Four-chamber heart modeling and automatic segmentation for 3-D cardiac CT volumes using marginal space learning and steerable features

Clinical Validation of a Deep Learning Algorithm for Automated Coronary Artery Disease Detection and Classification Using a Heterogeneous Multivendor Coronary Computed Tomography Angiography Data Set

PURPOSE

We sought to clinically validate a fully automated deep learning (DL) algorithm for coronary artery disease (CAD) detection and classification in a heterogeneous multivendor cardiac computed tomography angiography data set.

MATERIALS AND METHODS

In this single-centre retrospective study, we included patients who underwent cardiac computed tomography angiography scans between 2010 and 2020 with scanners from 4 vendors (Siemens Healthineers, Philips, General Electrics, and Canon). Coronary Artery Disease-Reporting and Data System (CAD-RADS) classification was performed by a DL algorithm and by an expert reader (reader 1, R1), the gold standard. Variability analysis was performed with a second reader (reader 2, R2) and the radiologic reports on a subset of cases. Statistical analysis was performed stratifying patients according to the presence of CAD (CAD-RADS >0) and obstructive CAD (CAD-RADS ≥3).

RESULTS

Two hundred ninety-six patients (average age: 53.66 ± 13.65, 169 males) were enrolled. For the detection of CAD only, the DL algorithm showed sensitivity, specificity, accuracy, and area under the curve of 95.3%, 79.7%, 87.5%, and 87.5%, respectively. For the detection of obstructive CAD, the DL algorithm showed sensitivity, specificity, accuracy, and area under the curve of 89.4%, 92.8%, 92.2%, and 91.1%, respectively. The variability analysis for the detection of obstructive CAD showed an accuracy of 92.5% comparing the DL algorithm with R1, and 96.2% comparing R1 with R2 and radiology reports. The time of analysis was lower using the DL algorithm compared with R1 ( P < 0.001).

CONCLUSIONS

The DL algorithm demonstrated robust performance and excellent agreement with the expert readers' analysis for the evaluation of CAD, which also corresponded with significantly reduced image analysis time.

Segmentation of the left atrial appendage based on fusion attention

In clinical practice, the morphology of the left atrial appendage (LAA) plays an important role in the selection of LAA closure devices for LAA closure procedures. The morphology determination is influenced by the segmentation results. The LAA occupies only a small part of the entire 3D medical image, and the segmentation results are more likely to be biased towards the background region, making the segmentation of the LAA challenging. In this paper, we propose a lightweight attention mechanism called fusion attention, which imitates human visual behavior. We process the 3D image of the LAA using a method that involves overview observation followed by detailed observation. In the overview observation stage, the image features are pooled along the three dimensions of length, width, and height. The obtained features from the three dimensions are then separately input into the spatial attention and channel attention modules to learn the regions of interest. In the detailed observation stage, the attention results from the previous stage are fused using element-wise multiplication and combined with the original feature map to enhance feature learning. The fusion attention mechanism was evaluated on a left atrial appendage dataset provided by Liaoning Provincial People's Hospital, resulting in an average Dice coefficient of 0.8855. The results indicate that the fusion attention mechanism achieves better segmentation results on 3D images compared to existing lightweight attention mechanisms.

Revolutionizing Cardiac Imaging: A Scoping Review of Artificial Intelligence in Echocardiography, CTA, and Cardiac MRI

BACKGROUND AND INTRODUCTION

Cardiac imaging is crucial for diagnosing heart disorders. Methods like X-rays, ultrasounds, CT scans, and MRIs provide detailed anatomical and functional heart images. AI can enhance these imaging techniques with its advanced learning capabilities.

METHOD

In this scoping review, following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) Guidelines, we searched PubMed, Scopus, Web of Science, and Google Scholar using related keywords on 16 April 2024. From 3679 articles, we first screened titles and abstracts based on the initial inclusion criteria and then screened the full texts. The authors made the final selections collaboratively.

RESULT

The PRISMA chart shows that 3516 articles were initially selected for evaluation after removing duplicates. Upon reviewing titles, abstracts, and quality, 24 articles were deemed eligible for the review. The findings indicate that AI enhances image quality, speeds up imaging processes, and reduces radiation exposure with sensitivity and specificity comparable to or exceeding those of qualified radiologists or cardiologists. Further research is needed to assess AI's applicability in various types of cardiac imaging, especially in rural hospitals where access to medical doctors is limited.

CONCLUSIONS

AI improves image quality, reduces human errors and radiation exposure, and can predict cardiac events with acceptable sensitivity and specificity.

Cardiac MRI segmentation using shifted-window multilayer perceptron mixer networks

Objectives. In this work, we proposed a deep-learning segmentation algorithm for cardiac magnetic resonance imaging to aid in contouring of the left ventricle, right ventricle, and Myocardium (Myo).Approach.We proposed a shifted window multilayer perceptron (Swin-MLP) mixer network which is built upon a 3D U-shaped symmetric encoder-decoder structure. We evaluated our proposed network using public data from 100 individuals. The network performance was quantitatively evaluated using 3D volume similarity between the ground truth contours and the predictions using Dice score coefficient, sensitivity, and precision as well as 2D surface similarity using Hausdorff distance (HD), mean surface distance (MSD) and residual mean square distance (RMSD). We benchmarked the performance against two other current leading edge networks known as Dynamic UNet and Swin-UNetr on the same public dataset.Results.The proposed network achieved the following volume similarity metrics when averaged over three cardiac segments: Dice = 0.952 ± 0.017, precision = 0.948 ± 0.016, sensitivity = 0.956 ± 0.022. The average surface similarities were HD = 1.521 ± 0.121 mm, MSD = 0.266 ± 0.075 mm, and RMSD = 0.668 ± 0.288 mm. The network shows statistically significant improvement in comparison to the Dynamic UNet and Swin-UNetr algorithms for most volumetric and surface metrics withp-value less than 0.05. Overall, the proposed Swin-MLP mixer network demonstrates better or comparable performance than competing methods.Significance.The proposed Swin-MLP mixer network demonstrates more accurate segmentation performance compared to current leading edge methods. This robust method demonstrates the potential to streamline clinical workflows for multiple applications.

