aortic-related diseases and the risk prediction of future adverse events. However, due to low vasculature contrast in such images, visual assessment of aortic morphology is challenging and highly depends on physicians' experience.

PURPOSE

The main objective of this study is to propose a novel multi-task framework based on deep learning for simultaneous aortic segmentation and localization of key landmarks on unenhanced chest CT. The secondary objective is to use the algorithm to measure quantitative features of thoracic aorta morphology.

METHODS

The proposed network is composed of two subnets to carry out segmentation and landmark detection, respectively. The segmentation subnet aims to demarcate the aortic sinuses of the Valsalva, aortic trunk and aortic branches, whereas the detection subnet is devised to locate five landmarks on the aorta to facilitate morphology measures. The networks share a common encoder and run decoders in parallel, taking full advantage of the synergy of the segmentation and landmark detection tasks. Furthermore, the volume of interest (VOI) module and the squeeze-and-excitation (SE) block with attention mechanisms are incorporated to further boost the capability of feature learning.

RESULTS

Benefiting from the multitask framework, we achieved a mean Dice score of 0.95, average symmetric surface distance of 0.53 mm, Hausdorff distance of 2.13 mm for aortic segmentation, and mean square error (MSE) of 3.23 mm for landmark localization in 40 testing cases.

CONCLUSION

We proposed a multitask learning framework which can perform segmentation of the thoracic aorta and localization of landmarks simultaneously and achieved good results. It can support quantitative measurement of aortic morphology for further analysis of aortic diseases, such as hypertension.

High-resolution structural-functional substrate-trigger characterization: Future roadmap for catheter ablation of ventricular tachycardia

Introduction

Patients with ventricular tachyarrhythmias (VT) are at high risk of sudden cardiac death. When appropriate, catheter ablation is modestly effective, with relatively high VT recurrence and complication rates. Personalized models that incorporate imaging and computational approaches have advanced VT management. However, 3D patient-specific functional electrical information is typically not considered. We hypothesize that incorporating non-invasive 3D electrical and structural characterization in a patient-specific model improves VT-substrate recognition and ablation targeting.

Materials and methods

In a 53-year-old male with ischemic cardiomyopathy and recurrent monomorphic VT, we built a structural-functional model based on high-resolution 3D late-gadolinium enhancement (LGE) cardiac magnetic resonance imaging (3D-LGE CMR), multi-detector computed tomography (CT), and electrocardiographic imaging (ECGI). Invasive data from high-density contact and pace mapping obtained during endocardial VT-substrate modification were also incorporated. The integrated 3D electro-anatomic model was analyzed off-line.

Results

Merging the invasive voltage maps and 3D-LGE CMR endocardial geometry led to a mean Euclidean node-to-node distance of 5 ± 2 mm. Inferolateral and apical areas of low bipolar voltage (<1.5 mV) were associated with high 3D-LGE CMR signal intensity (>0.4) and with higher transmurality of fibrosis. Areas of functional conduction delay or block (evoked delayed potentials, EDPs) were in close proximity to 3D-LGE CMR-derived heterogeneous tissue corridors. ECGI pinpointed the epicardial VT exit at ∼10 mm from the endocardial site of origin, both juxtaposed to the distal ends of two heterogeneous tissue corridors in the inferobasal left ventricle. Radiofrequency ablation at the entrances of these corridors, eliminating all EDPs, and at the VT site of origin rendered the patient non-inducible and arrhythmia-free until the present day (20 months follow-up). Off-line analysis in our model uncovered dynamic electrical instability of the LV inferolateral heterogeneous scar region which set the stage for an evolving VT circuit.

Discussion and conclusion

We developed a personalized 3D model that integrates high-resolution structural and electrical information and allows the investigation of their dynamic interaction during arrhythmia formation. This model enhances our mechanistic understanding of scar-related VT and provides an advanced, non-invasive roadmap for catheter ablation.

Weakly supervised inference of personalized heart meshes based on echocardiography videos

Echocardiography provides recordings of the heart chamber size and function and is a central tool for non-invasive diagnosis of heart diseases. It produces high-dimensional video data with substantial stochasticity in the measurements, which frequently prove difficult to interpret. To address this challenge, we propose an automated framework to enable the inference of a high resolution personalized 4D (3D plus time) surface mesh of the cardiac structures from 2D echocardiography video data. Inferring such shape models arises as a key step towards accurate personalized simulation that enables an automated assessment of the cardiac chamber morphology and function. The proposed method is trained using only unpaired echocardiography and heart mesh videos to find a mapping between these distinct visual domains in a self-supervised manner. The resulting model produces personalized 4D heart meshes, which exhibit a high correspondence with the input echocardiography videos. Furthermore, the 4D heart meshes enable the automatic extraction of echocardiographic variables, such as ejection fraction, myocardial muscle mass, and volumetric changes of chamber volumes over time with high temporal resolution.

Learned iterative segmentation of highly variable anatomy from limited data: Applications to whole heart segmentation for congenital heart disease

Training deep learning models that segment an image in one step typically requires a large collection of manually annotated images that captures the anatomical variability in a cohort. This poses challenges when anatomical variability is extreme but training data is limited, as when segmenting cardiac structures in patients with congenital heart disease (CHD). In this paper, we propose an iterative segmentation model and show that it can be accurately learned from a small dataset. Implemented as a recurrent neural network, the model evolves a segmentation over multiple steps, from a single user click until reaching an automatically determined stopping point. We develop a novel loss function that evaluates the entire sequence of output segmentations, and use it to learn model parameters. Segmentations evolve predictably according to growth dynamics encapsulated by training data, which consists of images, partially completed segmentations, and the recommended next step. The user can easily refine the final segmentation by examining those that are earlier or later in the output sequence. Using a dataset of 3D cardiac MR scans from patients with a wide range of CHD types, we show that our iterative model offers better generalization to patients with the most severe heart malformations.