Physical and Computational Modeling for Transcatheter Structural Heart Interventions

Structural heart disease interventions rely heavily on preprocedural planning and simulation to improve procedural outcomes and predict and prevent potential procedural complications. Modeling technologies, namely 3-dimensional (3D) printing and computational modeling, are nowadays increasingly used to predict the interaction between cardiac anatomy and implantable devices. Such models play a role in patient education, operator training, procedural simulation, and appropriate device selection. However, current modeling is often limited by the replication of a single static configuration within a dynamic cardiac cycle. Recognizing that health systems may face technical and economic limitations to the creation of "in-house" 3D-printed models, structural heart teams are pivoting to the use of computational software for modeling purposes.

Deep Learning-Based Automated Labeling of Coronary Segments for Structured Reporting of Coronary Computed Tomography Angiography in Accordance With Society of Cardiovascular Computed Tomography Guidelines

PURPOSE

To evaluate a novel deep learning (DL)-based automated coronary labeling approach for structured reporting of coronary artery disease according to the guidelines of the Society of Cardiovascular Computed Tomography (CT) on coronary CT angiography (CCTA).

PATIENTS AND METHODS

A retrospective cohort of 104 patients (60.3 ± 10.7 y, 61% males) who had undergone prospectively electrocardiogram-synchronized CCTA were included. Coronary centerlines were automatically extracted, labeled, and validated by 2 expert readers according to Society of Cardiovascular CT guidelines. The DL algorithm was trained on 706 radiologist-annotated cases for the task of automatically labeling coronary artery centerlines. The architecture leverages tree-structured long short-term memory recurrent neural networks to capture the full topological information of the coronary trees by using a two-step approach: a bottom-up encoding step, followed by a top-down decoding step. The first module encodes each sub-tree into fixed-sized vector representations. The decoding module then selectively attends to the aggregated global context to perform the local assignation of labels. To assess the performance of the software, percentage overlap was calculated between the labels of the algorithm and the expert readers.

RESULTS

A total number of 1491 segments were identified. The artificial intelligence-based software approach yielded an average overlap of 94.4% compared with the expert readers' labels ranging from 87.1% for the posterior descending artery of the right coronary artery to 100% for the proximal segment of the right coronary artery. The average computational time was 0.5 seconds per case. The interreader overlap was 96.6%.

CONCLUSIONS

The presented fully automated DL-based coronary artery labeling algorithm provides fast and precise labeling of the coronary artery segments bearing the potential to improve automated structured reporting for CCTA.

A Semi-automatic Pipeline for Generation of Large Cohorts of Four-Chamber Heart Meshes

Computational models for cardiac electro-mechanics have been increasingly used to further understand heart function. Small cohort and single patient computational studies provide useful insight into cardiac pathophysiology and response to therapy. However, these smaller studies have limited capability to capture the high level of anatomical variability seen in cardiology patients. Larger cohort studies are, on the other hand, more representative of the study population, but building several patient-specific anatomical meshes can be time-consuming and requires access to larger datasets of imaging data, image processing software to label anatomical structures and tools to create high fidelity anatomical meshes. Limited access to these tools and data might limit advances in this area of research. In this chapter, we present our semi-automatic pipeline to build patient-specific four-chamber heart meshes from CT imaging datasets, including ventricular myofibers and a set of universal ventricular and atrial coordinates. This pipeline was applied to CT images from both heart failure patients and healthy controls to generate cohorts of tetrahedral meshes suitable for electro-mechanics simulations. Both cohorts were made publicly available in order to promote computational studies employing large virtual cohorts.

S-Net: a multiple cross aggregation convolutional architecture for automatic segmentation of small/thin structures for cardiovascular applications

With the success of U-Net or its variants in automatic medical image segmentation, building a fully convolutional network (FCN) based on an encoder-decoder structure has become an effective end-to-end learning approach. However, the intrinsic property of FCNs is that as the encoder deepens, higher-level features are learned, and the receptive field size of the network increases, which results in unsatisfactory performance for detecting low-level small/thin structures such as atrial walls and small arteries. To address this issue, we propose to keep the different encoding layer features at their original sizes to constrain the receptive field from increasing as the network goes deeper. Accordingly, we develop a novel S-shaped multiple cross-aggregation segmentation architecture named S-Net, which has two branches in the encoding stage, i.e., a resampling branch to capture low-level fine-grained details and thin/small structures and a downsampling branch to learn high-level discriminative knowledge. In particular, these two branches learn complementary features by residual cross-aggregation; the fusion of the complementary features from different decoding layers can be effectively accomplished through lateral connections. Meanwhile, we perform supervised prediction at all decoding layers to incorporate coarse-level features with high semantic meaning and fine-level features with high localization capability to detect multi-scale structures, especially for small/thin volumes fully. To validate the effectiveness of our S-Net, we conducted extensive experiments on the segmentation of cardiac wall and intracranial aneurysm (IA) vasculature, and quantitative and qualitative evaluations demonstrated the superior performance of our method for predicting small/thin structures in medical images.

Artificial intelligence in cardiac computed tomography

Artificial Intelligence (AI) is a broad discipline of computer science and engineering. Modern application of AI encompasses intelligent models and algorithms for automated data analysis and processing, data generation, and prediction with applications in visual perception, speech understanding, and language translation. AI in healthcare uses machine learning (ML) and other predictive analytical techniques to help sort through vast amounts of data and generate outputs that aid in diagnosis, clinical decision support, workflow automation, and prognostication. Coronary computed tomography angiography (CCTA) is an ideal union for these applications due to vast amounts of data generation and analysis during cardiac segmentation, coronary calcium scoring, plaque quantification, adipose tissue quantification, peri-operative planning, fractional flow reserve quantification, and cardiac event prediction. In the past 5 years, there has been an exponential increase in the number of studies exploring the use of AI for cardiac computed tomography (CT) image acquisition, de-noising, analysis, and prognosis. Beyond image processing, AI has also been applied to improve the imaging workflow in areas such as patient scheduling, urgent result notification, report generation, and report communication. In this review, we discuss algorithms applicable to AI and radiomic analysis; we then present a summary of current and emerging clinical applications of AI in cardiac CT. We conclude with AI's advantages and limitations in this new field.

Aortic Annulus Detection Based on Deep Learning for Transcatheter Aortic Valve Replacement Using Cardiac Computed Tomography

BACKGROUND

To propose a deep learning architecture for automatically detecting the complex structure of the aortic annulus plane using cardiac computed tomography (CT) for transcatheter aortic valve replacement (TAVR).