Assessing the Accuracy of an Artificial Intelligence-Based Segmentation Algorithm for the Thoracic Aorta in Computed Tomography Applications

The aim was to evaluate the accuracy of a prototypical artificial intelligence-based algorithm for automated segmentation and diameter measurement of the thoracic aorta (TA) using CT. One hundred twenty-two patients who underwent dual-source CT were retrospectively included. Ninety-three of these patients had been administered intravenous iodinated contrast. Images were evaluated using the prototypical algorithm, which segments the TA and determines the corresponding diameters at predefined anatomical locations based on the American Heart Association guidelines. The reference standard was established by two radiologists individually in a blinded, randomized fashion. Equivalency was tested and inter-reader agreement was assessed using intra-class correlation (ICC). In total, 99.2% of the parameters measured by the prototype were assessable. In nine patients, the prototype failed to determine one diameter along the vessel. Measurements along the TA did not differ between the algorithm and readers (p > 0.05), establishing equivalence. Inter-reader agreement between the algorithm and readers (ICC ≥ 0.961; 95% CI: 0.940−0.974), and between the readers was excellent (ICC ≥ 0.879; 95% CI: 0.818−0.92). The evaluated prototypical AI-based algorithm accurately measured TA diameters at each region of interest independent of the use of either contrast utilization or pathology. This indicates that the prototypical algorithm has substantial potential as a valuable tool in the rapid clinical evaluation of aortic pathology.

Scale-Resolving Simulations of Steady and Pulsatile Flow Through Healthy and Stenotic Heart Valves

Blood-flow downstream of stenotic and healthy aortic valves exhibits intermittent random fluctuations in the velocity field which are associated with turbulence. Such flows warrant the use of computationally demanding scale-resolving models. The aim of this work was to compute and quantify this turbulent flow in healthy and stenotic heart valves for steady and pulsatile flow conditions. Large eddy simulations (LESs) and Reynolds-averaged Navier-Stokes (RANS) simulations were used to compute the flow field at inlet Reynolds numbers of 2700 and 5400 for valves with an opening area of 70 mm2 and 175 mm2 and their projected orifice-plate type counterparts. Power spectra and turbulent kinetic energy were quantified on the centerline. Projected geometries exhibited an increased pressure-drop (>90%) and elevated turbulent kinetic energy levels (>147%). Turbulence production was an order of magnitude higher in stenotic heart valves compared to healthy valves. Pulsatile flow stabilizes flow in the acceleration phase, whereas onset of deceleration triggered (healthy valve) or amplified (stenotic valve) turbulence. Simplification of the aortic valve by projecting the orifice area should be avoided in computational fluid dynamics (CFD). RANS simulations may be used to predict the transvalvular pressure-drop, but scale-resolving models are recommended when detailed information of the flow field is required.

A deep-learning approach for direct whole-heart mesh reconstruction

Automated construction of surface geometries of cardiac structures from volumetric medical images is important for a number of clinical applications. While deep-learning-based approaches have demonstrated promising reconstruction precision, these approaches have mostly focused on voxel-wise segmentation followed by surface reconstruction and post-processing techniques. However, such approaches suffer from a number of limitations including disconnected regions or incorrect surface topology due to erroneous segmentation and stair-case artifacts due to limited segmentation resolution. We propose a novel deep-learning-based approach that directly predicts whole heart surface meshes from volumetric CT and MR image data. Our approach leverages a graph convolutional neural network to predict deformation on mesh vertices from a pre-defined mesh template to reconstruct multiple anatomical structures in a 3D image volume. Our method demonstrated promising performance of generating whole heart reconstructions with as good or better accuracy than prior deep-learning-based methods on both CT and MR data. Furthermore, by deforming a template mesh, our method can generate whole heart geometries with better anatomical consistency and produce high-resolution geometries from lower resolution input image data. Our method was also able to produce temporally-consistent surface mesh predictions for heart motion from CT or MR cine sequences, and therefore can potentially be applied for efficiently constructing 4D whole heart dynamics. Our code and pre-trained networks are available at https://github.com/fkong7/MeshDeformNet.

The impact of shape uncertainty on aortic-valve pressure-drop computations

Patient-specific image-based computational fluid dynamics (CFD) is widely adopted in the cardiovascular research community to study hemodynamics, and will become increasingly important for personalized medicine. However, segmentation of the flow domain is not exact and geometric uncertainty can be expected which propagates through the computational model, leading to uncertainty in model output. Seventy-four aortic-valves were segmented from computed tomography images at peak systole. Statistical shape modeling was used to obtain an approximate parameterization of the original segmentations. This parameterization was used to train a meta-model that related the first five shape mode coefficients and flowrate to the CFD-computed transvalvular pressure-drop. Consequently, shape uncertainty in the order of 0.5 and 1.0 mm was emulated by introducing uncertainty in the shape mode coefficients. A global variance-based sensitivity analysis was performed to quantify output uncertainty and to determine relative importance of the shape modes. The first shape mode captured the opening/closing behavior of the valve and uncertainty in this mode coefficient accounted for more than 90% of the output variance. However, sensitivity to shape uncertainty is patient-specific, and the relative importance of the fourth shape mode coefficient tended to increase with increases in valvular area. These results show that geometric uncertainty in the order of image voxel size may lead to substantial uncertainty in CFD-computed transvalvular pressure-drops. Moreover, this illustrates that it is essential to assess the impact of geometric uncertainty on model output, and that this should be thoroughly quantified for applications that wish to use image-based CFD models.