METHODS

This study retrospectively reviewed consecutive patients who underwent TAVR between January 2017 and July 2020 at a tertiary medical center. Annulus Detection Permuted AdaIN network (ADPANet) based on a three-dimensional (3D) U-net architecture was developed to detect and localize the aortic annulus plane using cardiac CT. Patients (N = 72) who underwent TAVR between January 2017 and July 2020 at a tertiary medical center were enrolled. Ground truth using a limited dataset was delineated manually by three cardiac radiologists. Training, tuning, and testing sets (70:10:20) were used to build the deep learning model. The performance of ADPANet for detecting the aortic annulus plane was analyzed using the root mean square error (RMSE) and dice similarity coefficient (DSC).

RESULTS

In this study, the total dataset consisted of 72 selected scans from patients who underwent TAVR. The RMSE and DSC values for the aortic annulus plane using ADPANet were 55.078 ± 35.794 and 0.496 ± 0.217, respectively.

CONCLUSION

Our deep learning framework was feasible to detect the 3D complex structure of the aortic annulus plane using cardiac CT for TAVR. The performance of our algorithms was higher than other convolutional neural networks.

AI-based, automated chamber volumetry from gated, non-contrast CT

BACKGROUND

Accurate chamber volumetry from gated, non-contrast cardiac CT (NCCT) scans can be useful for potential screening of heart failure.

OBJECTIVES

To validate a new, fully automated, AI-based method for cardiac volume and myocardial mass quantification from NCCT scans compared to contrasted CT Angiography (CCTA).

METHODS

Of a retrospectively collected cohort of 1051 consecutive patients, 420 patients had both NCCT and CCTA scans at mid-diastolic phase, excluding patients with cardiac devices. Ground truth values were obtained from the CCTA scans.

RESULTS

The NCCT volume computation shows good agreement with ground truth values. Volume differences [95% CI ] and correlation coefficients were: -9.6 [-45; 26] mL, r ​= ​0.98 for LV Total, -5.4 [-24; 13] mL, r ​= ​0.95 for LA, -8.7 [-45; 28] mL, r ​= ​0.94 for RV, -5.2 [-27; 17] mL, r ​= ​0.92 for RA, -3.2 [-42; 36] mL, r ​= ​0.91 for LV blood pool, and -6.7 [-39; 26] g, r ​= ​0.94 for LV wall mass, respectively. Mean relative volume errors of less than 7% were obtained for all chambers.

CONCLUSIONS

Fully automated assessment of chamber volumes from NCCT scans is feasible and correlates well with volumes obtained from contrast study.

Automated Detection of Apical Foreshortening in Echocardiography Using Statistical Shape Modelling

OBJECTIVE

Automated detection of foreshortening, a common challenge in routine 2-D echocardiography, has the potential to improve quality of acquisitions and reduce the variability of left ventricular measurements. Acquiring and labelling the required training data is challenging due to the time-intensive and highly subjective nature of foreshortened apical views. We aimed to develop an automatic pipeline for the detection of foreshortening. To this end, we propose a method to generate synthetic apical-four-chamber (A4C) views with matching ground truth foreshortening labels.

METHODS

A statistical shape model of the four chambers of the heart was used to synthesise idealised A4C views with varying degrees of foreshortening. Contours of the left ventricular endocardium were segmented in the images, and a partial least squares (PLS) model was trained to learn the morphological traits of foreshortening. The predictive capability of the learned synthetic features was evaluated on an independent set of manually labelled and automatically curated real echocardiographic A4C images.

RESULTS

Acceptable classification accuracy for identification of foreshortened views in the testing set was achieved using logistic regression based on 11 PLS shape modes, with a sensitivity, specificity and area under the receiver operating characteristic curve of 0.84, 0.82 and 0.84, respectively. Both synthetic and real cohorts showed interpretable traits of foreshortening within the first two PLS shape modes, reflected as a shortening in the long-axis length and apical rounding.

CONCLUSION

A contour shape model trained only on synthesized A4C views allowed accurate prediction of foreshortening in real echocardiographic images.

Latest Developments in Adapting Deep Learning for Assessing TAVR Procedures and Outcomes

Aortic valve defects are among the most prevalent clinical conditions. A severely damaged or non-functioning aortic valve is commonly replaced with a bioprosthetic heart valve (BHV) via the transcatheter aortic valve replacement (TAVR) procedure. Accurate pre-operative planning is crucial for a successful TAVR outcome. Assessment of computational fluid dynamics (CFD), finite element analysis (FEA), and fluid-solid interaction (FSI) analysis offer a solution that has been increasingly utilized to evaluate BHV mechanics and dynamics. However, the high computational costs and the complex operation of computational modeling hinder its application. Recent advancements in the deep learning (DL) domain can offer a real-time surrogate that can render hemodynamic parameters in a few seconds, thus guiding clinicians to select the optimal treatment option. Herein, we provide a comprehensive review of classical computational modeling approaches, medical imaging, and DL approaches for planning and outcome assessment of TAVR. Particularly, we focus on DL approaches in previous studies, highlighting the utilized datasets, deployed DL models, and achieved results. We emphasize the critical challenges and recommend several future directions for innovative researchers to tackle. Finally, an end-to-end smart DL framework is outlined for real-time assessment and recommendation of the best BHV design for TAVR. Ultimately, deploying such a framework in future studies will support clinicians in minimizing risks during TAVR therapy planning and will help in improving patient care.

The Association of Coronary Fat Attenuation Index Quantified by Automated Software on Coronary Computed Tomography Angiography with Adverse Events in Patients with Less than Moderate Coronary Artery Stenosis

OBJECTIVE

This study analyzed the relationship between the coronary FAI on CCTA and coronary adverse events in patients with moderate coronary artery disease based on machine learning.

METHODS

A total of 172 patients with coronary artery disease with moderate or lower coronary artery stenosis were included. According to whether the patients had coronary adverse events, the patients were divided into an adverse group and a non-adverse group. The coronary FAI of patients was quantified via machine learning, and significant differences between the two groups were analyzed via t-test.

RESULTS

The age difference between the two groups was statistically significant (p < 0.001). The group that had adverse reactions was older, and there was no statistically significant difference between the two groups in terms of sex and smoking status. There was no statistical significance in the blood biochemical indexes between the two groups (p > 0.05). There was a significant difference in the FAIs between the two groups (p < 0.05), with the FAI of the defective group being greater than that of the nonperforming group. Taking the age of patients as a covariate, an analysis of covariance showed that after excluding the influence of age, the FAIs between the two groups were still significantly different (p < 0.001).

Additional coils mitigate elevated defibrillation threshold in right-sided implantable cardioverter defibrillator generator placement: a simulation study

AIMS

The standard implantable cardioverter defibrillator (ICD) generator (can) is placed in the left pectoral area; however, in certain circumstances, right-sided cans may be required which may increase defibrillation threshold (DFT) due to suboptimal shock vectors. We aim to quantitatively assess whether the potential increase in DFT of right-sided can configurations may be mitigated by alternate positioning of the right ventricular (RV) shocking coil or adding coils in the superior vena cava (SVC) and coronary sinus (CS).

METHODS AND RESULTS

A cohort of CT-derived torso models was used to assess DFT of ICD configurations with right-sided cans and alternate positioning of RV shock coils. Efficacy changes with additional coils in the SVC and CS were evaluated. A right-sided can with an apical RV shock coil significantly increased DFT compared to a left-sided can [19.5 (16.4, 27.1) J vs. 13.3 (11.7, 19.9) J, P < 0.001]. Septal positioning of the RV coil led to a further DFT increase when using a right-sided can [26.7 (18.1, 36.1) J vs. 19.5 (16.4, 27.1) J, P < 0.001], but not a left-sided can [12.1 (8.1, 17.6) J vs. 13.3 (11.7, 19.9) J, P = 0.099). Defibrillation threshold of a right-sided can with apical or septal coil was reduced the most by adding both SVC and CS coils [19.5 (16.4, 27.1) J vs. 6.6 (3.9, 9.9) J, P < 0.001, and 26.7 (18.1, 36.1) J vs. 12.1 (5.7, 13.5) J, P < 0.001].

CONCLUSION

Right-sided, compared to left-sided, can positioning results in a 50% increase in DFT. For right-sided cans, apical shock coil positioning produces a lower DFT than septal positions. Elevated right-sided can DFTs may be mitigated by utilizing additional coils in SVC and CS.

Artificial intelligence in cardiovascular imaging: enhancing image analysis and risk stratification

In this review, we summarize state-of-the-art artificial intelligence applications for non-invasive cardiovascular imaging modalities including CT, MRI, echocardiography, and nuclear myocardial perfusion imaging.

Current and Future Applications of Artificial Intelligence in Cardiac CT

PURPOSE OF REVIEW

In this review, we aim to summarize state-of-the-art artificial intelligence (AI) approaches applied to cardiovascular CT and their future implications.

RECENT FINDINGS

Recent studies have shown that deep learning networks can be applied for rapid automated segmentation of coronary plaque from coronary CT angiography, with AI-enabled measurement of total plaque volume predicting future heart attack. AI has also been applied to automate assessment of coronary artery calcium on cardiac and ungated chest CT and to automate the measurement of epicardial fat. Additionally, AI-based prediction models integrating clinical and imaging parameters have been shown to improve prediction of cardiac events compared to traditional risk scores. Artificial intelligence applications have been applied in all aspects of cardiovascular CT - in image acquisition, reconstruction and denoising, segmentation and quantitative analysis, diagnosis and decision assistance and to integrate prognostic risk from clinical data and images. Further incorporation of artificial intelligence in cardiovascular imaging holds important promise to enhance cardiovascular CT as a precision medicine tool.

Non-invasive localization of post-infarct ventricular tachycardia exit sites to guide ablation planning: a computational deep learning platform utilizing the 12-lead electrocardiogram and intracardiac electrograms from implanted devices

AIMS

Existing strategies that identify post-infarct ventricular tachycardia (VT) ablation target either employ invasive electrophysiological (EP) mapping or non-invasive modalities utilizing the electrocardiogram (ECG). Their success relies on localizing sites critical to the maintenance of the clinical arrhythmia, not always recorded on the 12-lead ECG. Targeting the clinical VT by utilizing electrograms (EGM) recordings stored in implanted devices may aid ablation planning, enhancing safety and speed and potentially reducing the need of VT induction. In this context, we aim to develop a non-invasive computational-deep learning (DL) platform to localize VT exit sites from surface ECGs and implanted device intracardiac EGMs.

METHODS AND RESULTS

A library of ECGs and EGMs from simulated paced beats and representative post-infarct VTs was generated across five torso models. Traces were used to train DL algorithms to localize VT sites of earliest systolic activation; first tested on simulated data and then on a clinically induced VT to show applicability of our platform in clinical settings. Localization performance was estimated via localization errors (LEs) against known VT exit sites from simulations or clinical ablation targets. Surface ECGs successfully localized post-infarct VTs from simulated data with mean LE = 9.61 ± 2.61 mm across torsos. VT localization was successfully achieved from implanted device intracardiac EGMs with mean LE = 13.10 ± 2.36 mm. Finally, the clinically induced VT localization was in agreement with the clinical ablation volume.

CONCLUSION

The proposed framework may be utilized for direct localization of post-infarct VTs from surface ECGs and/or implanted device EGMs, or in conjunction with efficient, patient-specific modelling, enhancing safety and speed of ablation planning.

How scan parameter choice affects deep learning-based coronary artery disease assessment from computed tomography