Automated segmentation and quantification of the healthy and diseased aorta in CT angiographies using a dedicated deep learning approach

OBJECTIVES

To develop and validate a deep learning-based algorithm for segmenting and quantifying the physiological and diseased aorta in computed tomography angiographies.

METHODS

CTA exams of the aorta of 191 patients (68.1 ± 14 years, 128 male), performed between 2015 and 2018, were retrospectively identified from our imaging archive and manually segmented by two investigators. A 3D U-Net model was trained on the data, which was divided into a training, a validation, and a test group at a ratio of 7:1:2. Cases in the test group (n = 41) were evaluated to compare manual and automatic segmentations. Dice similarity coefficient (DSC), mean surface distance (MSD), and Hausdorff surface distance (HSD) were extracted. Maximum diameter, effective diameter, and area were quantified and compared between both segmentations at eight anatomical landmarks, and at the maximum area of an aneurysms if present (n = 14). Statistics included error calculation, intraclass correlation coefficient, and Bland-Altman analysis.

RESULTS

A DSC of 0.95 [0.94; 0.95] and an MSD of 0.76 [0.06; 0.99] indicated close agreement between segmentations. HSD was 8.00 [4.47; 10.00]. The largest absolute errors were found in the ascending aorta with 0.8 ± 1.5 mm for maximum diameter and at the coeliac trunk with - 30.0 ± 81.6 mm² for area. Results for absolute errors in aneurysms were - 0.5 ± 2.3 mm for maximum diameter, 0.3 ± 1.6 mm for effective diameter, and 64.9 ± 114.9 mm² for area. ICC showed excellent agreement (> 0.9; p < 0.05) between quantitative measurements.

CONCLUSIONS

Automated segmentation of the aorta on CTA data using a deep learning algorithm is feasible and allows for accurate quantification of the aortic lumen even if the vascular architecture is altered by disease.

KEY POINTS

• A deep learning-based algorithm can automatically segment the aorta, mostly within acceptable margins of error, even if the vascular architecture is altered by disease. • Quantifications performed in the segmentations were mostly within clinically acceptable limits, even in pathologically altered segments of the aorta.

Automated cardiac volume assessment and cardiac long- and short-axis imaging plane prediction from electrocardiogram-gated computed tomography volumes enabled by deep learning

AIMS

To develop an automated method for bloodpool segmentation and imaging plane re-slicing of cardiac computed tomography (CT) via deep learning (DL) for clinical use in coronary artery disease (CAD) wall motion assessment and reproducible longitudinal imaging.

METHODS AND RESULTS

One hundred patients who underwent clinically indicated cardiac CT scans with manually segmented left ventricle (LV) and left atrial (LA) chambers were used for training. For each patient, long-axis (LAX) and short-axis planes were manually defined by an imaging expert. A DL model was trained to predict bloodpool segmentations and imaging planes. Deep learning bloodpool segmentations showed close agreement with manual LV [median Dice: 0.91, Hausdorff distance (HD): 6.18 mm] and LA (Dice: 0.93, HD: 7.35 mm) segmentations and a strong correlation with manual ejection fraction (Pearson r: 0.95 LV, 0.92 LA). Predicted planes had low median location (6.96 mm) and angular orientation (7.96 ° ) errors which were comparable to inter-reader differences (P > 0.71). 84-97% of DL-prescribed LAX planes correctly intersected American Heart Association segments, which was comparable (P > 0.05) to manual slicing. In a test cohort of 144 patients, we evaluated the ability of the DL approach to provide diagnostic imaging planes. Visual scoring by two blinded experts determined ≥94% of DL-predicted planes to be diagnostically adequate. Further, DL-enabled visualization of LV wall motion abnormalities due to CAD and provided reproducible planes upon repeat imaging.

CONCLUSION

A volumetric, DL approach provides multiple chamber segmentations and can re-slice the imaging volume along standardized cardiac imaging planes for reproducible wall motion abnormality and functional assessment.

Tissue clearing and imaging methods for cardiovascular development

Tissue imaging in 3D using visible light is limited and various clearing techniques were developed to increase imaging depth, but none provides universal solution for all tissues at all developmental stages. In this review, we focus on different tissue clearing methods for 3D imaging of heart and vasculature, based on chemical composition (solvent-based, simple immersion, hyperhydration, and hydrogel embedding techniques). We discuss in detail compatibility of various tissue clearing techniques with visualization methods: fluorescence preservation, immunohistochemistry, nuclear staining, and fluorescent dyes vascular perfusion. We also discuss myocardium visualization using autofluorescence, tissue shrinking, and expansion. Then we overview imaging methods used to study cardiovascular system and live imaging. We discuss heart and vessels segmentation methods and image analysis. The review covers the whole process of cardiovascular system 3D imaging, starting from tissue clearing and its compatibility with various visualization methods to the types of imaging methods and resulting image analysis.