Recently, algorithms capable of assessing the severity of Coronary Artery Disease (CAD) in form of the Coronary Artery Disease-Reporting and Data System (CAD-RADS) grade from Coronary Computed Tomography Angiography (CCTA) scans using Deep Learning (DL) were proposed. Before considering to apply these algorithms in clinical practice, their robustness regarding different commonly used Computed Tomography (CT)-specific image formation parameters-including denoising strength, slab combination, and reconstruction kernel-needs to be evaluated. For this study, we reconstructed a data set of 500 patient CCTA scans under seven image formation parameter configurations. We select one default configuration and evaluate how varying individual parameters impacts the performance and stability of a typical algorithm for automated CAD assessment from CCTA. This algorithm consists of multiple preprocessing and a DL prediction step. We evaluate the influence of the parameter changes on the entire pipeline and additionally on only the DL step by propagating the centerline extraction results of the default configuration to all others. We consider the standard deviation of the CAD severity prediction grade difference between the default and variation configurations to assess the stability w.r.t. parameter changes. For the full pipeline we observe slight instability (± 0.226 CAD-RADS) for all variations. Predictions are more stable with centerlines propagated from the default to the variation configurations (± 0.122 CAD-RADS), especially for differing denoising strengths (± 0.046 CAD-RADS). However, stacking slabs with sharp boundaries instead of mixing slabs in overlapping regions (called true stack ± 0.313 CAD-RADS) and increasing the sharpness of the reconstruction kernel (± 0.150 CAD-RADS) leads to unstable predictions. Regarding the clinically relevant tasks of excluding CAD (called rule-out; AUC default 0.957, min 0.937) and excluding obstructive CAD (called hold-out; AUC default 0.971, min 0.964) the performance remains on a high level for all variations. Concluding, an influence of reconstruction parameters on the predictions is observed. Especially, scans reconstructed with the true stack parameter need to be treated with caution when using a DL-based method. Also, reconstruction kernels which are underrepresented in the training data increase the prediction uncertainty.

DPC-MSGATNet: dual-path chain multi-scale gated axial-transformer network for four-chamber view segmentation in fetal echocardiography

Echocardiography is essential in evaluating fetal cardiac anatomical structures and functions when clinicians conduct early treatment and screening for congenital heart defects, a common and intricate fetal malformation. Nevertheless, the prenatal detection rate of fetal CHD remains low since the peculiarities of fetal cardiac structures and the variousness of fetal CHD. Precisely segmenting four cardiac chambers can assist clinicians in analyzing cardiac morphology and further facilitate CHD diagnosis. Hence, we design a dual-path chain multi-scale gated axial-transformer network (DPC-MSGATNet) that simultaneously models global dependencies and local visual cues for fetal ultrasound (US) four-chamber (FC) views and further accurately segments four chambers. Our DPC-MSGATNet includes a global and a local branch that simultaneously operates on an entire FC view and image patches to learn multi-scale representations. We design a plug-and-play module, Interactive dual-path chain gated axial-transformer (IDPCGAT), to enhance the interactions between global and local branches. In IDPCGAT, the multi-scale representations from the two branches can complement each other, capturing the same region’s salient features and suppressing feature responses to maintain only the activations associated with specific targets. Extensive experiments demonstrate that the DPC-MSGATNet exceeds seven state-of-the-art convolution- and transformer-based methods by a large margin in terms of F1 and IoU scores on our fetal FC view dataset, achieving a F1 score of 96.87$$\%$$ % and an IoU score of 93.99$$\%$$ % . The codes and datasets can be available at https://github.comQiaoSiBo/DPC-MSGATNet.

Global Sensitivity Analysis of Four Chamber Heart Hemodynamics Using Surrogate Models

Computational Fluid Dynamics (CFD) is used to assist in designing artificial valves and planning procedures, focusing on local flow features. However, assessing the impact on overall cardiovascular function or predicting longer-term outcomes may requires more comprehensive whole heart CFD models. Fitting such models to patient data requires numerous computationally expensive simulations, and depends on specific clinical measurements to constrain model parameters, hampering clinical adoption. Surrogate models can help to accelerate the fitting process while accounting for the added uncertainty. We create a validated patient-specific four-chamber heart CFD model based on the Navier-Stokes-Brinkman (NSB) equations and test Gaussian Process Emulators (GPEs) as a surrogate model for performing a variance-based global sensitivity analysis (GSA). GSA identified preload as the dominant driver of flow in both the right and left side of the heart, respectively. Left-right differences were seen in terms of vascular outflow resistances, with pulmonary artery resistance having a much larger impact on flow than aortic resistance. Our results suggest that GPEs can be used to identify parameters in personalized whole heart CFD models, and highlight the importance of accurate preload measurements.

Automated Dual-energy Computed Tomography-based Extracellular Volume Estimation for Myocardial Characterization in Patients With Ischemic and Nonischemic Cardiomyopathy

OBJECTIVES

We aimed to validate and test a prototype algorithm for automated dual-energy computed tomography (DECT)-based myocardial extracellular volume (ECV) assessment in patients with various cardiomyopathies.

METHODS

This retrospective study included healthy subjects (n=9; 61±10 y) and patients with cardiomyopathy (n=109, including a validation cohort n=60; 68±9 y; and a test cohort n=49; 69±11 y), who had previously undergone cardiac DECT. Myocardial ECV was calculated using a prototype-based fully automated algorithm and compared with manual assessment. Receiver-operating characteristic analysis was performed to test the algorithm's ability to distinguish healthy subjects and patients with cardiomyopathy.

RESULTS

The fully automated method led to a significant reduction of postprocessing time compared with manual assessment (2.2±0.4 min and 9.4±0.7 min, respectively, P <0.001). There was no significant difference in ECV between the automated and manual methods ( P =0.088). The automated method showed moderate correlation and agreement with the manual technique ( r =0.68, intraclass correlation coefficient=0.66). ECV was significantly higher in patients with cardiomyopathy compared with healthy subjects, regardless of the method used ( P <0.001). In the test cohort, the automated method yielded an area under the curve of 0.98 for identifying patients with cardiomyopathies.

CONCLUSION

Automated ECV estimation based on DECT showed moderate agreement with the manual method and matched with previously reported ECV values for healthy volunteers and patients with cardiomyopathy. The automatically derived ECV demonstrated an excellent diagnostic performance to discriminate between healthy and diseased myocardium, suggesting that it could be an effective initial screening tool while significantly reducing the time of assessment.

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.

Machine Learning-Based Automated Diagnostic Systems Developed for Heart Failure Prediction Using Different Types of Data Modalities: A Systematic Review and Future Directions

One of the leading causes of deaths around the globe is heart disease. Heart is an organ that is responsible for the supply of blood to each part of the body. Coronary artery disease (CAD) and chronic heart failure (CHF) often lead to heart attack. Traditional medical procedures (angiography) for the diagnosis of heart disease have higher cost as well as serious health concerns. Therefore, researchers have developed various automated diagnostic systems based on machine learning (ML) and data mining techniques. ML-based automated diagnostic systems provide an affordable, efficient, and reliable solutions for heart disease detection. Various ML, data mining methods, and data modalities have been utilized in the past. Many previous review papers have presented systematic reviews based on one type of data modality. This study, therefore, targets systematic review of automated diagnosis for heart disease prediction based on different types of modalities, i.e., clinical feature-based data modality, images, and ECG. Moreover, this paper critically evaluates the previous methods and presents the limitations in these methods. Finally, the article provides some future research directions in the domain of automated heart disease detection based on machine learning and multiple of data modalities.

Automatic Aortic Valve Cusps Segmentation from CT Images Based on the Cascading Multiple Deep Neural Networks

Accurate morphological information on aortic valve cusps is critical in treatment planning. Image segmentation is necessary to acquire this information, but manual segmentation is tedious and time consuming. In this paper, we propose a fully automatic aortic valve cusps segmentation method from CT images by combining two deep neural networks, spatial configuration-Net for detecting anatomical landmarks and U-Net for segmentation of aortic valve components. A total of 258 CT volumes of end systolic and end diastolic phases, which include cases with and without severe calcifications, were collected and manually annotated for each aortic valve component. The collected CT volumes were split 6:2:2 for the training, validation and test steps, and our method was evaluated by five-fold cross validation. The segmentation was successful for all CT volumes with 69.26 s as mean processing time. For the segmentation results of the aortic root, the right-coronary cusp, the left-coronary cusp and the non-coronary cusp, mean Dice Coefficient were 0.95, 0.70, 0.69, and 0.67, respectively. There were strong correlations between measurement values automatically calculated based on the annotations and those based on the segmentation results. The results suggest that our method can be used to automatically obtain measurement values for aortic valve morphology.

Cardiac CT motion artifact grading via semi-automatic labeling and vessel tracking using synthetic image-augmented training data

Cardiac CT provides critical information for the evaluation of cardiovascular diseases. However, involuntary patient motion and physiological movement of the organs during CT scanning cause motion blur in the reconstructed CT images, degrading both cardiac CT image quality and its diagnostic value. In this paper, we propose and demonstrate an effective and efficient method for CT coronary angiography image quality grading via semi-automatic labeling and vessel tracking. These algorithms produce scores that accord with those of expert readers to within 0.85 points on a 5-point scale. We also train a neural network model to perform fully-automatic motion artifact grading. We demonstrate, using XCAT simulation tools to generate realistic phantom CT data, that supplementing clinical data with synthetic data improves the scoring performance of this network. With respect to ground truth scores assigned by expert operators, the mean square error of grading motion of the right coronary artery is reduced by 36% by synthetic data supplementation. This demonstrates that augmentation of clinical training data with realistically synthesized images can potentially reduce the number of clinical studies needed to train the network.

Deep learning-based whole-heart segmentation in 4D contrast-enhanced cardiac CT

Automatic cardiac chamber and left ventricular (LV) myocardium segmentation over the cardiac cycle significantly extends the utilization of contrast-enhanced cardiac CT, potentially enabling in-depth assessment of cardiac function. Therefore, we evaluate an automatic method for cardiac chamber and LV myocardium segmentation in 4D cardiac CT. In this study, 4D contrast-enhanced cardiac CT scans of 1509 patients selected for transcatheter aortic valve implantation with 21,605 3D images, were divided into development (N = 12) and test set (N = 1497). 3D convolutional neural networks were trained with end-systolic (ES) and end-diastolic (ED) images. Dice similarity coefficient (DSC) and average symmetric surface distance (ASSD) were computed for 3D segmentations at ES and ED in the development set via cross-validation, and for 2D segmentations in four cardiac phases for 81 test set patients. Segmentation quality in the full test set of 1497 patients was assessed visually on a three-point scale per structure based on estimated overlap with the ground truth. Automatic segmentation resulted in a mean DSC of 0.89 ± 0.10 and ASSD of 1.43 ± 1.45 mm in 12 patients in 3D, and a DSC of 0.89 ± 0.08 and ASSD of 1.86 ± 1.20 mm in 81 patients in 2D. The qualitative evaluation in the whole test set of 1497 patients showed that automatic segmentations were assigned grade 1 (clinically useful) in 98.5%, 92.2%, 83.1%, 96.3%, and 91.6% of cases for LV cavity and myocardium, right ventricle, left atrium, and right atrium. Our automatic method using convolutional neural networks performed clinically useful segmentation across the cardiac cycle in a large set of 4D cardiac CT images, potentially enabling in-depth assessment of cardiac function.

Artificial intelligence in cardiovascular CT: Current status and future implications

Artificial intelligence (AI) refers to the use of computational techniques to mimic human thought processes and learning capacity. The past decade has seen a rapid proliferation of AI developments for cardiovascular computed tomography (CT). These algorithms aim to increase efficiency, objectivity, and performance in clinical tasks such as image quality improvement, structure segmentation, quantitative measurements, and outcome prediction. By doing so, AI has the potential to streamline clinical workflow, increase interpretative speed and accuracy, and inform subsequent clinical pathways. This review covers state-of-the-art AI techniques in cardiovascular CT and the future role of AI as a clinical support tool.

A category attention instance segmentation network for four cardiac chambers segmentation in fetal echocardiography

Fetal echocardiography is an essential and comprehensive examination technique for the detection of fetal heart anomalies. Accurate cardiac chambers segmentation can assist cardiologists to analyze cardiac morphology and facilitate heart disease diagnosis. Previous research mainly focused on the segmentation of single cardiac chambers, such as left ventricle (LV) segmentation or left atrium (LA) segmentation. We propose a generic framework based on instance segmentation to segment the four cardiac chambers accurately and simultaneously. The proposed Category Attention Instance Segmentation Network (CA-ISNet) has three branches: a category branch for predicting the semantic category, a mask branch for segmenting the cardiac chambers, and a category attention branch for learning category information of instances. The category attention branch is used to correct instance misclassification of the category branch. In our collected dataset, which contains echocardiography images with four-chamber views of 319 fetuses, experimental results show our method can achieve superior segmentation performance against state-of-the-art methods. Specifically, using fivefold cross-validation, our model achieves Dice coefficients of 0.7956, 0.7619, 0.8199, 0.7470 for the four cardiac chambers, and with an average precision of 45.64%.