Deep learning approach for the segmentation of aneurysmal ascending aorta

Diagnosis of ascending thoracic aortic aneurysm (ATAA) is based on the measurement of the maximum aortic diameter, but size is not a good predictor of the risk of adverse events. There is growing interest in the development of novel image-derived risk strategies to improve patient risk management towards a highly individualized level. In this study, the feasibility and efficacy of deep learning for the automatic segmentation of ATAAs was investigated using UNet, ENet, and ERFNet techniques. Specifically, CT angiography done on 72 patients with ATAAs and different valve morphology (i.e., tricuspid aortic valve, TAV, and bicuspid aortic valve, BAV) were semi-automatically segmented with Mimics software (Materialize NV, Leuven, Belgium), and then used for training of the tested deep learning models. The segmentation performance in terms of accuracy and time inference were compared using several parameters. All deep learning models reported a dice score higher than 88%, suggesting a good agreement between predicted and manual ATAA segmentation. We found that the ENet and UNet are more accurate than ERFNet, with the ENet much faster than UNet. This study demonstrated that deep learning models can rapidly segment and quantify the 3D geometry of ATAAs with high accuracy, thereby facilitating the expansion into clinical workflow of personalized approach to the management of patients with ATAAs.

Fully automated left atrium segmentation from anatomical cine long-axis MRI sequences using deep convolutional neural network with unscented Kalman filter

This study proposes a fully automated approach for the left atrial segmentation from routine cine long-axis cardiac magnetic resonance image sequences using deep convolutional neural networks and Bayesian filtering. The proposed approach consists of a classification network that automatically detects the type of long-axis sequence and three different convolutional neural network models followed by unscented Kalman filtering (UKF) that delineates the left atrium. Instead of training and predicting all long-axis sequence types together, the proposed approach first identifies the image sequence type as to 2, 3 and 4 chamber views, and then performs prediction based on neural nets trained for that particular sequence type. The datasets were acquired retrospectively and ground truth manual segmentation was provided by an expert radiologist. In addition to neural net based classification and segmentation, another neural net is trained and utilized to select image sequences for further processing using UKF to impose temporal consistency over cardiac cycle. A cyclic dynamic model with time-varying angular frequency is introduced in UKF to characterize the variations in cardiac motion during image scanning. The proposed approach was trained and evaluated separately with varying amount of training data with images acquired from 20, 40, 60 and 80 patients. Evaluations over 1515 images with equal number of images from each chamber group acquired from an additional 20 patients demonstrated that the proposed model outperformed state-of-the-art and yielded a mean Dice coefficient value of 94.1%, 93.7% and 90.1% for 2, 3 and 4-chamber sequences, respectively, when trained with datasets from 80 patients.

Combining statistical shape modeling, CFD, and meta-modeling to approximate the patient-specific pressure-drop across the aortic valve in real-time

BACKGROUND

Advances in medical imaging, segmentation techniques, and high performance computing have stimulated the use of complex, patient-specific, three-dimensional Computational Fluid Dynamics (CFD) simulations. Patient-specific, CFD-compatible geometries of the aortic valve are readily obtained. CFD can then be used to obtain the patient-specific pressure-flow relationship of the aortic valve. However, such CFD simulations are computationally expensive, and real-time alternatives are desired.

AIM

The aim of this work is to evaluate the performance of a meta-model with respect to high-fidelity, three-dimensional CFD simulations of the aortic valve.

METHODS

Principal component analysis was used to build a statistical shape model (SSM) from a population of 74 iso-topological meshes of the aortic valve. Synthetic meshes were created with the SSM, and steady-state CFD simulations at flow-rates between 50 and 650 mL/s were performed to build a meta-model. The meta-model related the statistical shape variance, and flow-rate to the pressure-drop.

RESULTS

Even though the first three shape modes account for only 46% of shape variance, the features relevant for the pressure-drop seem to be captured. The three-mode shape-model approximates the pressure-drop with an average error of 8.8% to 10.6% for aortic valves with a geometric orifice area below 150 mm² . The proposed methodology was least accurate for aortic valve areas above 150 mm² . Further reduction to a meta-model introduces an additional 3% error.

CONCLUSIONS

Statistical shape modeling can be used to capture shape variation of the aortic valve. Meta-models trained by SSM-based CFD simulations can provide an estimate of the pressure-flow relationship in real-time.

Towards improving the accuracy of aortic transvalvular pressure gradients: rethinking Bernoulli

The transvalvular pressure gradient (TPG) is commonly estimated using the Bernoulli equation. However, the method is known to be inaccurate. Therefore, an adjusted Bernoulli model for accurate TPG assessment was developed and evaluated. Numerical simulations were used to calculate TPG_CFD in patient-specific geometries of aortic stenosis as ground truth. Geometries, aortic valve areas (AVA), and flow rates were derived from computed tomography scans. Simulations were divided in a training data set (135 cases) and a test data set (36 cases). The training data was used to fit an adjusted Bernoulli model as a function of AVA and flow rate. The model-predicted TPG_Model was evaluated using the test data set and also compared against the common Bernoulli equation (TPG_B). TPG_B and TPG_Model both correlated well with TPG_CFD (r > 0.94), but significantly overestimated it. The average difference between TPG_Model and TPG_CFD was much lower: 3.3 mmHg vs. 17.3 mmHg between TPG_B and TPG_CFD. Also, the standard error of estimate was lower for the adjusted model: SEE_Model = 5.3 mmHg vs. SEE_B = 22.3 mmHg. The adjusted model's performance was more accurate than that of the conventional Bernoulli equation. The model might help to improve non-invasive assessment of TPG. Graphical abstract Processing pipeline for the definition of an adjusted Bernoulli model for the assessment of transvalvular pressure gradient. Using CT image data, the patient specific geometry of the stenosed AVs were reconstructed. Using this segmentation, the AVA as well as the volume flow rate was calculated and used for model definition. This novel model was compared against classical approaches on a test data set, which was not used for the model definition.