Systematic Comparison of Left Ventricular Geometry Between 3D-Echocardiography and Cardiac Magnetic Resonance Imaging

Aims: Left ventricular (LV) volumes estimated using three-dimensional echocardiography (3D-echo) have been reported to be smaller than those measured using cardiac magnetic resonance (CMR) imaging, but the underlying causes are not well-understood. We investigated differences in regional LV anatomy derived from these modalities and related subsequent findings to image characteristics. Methods and Results: Seventy participants (18 patients and 52 healthy participants) were imaged with 3D-echo and CMR (<1 h apart). Three-dimensional left ventricular models were constructed at end-diastole (ED) and end-systole (ES) from both modalities using previously validated software, enabling the fusion of CMR with 3D-echo by rigid registration. Regional differences were evaluated as mean surface distances for each of the 17 American Heart Association segments, and by comparing contours superimposed on images from each modality. In comparison to CMR-derived models, 3D-echo models underestimated LV end-diastolic volume (EDV) by -16 ± 22, -1 ± 25, and -18 ± 24 ml across three independent analysis methods. Average surface distance errors were largest in the basal-anterolateral segment (11-15 mm) and smallest in the mid-inferoseptal segment (6 mm). Larger errors were associated with signal dropout in anterior regions and the appearance of trabeculae at the lateral wall. Conclusions: Fusion of CMR and 3D-echo provides insight into the causes of volume underestimation by 3D-echo. Systematic signal dropout and differences in appearances of trabeculae lead to discrepancies in the delineation of LV geometry at anterior and lateral regions. A better understanding of error sources across modalities may improve correlation of clinical indices between 3D-echo and CMR.

Automatic quantification of epicardial adipose tissue volume

PURPOSE

Epicardial fat is the adipose tissue between the serosal pericardial wall layer and the visceral layer. It is distributed mainly around the atrioventricular groove, atrial septum, ventricular septum and coronary arteries. Studies have shown that the density, thickness, volume and other characteristics of epicardial adipose tissue (EAT) are independently correlated with a variety of cardiovascular diseases. Given this association, the accurate determination of EAT volume is an essential aim of future research. Therefore, the purpose of this study was to establish a framework for fully automatic EAT segmentation and quantification in coronary computed tomography angiography (CCTA) scans.

METHODS

A set of 103 scans are randomly selected from our medical center. An automatic pipeline has been developed to segment and quantify the volume of EAT. First, a multi-slice deep neural network is used to simultaneously segment the pericardium in multiple adjacent slices. Then a deformable model is employed to reduce false positive and negative regions in the segmented binary pericardial images. Finally, the pericardium mask is used to define the region of interest (ROI) and the threshold method is utilized to extract the pixels ranging from -175 Hounsfield units (HU) to -15 HU for the segmentation of EAT.

RESULTS

The Dice indices of the pericardial segmentation using the proposed method with respect to the manual delineation results of two radiology experts were 97.1%  ±  0.7% and 96.9%  ±  0.6%, respectively. The inter-observer variability was also assessed, resulting in a Dice index of 97.0%  ±  0.7%. For the EAT segmentation results, the Dice indices between the proposed method and the two radiology experts were 93.4%  ±  1.5% and 93.3%  ±  1.3%, respectively, and the same measurement between the experts themselves was 93.6%  ±  1.9%. The Pearson's correlation coefficients between the EAT volumes computed from the results of the proposed method and the manual delineation by the two experts were 1.00 and 0.99 and the same coefficients between the experts was 0.99.

CONCLUSIONS

This work describes the development of a fully automatic EAT segmentation and quantification method from CCTA scans and the results compare favorably with the assessments of two independent experts. The proposed method is also packaged with a graphical user interface which can be found at https://github.com/MountainAndMorning/EATSeg.

Position paper of the EACVI and EANM on artificial intelligence applications in multimodality cardiovascular imaging using SPECT/CT, PET/CT, and cardiac CT

In daily clinical practice, clinicians integrate available data to ascertain the diagnostic and prognostic probability of a disease or clinical outcome for their patients. For patients with suspected or known cardiovascular disease, several anatomical and functional imaging techniques are commonly performed to aid this endeavor, including coronary computed tomography angiography (CCTA) and nuclear cardiology imaging. Continuous improvement in positron emission tomography (PET), single-photon emission computed tomography (SPECT), and CT hardware and software has resulted in improved diagnostic performance and wide implementation of these imaging techniques in daily clinical practice. However, the human ability to interpret, quantify, and integrate these data sets is limited. The identification of novel markers and application of machine learning (ML) algorithms, including deep learning (DL) to cardiovascular imaging techniques will further improve diagnosis and prognostication for patients with cardiovascular diseases. The goal of this position paper of the European Association of Nuclear Medicine (EANM) and the European Association of Cardiovascular Imaging (EACVI) is to provide an overview of the general concepts behind modern machine learning-based artificial intelligence, highlights currently prefered methods, practices, and computational models, and proposes new strategies to support the clinical application of ML in the field of cardiovascular imaging using nuclear cardiology (hybrid) and CT techniques.

Polygonal surface processing and mesh generation tools for the numerical simulation of the cardiac function

In order to simulate the cardiac function for a patient-specific geometry, the generation of the computational mesh is crucially important. In practice, the input is typically a set of unprocessed polygonal surfaces coming either from a template geometry or from medical images. These surfaces need ad-hoc processing to be suitable for a volumetric mesh generation. In this work we propose a set of new algorithms and tools aiming to facilitate the mesh generation process. In particular, we focus on different aspects of a cardiac mesh generation pipeline: (1) specific polygonal surface processing for cardiac geometries, like connection of different heart chambers or segmentation outputs; (2) generation of accurate boundary tags; (3) definition of mesh-size functions dependent on relevant geometric quantities; (4) processing and connecting together several volumetric meshes. The new algorithms-implemented in the open-source software vmtk-can be combined with each other allowing the creation of personalized pipelines, that can be optimized for each cardiac geometry or for each aspect of the cardiac function to be modeled. Thanks to these features, the proposed tools can significantly speed-up the mesh generation process for a large range of cardiac applications, from single-chamber single-physics simulations to multi-chambers multi-physics simulations. We detail all the proposed algorithms motivating them in the cardiac context and we highlight their flexibility by showing different examples of cardiac mesh generation pipelines.