Automated Segmentation of Left Ventricular Myocardium on Cardiac Computed Tomography Using Deep Learning

OBJECTIVE

To evaluate the accuracy of a deep learning-based automated segmentation of the left ventricle (LV) myocardium using cardiac CT.

MATERIALS AND METHODS

To develop a fully automated algorithm, 100 subjects with coronary artery disease were randomly selected as a development set (50 training / 20 validation / 30 internal test). An experienced cardiac radiologist generated the manual segmentation of the development set. The trained model was evaluated using 1000 validation set generated by an experienced technician. Visual assessment was performed to compare the manual and automatic segmentations. In a quantitative analysis, sensitivity and specificity were calculated according to the number of pixels where two three-dimensional masks of the manual and deep learning segmentations overlapped. Similarity indices, such as the Dice similarity coefficient (DSC), were used to evaluate the margin of each segmented masks.

RESULTS

The sensitivity and specificity of automated segmentation for each segment (1-16 segments) were high (85.5-100.0%). The DSC was 88.3 ± 6.2%. Among randomly selected 100 cases, all manual segmentation and deep learning masks for visual analysis were classified as very accurate to mostly accurate and there were no inaccurate cases (manual vs. deep learning: very accurate, 31 vs. 53; accurate, 64 vs. 39; mostly accurate, 15 vs. 8). The number of very accurate cases for deep learning masks was greater than that for manually segmented masks.

CONCLUSION

We present deep learning-based automatic segmentation of the LV myocardium and the results are comparable to manual segmentation data with high sensitivity, specificity, and high similarity scores.

Automatic segmentation of multiple cardiovascular structures from cardiac computed tomography angiography images using deep learning

OBJECTIVES

To develop, demonstrate and evaluate an automated deep learning method for multiple cardiovascular structure segmentation.

BACKGROUND

Segmentation of cardiovascular images is resource-intensive. We design an automated deep learning method for the segmentation of multiple structures from Coronary Computed Tomography Angiography (CCTA) images.

METHODS

Images from a multicenter registry of patients that underwent clinically-indicated CCTA were used. The proximal ascending and descending aorta (PAA, DA), superior and inferior vena cavae (SVC, IVC), pulmonary artery (PA), coronary sinus (CS), right ventricular wall (RVW) and left atrial wall (LAW) were annotated as ground truth. The U-net-derived deep learning model was trained, validated and tested in a 70:20:10 split.

RESULTS

The dataset comprised 206 patients, with 5.130 billion pixels. Mean age was 59.9 ± 9.4 yrs., and was 42.7% female. An overall median Dice score of 0.820 (0.782, 0.843) was achieved. Median Dice scores for PAA, DA, SVC, IVC, PA, CS, RVW and LAW were 0.969 (0.979, 0.988), 0.953 (0.955, 0.983), 0.937 (0.934, 0.965), 0.903 (0.897, 0.948), 0.775 (0.724, 0.925), 0.720 (0.642, 0.809), 0.685 (0.631, 0.761) and 0.625 (0.596, 0.749) respectively. Apart from the CS, there were no significant differences in performance between sexes or age groups.

CONCLUSIONS

An automated deep learning model demonstrated segmentation of multiple cardiovascular structures from CCTA images with reasonable overall accuracy when evaluated on a pixel level.

Increased Accuracy, Confidence, and Efficiency in Anterior Ethmoidal Artery Identification with Segmented Image Guidance

OBJECTIVE

To determine whether using image guidance technology with 3-dimensional image segmentation increases the endoscopic surgeon's accuracy, efficiency, and confidence in identifying the anterior ethmoidal artery.

METHODS

This is a cross-sectional study of attending physicians and residents at an academic medical center. Because identification of the anterior ethmoidal artery during image-guided surgery can be challenging, we studied the effect of anterior ethmoidal artery image segmentation (ie, partitioning and coloring) on surgeon test performance. A computerized test was administered to 16 surgeons who were asked to identify the anterior ethmoidal artery on multiplanar computed tomographic images and to answer multiple-choice questions. Half the questions showed segmented images of the anterior ethmoidal artery, and half showed images without segmentation. Efficiency and accuracy of identification and subjective surgeon confidence were determined for each question. Descriptive statistics were used to compare test performance for identification of the anterior ethmoidal artery on images with or without segmentation.

RESULTS

Percentage of correct answers ( P < .001), efficiency ( P < .001), and confidence ( P < .001) in identification of the anterior ethmoidal artery were significantly better with segmented computed tomographic images.

DISCUSSION

We demonstrated that use of segmented images improves surgeons' accuracy, confidence, and efficiency for identification of the anterior ethmoidal artery.

IMPLICATIONS FOR PRACTICE

We describe how segmentation can allow surgeons to improve the surgical course by increasing their accuracy, confidence, and efficiency in identifying the anterior ethmoidal artery.

Automated 3D segmentation and diameter measurement of the thoracic aorta on non-contrast enhanced CT

Objectives

To develop and evaluate a fully automatic method to measure diameters of the ascending and descending aorta on non-ECG-gated, non-contrast computed tomography (CT) scans.

Material and methods

The method combines multi-atlas registration to obtain seed points, aorta centerline extraction, and an optimal surface segmentation approach to extract the aorta surface around the centerline. From the extracted 3D aorta segmentation, the diameter of the ascending and descending aorta was calculated at cross-sectional slices perpendicular to the extracted centerline, at the level of the pulmonary artery bifurcation, and at 1-cm intervals up to 3 cm above and below this level. Agreement with manual annotations was evaluated by dice similarity coefficient (DSC) for segmentation overlap, mean surface distance (MSD), and intra-class correlation (ICC) of diameters on 100 CT scans from a lung cancer screening trial. Repeatability of the diameter measurements was evaluated on 617 baseline-one year follow-up CT scan pairs.

Results

The agreement between manual and automatic segmentations was good with 0.95 ± 0.01 DSC and 0.56 ± 0.08 mm MSD. ICC between the diameters derived from manual and from automatic segmentations was 0.97, with the per-level ICC ranging from 0.87 to 0.94. An ICC of 0.98 for all measurements and per-level ICC ranging from 0.91 to 0.96 were obtained for repeatability.

Conclusion

This fully automatic method can assess diameters in the thoracic aorta reliably even in non-ECG-gated, non-contrast CT scans. This could be a promising tool to assess aorta dilatation in screening and in clinical practice.

Key Points

• Fully automatic method to assess thoracic aorta diameters.

• High agreement between fully automatic method and manual segmentations.

• Method is suitable for non-ECG-gated CT and can therefore be used in screening.

Electronic supplementary material

The online version of this article (10.1007/s00330-018-5931-z) contains supplementary material, which is available to authorized users.

Automated segmentation of the atrial region and fossa ovalis towards computer-aided planning of inter-atrial wall interventions

BACKGROUND AND OBJECTIVE

Image-fusion strategies have been applied to improve inter-atrial septal (IAS) wall minimally-invasive interventions. Hereto, several landmarks are initially identified on richly-detailed datasets throughout the planning stage and then combined with intra-operative images, enhancing the relevant structures and easing the procedure. Nevertheless, such planning is still performed manually, which is time-consuming and not necessarily reproducible, hampering its regular application. In this article, we present a novel automatic strategy to segment the atrial region (left/right atrium and aortic tract) and the fossa ovalis (FO).

METHODS

The method starts by initializing multiple 3D contours based on an atlas-based approach with global transforms only and refining them to the desired anatomy using a competitive segmentation strategy. The obtained contours are then applied to estimate the FO by evaluating both IAS wall thickness and the expected FO spatial location.

RESULTS

The proposed method was evaluated in 41 computed tomography datasets, by comparing the atrial region segmentation and FO estimation results against manually delineated contours. The automatic segmentation method presented a performance similar to the state-of-the-art techniques and a high feasibility, failing only in the segmentation of one aortic tract and of one right atrium. The FO estimation method presented an acceptable result in all the patients with a performance comparable to the inter-observer variability. Moreover, it was faster and fully user-interaction free.

CONCLUSIONS

Hence, the proposed method proved to be feasible to automatically segment the anatomical models for the planning of IAS wall interventions, making it exceptionally attractive for use in the clinical practice.

Efficient workflow for automatic segmentation of the right heart based on 2D echocardiography

The present study aimed to present a workflow algorithm for automatic processing of 2D echocardiography images. The workflow was based on several sequential steps. For each step, we compared different approaches. Epicardial 2D echocardiography datasets were acquired during various open-chest beating-heart surgical procedures in three porcine hearts. We proposed a metric called the global index that is a weighted average of several accuracy coefficients, indices and the mean processing time. This metric allows the estimation of the speed and accuracy for processing each image. The global index ranges from 0 to 1, which facilitates comparison between different approaches. The second step involved comparison among filtering, sharpening and segmentation techniques. During the noise reduction step, we compared the median filter, total variation filter, bilateral filter, curvature flow filter, non-local means filter and mean shift filter. To clarify the endocardium borders of the right heart, we used the linear sharpen. Lastly, we applied watershed segmentation, clusterisation, region-growing, morphological segmentation, image foresting segmentation and isoline delineation. We assessed all the techniques and identified the most appropriate workflow for echocardiography image segmentation of the right heart. For successful processing and segmentation of echocardiography images with minimal error, we found that the workflow should include the total variation filter/bilateral filter, linear sharpen technique, isoline delineation/region-growing segmentation and morphological post-processing. We presented an efficient and accurate workflow for the precise diagnosis of cardiovascular diseases. We introduced the global index metric for image pre-processing and segmentation estimation.

A competitive strategy for atrial and aortic tract segmentation based on deformable models

Multiple strategies have previously been described for atrial region (i.e. atrial bodies and aortic tract) segmentation. Although these techniques have proven their accuracy, inadequate results in the mid atrial walls are common, restricting their application for specific cardiac interventions. In this work, we introduce a novel competitive strategy to perform atrial region segmentation with correct delineation of the thin mid walls, and integrated it into the B-spline Explicit Active Surfaces framework. A double-stage segmentation process is used, which starts with a fast contour growing followed by a refinement stage with local descriptors. Independent functions are used to define each region, being afterward combined to compete for the optimal boundary. The competition locally constrains the surface evolution, prevents overlaps and allows refinement to the walls. Three different scenarios were used to demonstrate the advantages of the proposed approach, through the evaluation of its segmentation accuracy, and its performance for heterogeneous mid walls. Both computed tomography and magnetic resonance imaging datasets were used, presenting results similar to the state-of-the-art methods for both atria and aorta. The competitive strategy showed its superior performance with statistically significant differences against the traditional free-evolution approach in cases with bad image quality or missed atrial/aortic walls. Moreover, only the competitive approach was able to accurately segment the atrial/aortic wall. Overall, the proposed strategy showed to be suitable for atrial region segmentation with a correct segmentation of the mid thin walls, demonstrating its added value with respect to the traditional techniques.

Automated detection of coarctation of aorta in neonates from two-dimensional echocardiograms

Coarctation of aorta (CoA) is a critical congenital heart defect (CCHD) that requires accurate and immediate diagnosis and treatment. Current newborn screening methods to detect CoA lack both in sensitivity and specificity, and when suspected in a newborn, it must be confirmed using specialized imaging and expert diagnosis, both of which are usually unavailable at tertiary birthing centers. We explore the feasibility of applying machine learning methods to reliably determine the presence of this difficult-to-diagnose cardiac abnormality from ultrasound image data. We propose a framework that uses deep learning-based machine learning methods for fully automated detection of CoA from two-dimensional ultrasound clinical data acquired in the parasternal long axis view, the apical four chamber view, and the suprasternal notch view. On a validation set consisting of 26 CoA and 64 normal patients our algorithm achieved a total error rate of 12.9% (11.5% false-negative error and 13.6% false-positive error) when combining decisions of classifiers over three standard echocardiographic view planes. This compares favorably with published results that combine clinical assessments with pulse oximetry to detect CoA (71% sensitivity).

Cardiac atlas development and validation for automatic segmentation of cardiac substructures

PURPOSE

To develop and validate a set of atlases for auto-contouring cardiac substructures.

METHODS

Eight radiation oncologists manually and independently delineated 15 cardiac substructures from noncontrast CT images of 6 patients by referring to their respective fused contrast CT images. Individual contours were fused together for each structure, edited by 2 physicians, and became atlases to delineate other 6 patients. The auto-delineated contours of the 6 additional patients became templates for manual contouring. These 12 patients with well-defined contours composed the final atlases for multi-atlas segmentation.

RESULTS

The average time for manually contouring the 15 cardiac substructures was about 40min. Inter-observer variability was small for the heart, the chambers, and the aorta compared with that for other structures that were not clearly distinguishable in CT images. The mean dice similarity coefficient and mean surface distance of auto-segmented contours were within one standard deviation of expert contouring variability. Good agreement between auto-segmented and manual contours was observed for the heart, the chambers, and the great vessels. Independent validation on other 19 patients showed reasonable agreement for the heart chambers.

CONCLUSIONS

A set of cardiac atlases was created for auto-contouring from noncontrast CT images. The accuracy of auto-contouring for the heart, chambers, and great vessels was validated for potential clinical use.

Optimal C-arm angulation during transcatheter aortic valve replacement: Accuracy of a rotational C-arm computed tomography based three dimensional heart model

AIM

To investigate the accuracy of a rotational C-arm CT-based 3D heart model to predict an optimal C-arm configuration during transcatheter aortic valve replacement (TAVR).

METHODS

Rotational C-arm CT (RCT) under rapid ventricular pacing was performed in 57 consecutive patients with severe aortic stenosis as part of the pre-procedural cardiac catheterization. With prototype software each RCT data set was segmented using a 3D heart model. From that the line of perpendicularity curve was obtained that generates a perpendicular view of the aortic annulus according to the right-cusp rule. To evaluate the accuracy of a model-based overlay we compared model- and expert-derived aortic root diameters.

RESULTS

For all 57 patients in the RCT cohort diameter measurements were obtained from two independent operators and were compared to the model-based measurements. The inter-observer variability was measured to be in the range of 0°-12.96° of angular C-arm displacement for two independent operators. The model-to-operator agreement was 0°-13.82°. The model-based and expert measurements of aortic root diameters evaluated at the aortic annulus (r = 0.79, P < 0.01), the aortic sinus (r = 0.93, P < 0.01) and the sino-tubular junction (r = 0.92, P < 0.01) correlated on a high level and the Bland-Altman analysis showed good agreement. The interobserver measurements did not show a significant bias.

CONCLUSION

Automatic segmentation of the aortic root using an anatomical model can accurately predict an optimal C-arm configuration, potentially simplifying current clinical workflows before and during TAVR.

Four challenges in medical image analysis from an industrial perspective

Today's medical imaging systems produce a huge amount of images containing a wealth of information. However, the information is hidden in the data and image analysis algorithms are needed to extract it, to make it readily available for medical decisions and to enable an efficient work flow. Advances in medical image analysis over the past 20 years mean there are now many algorithms and ideas available that allow to address medical image analysis tasks in commercial solutions with sufficient performance in terms of accuracy, reliability and speed. At the same time new challenges have arisen. Firstly, there is a need for more generic image analysis technologies that can be efficiently adapted for a specific clinical task. Secondly, efficient approaches for ground truth generation are needed to match the increasing demands regarding validation and machine learning. Thirdly, algorithms for analyzing heterogeneous image data are needed. Finally, anatomical and organ models play a crucial role in many applications, and algorithms to construct patient-specific models from medical images with a minimum of user interaction are needed. These challenges are complementary to the on-going need for more accurate, more reliable and faster algorithms, and dedicated algorithmic solutions for specific applications.

Multiphysics and multiscale modelling, data-model fusion and integration of organ physiology in the clinic: ventricular cardiac mechanics

With heart and cardiovascular diseases continually challenging healthcare systems worldwide, translating basic research on cardiac (patho)physiology into clinical care is essential. Exacerbating this already extensive challenge is the complexity of the heart, relying on its hierarchical structure and function to maintain cardiovascular flow. Computational modelling has been proposed and actively pursued as a tool for accelerating research and translation. Allowing exploration of the relationships between physics, multiscale mechanisms and function, computational modelling provides a platform for improving our understanding of the heart. Further integration of experimental and clinical data through data assimilation and parameter estimation techniques is bringing computational models closer to use in routine clinical practice. This article reviews developments in computational cardiac modelling and how their integration with medical imaging data is providing new pathways for translational cardiac modelling.

Pulmonary hypertension and right ventricular dysfunction in patients with left to right shunt coronary artery fistula: evaluation with cardiac CT

In this study, we aimed to evaluate whether patients with left to right shunt coronary artery fistula (LRSCAF) are predisposed to developing pulmonary hypertension and right ventricular dysfunction compared with healthy individuals. The value of cardiac CT findings in determining the necessity of intervention for these patients was investigated. We retrospectively studied 19 patients with LRSCAF and 19 healthy patients. Several parameters were observed on cardiac CT by two radiologists, including pulmonary trunk diameter (PA diameter), right ventricular diameter (RVD), left ventricular diameter (LVD), RVD/LVD ratio, septal bowing and CT score of right ventricular dysfunction (CSRVD). Data from both groups were compared. The inter- and intra-observer variabilities and correlations were examined. The disease group was further divided into intervention (n = 9) and non-intervention (n = 10) groups, and their data were compared. All cardiac CT findings showed significant intra- and inter-observer correlation without significant variability. Mann-Whitney U tests and χ(2) analysis showed that PA diameter, RVD/LVD ratio acquired from two observers, and CSRVD were higher in the disease group than in the control group (all P values < 0.05 for χ(2) and almost all P values < 0.05 for Mann-Whitney U). The RVD/LVD ratio and CSRVD were higher in the intervention group than in the non-intervention group (all P values < 0.05). Receiver operating curve analysis identified RVD/LVD = 1.036 and CSRVD = 3.5 as the best cut-off values to determine the necessity of further intervention. Patients with LRSCAF are more predisposed to pulmonary hypertension and right ventricular dysfunction compared with the normal population. RVD/LVD > 1.0 and CSRVD ≥ 4.0 may determine the necessity of intervention for patients with LRSCAF.

Left-Atrial Segmentation From 3-D Ultrasound Using B-Spline Explicit Active Surfaces With Scale Uncoupling

Segmentation of the left atrium (LA) of the heart allows quantification of LA volume dynamics which can give insight into cardiac function. However, very little attention has been given to LA segmentation from three-dimensional (3-D) ultrasound (US), most efforts being focused on the segmentation of the left ventricle (LV). The B-spline explicit active surfaces (BEAS) framework has been shown to be a very robust and efficient methodology to perform LV segmentation. In this study, we propose an extension of the BEAS framework, introducing B-splines with uncoupled scaling. This formulation improves the shape support for less regular and more variable structures, by giving independent control over smoothness and number of control points. Semiautomatic segmentation of the LA endocardium using this framework was tested in a setup requiring little user input, on 20 volumetric sequences of echocardiographic data from healthy subjects. The segmentation results were evaluated against manual reference delineations of the LA. Relevant LA morphological and functional parameters were derived from the segmented surfaces, in order to assess the performance of the proposed method on its clinical usage. The results showed that the modified BEAS framework is capable of accurate semiautomatic LA segmentation in 3-D transthoracic US, providing reliable quantification of the LA morphology and function.

Multi-modal data fusion for Cardiac Resynchronization Therapy planning and assistance

Cardiac Resynchronization Therapy (CRT) has been validated as an efficient treatment for selected patients suffering from heart failure with cardiac dyssynchrony. In case of bi-ventricular stimulation, the response to the therapy may be improved by an optimal choice of the left ventricle (LV) pacing sites. The characterization of LV properties to select the best candidate sites and to precise their access modes would be useful for the clinician in pre- and per-operative stages. For that purpose, we propose a new pre-operative analysis solution integrating previously developed multi-modal data registration methods and a new segmentation process of their coronary venous access. Moreover, a novel visualization interface is proposed to help the clinician to visualize the most relevant pacing sites and their access during the implantation in the operating room. This work is illustrated on real CRT data patients.

Cardiac Multi-detector CT Segmentation Based on Multiscale Directional Edge Detector and 3D Level Set

Extraction of the cardiac surfaces of interest from multi-detector computed tomographic (MDCT) data is a pre-requisite step for cardiac analysis, as well as for image guidance procedures. Most of the existing methods need manual corrections, which is time-consuming. We present a fully automatic segmentation technique for the extraction of the right ventricle, left ventricular endocardium and epicardium from MDCT images. The method consists in a 3D level set surface evolution approach coupled to a new stopping function based on a multiscale directional second derivative Gaussian filter, which is able to stop propagation precisely on the real boundary of the structures of interest. We validated the segmentation method on 18 MDCT volumes from healthy and pathologic subjects using manual segmentation performed by a team of expert radiologists as gold standard. Segmentation errors were assessed for each structure resulting in a surface-to-surface mean error below 0.5 mm and a percentage of surface distance with errors less than 1 mm above 80%. Moreover, in comparison to other segmentation approaches, already proposed in previous work, our method presented an improved accuracy (with surface distance errors less than 1 mm increased of 8-20% for all structures). The obtained results suggest that our approach is accurate and effective for the segmentation of ventricular cavities and myocardium from MDCT images.