Linking statistical shape models and simulated function in the healthy adult human heart

Cardiac anatomy plays a crucial role in determining cardiac function. However, there is a poor understanding of how specific and localised anatomical changes affect different cardiac functional outputs. In this work, we test the hypothesis that in a statistical shape model (SSM), the modes that are most relevant for describing anatomy are also most important for determining the output of cardiac electromechanics simulations. We made patient-specific four-chamber heart meshes (n = 20) from cardiac CT images in asymptomatic subjects and created a SSM from 19 cases. Nine modes captured 90% of the anatomical variation in the SSM. Functional simulation outputs correlated best with modes 2, 3 and 9 on average (R = 0.49 ± 0.17, 0.37 ± 0.23 and 0.34 ± 0.17 respectively). We performed a global sensitivity analysis to identify the different modes responsible for different simulated electrical and mechanical measures of cardiac function. Modes 2 and 9 were the most important for determining simulated left ventricular mechanics and pressure-derived phenotypes. Mode 2 explained 28.56 ± 16.48% and 25.5 ± 20.85, and mode 9 explained 12.1 ± 8.74% and 13.54 ± 16.91% of the variances of mechanics and pressure-derived phenotypes, respectively. Electrophysiological biomarkers were explained by the interaction of 3 ± 1 modes. In the healthy adult human heart, shape modes that explain large portions of anatomical variance do not explain equivalent levels of electromechanical functional variation. As a result, in cardiac models, representing patient anatomy using a limited number of modes of anatomical variation can cause a loss in accuracy of simulated electromechanical function.

Multitask Learning for Estimating Multitype Cardiac Indices in MRI and CT Based on Adversarial Reverse Mapping

The estimation of multitype cardiac indices from cardiac magnetic resonance imaging (MRI) and computed tomography (CT) images attracts great attention because of its clinical potential for comprehensive function assessment. However, the most exiting model can only work in one imaging modality (MRI or CT) without transferable capability. In this article, we propose the multitask learning method with the reverse inferring for estimating multitype cardiac indices in MRI and CT. Different from the existing forward inferring methods, our method builds a reverse mapping network that maps the multitype cardiac indices to cardiac images. The task dependencies are then learned and shared to multitask learning networks using an adversarial training approach. Finally, we transfer the parameters learned from MRI to CT. A series of experiments were conducted in which we first optimized the performance of our framework via ten-fold cross-validation of over 2900 cardiac MRI images. Then, the fine-tuned network was run on an independent data set with 2360 cardiac CT images. The results of all the experiments conducted on the proposed adversarial reverse mapping show excellent performance in estimating multitype cardiac indices.

Automated interpretation of congenital heart disease from multi-view echocardiograms

Congenital heart disease (CHD) is the most common birth defect and the leading cause of neonate death in China. Clinical diagnosis can be based on the selected 2D key-frames from five views. Limited by the availability of multi-view data, most methods have to rely on the insufficient single view analysis. This study proposes to automatically analyze the multi-view echocardiograms with a practical end-to-end framework. We collect the five-view echocardiograms video records of 1308 subjects (including normal controls, ventricular septal defect (VSD) patients and atrial septal defect (ASD) patients) with both disease labels and standard-view key-frame labels. Depthwise separable convolution-based multi-channel networks are adopted to largely reduce the network parameters. We also approach the imbalanced class problem by augmenting the positive training samples. Our 2D key-frame model can diagnose CHD or negative samples with an accuracy of 95.4%, and in negative, VSD or ASD classification with an accuracy of 92.3%. To further alleviate the work of key-frame selection in real-world implementation, we propose an adaptive soft attention scheme to directly explore the raw video data. Four kinds of neural aggregation methods are systematically investigated to fuse the information of an arbitrary number of frames in a video. Moreover, with a view detection module, the system can work without the view records. Our video-based model can diagnose with an accuracy of 93.9% (binary classification), and 92.1% (3-class classification) in a collected 2D video testing set, which does not need key-frame selection and view annotation in testing. The detailed ablation study and the interpretability analysis are provided. The presented model has high diagnostic rates for VSD and ASD that can be potentially applied to the clinical practice in the future. The short-term automated machine learning process can partially replace and promote the long-term professional training of primary doctors, improving the primary diagnosis rate of CHD in China, and laying the foundation for early diagnosis and timely treatment of children with CHD.

MAD-UNet: A deep U-shaped network combined with an attention mechanism for pancreas segmentation in CT images

PURPOSE

Pancreas segmentation is a difficult task because of the high intrapatient variability in the shape, size, and location of the organ, as well as the low contrast and small footprint of the CT scan. At present, the U-Net model is likely to lead to the problems of intraclass inconsistency and interclass indistinction in pancreas segmentation. To solve this problem, we improved the contextual and semantic feature information acquisition method of the biomedical image segmentation model (U-Net) based on a convolutional network and proposed an improved segmentation model called the multiscale attention dense residual U-shaped network (MAD-UNet).

METHODS

There are two aspects considered in this method. First, we adopted dense residual blocks and weighted binary cross-entropy to enhance the semantic features to learn the details of the pancreas. Using such an approach can reduce the effects of intraclass inconsistency. Second, we used an attention mechanism and multiscale convolution to enrich the contextual information and suppress learning in unrelated areas. We let the model be more sensitive to pancreatic marginal information and reduced the impact of interclass indistinction.

RESULTS

We evaluated our model using fourfold cross-validation on 82 abdominal enhanced three-dimensional (3D) CT scans from the National Institutes of Health (NIH-82) and 281 3D CT scans from the 2018 MICCAI segmentation decathlon challenge (MSD). The experimental results showed that our method achieved state-of-the-art performance on the two pancreatic datasets. The mean Dice coefficients were 86.10% ± 3.52% and 88.50% ± 3.70%.

CONCLUSIONS

Our model can effectively solve the problems of intraclass inconsistency and interclass indistinction in the segmentation of the pancreas, and it has value in clinical application. Code is available at https://github.com/Mrqins/pancreas-segmentation.

One Shot Segmentation: Unifying Rigid Detection and Non-Rigid Segmentation Using Elastic Regularization

This paper proposes a novel approach for the non-rigid segmentation of deformable objects in image sequences, which is based on one-shot segmentation that unifies rigid detection and non-rigid segmentation using elastic regularization. The domain of application is the segmentation of a visual object that temporally undergoes a rigid transformation (e.g., affine transformation) and a non-rigid transformation (i.e., contour deformation). The majority of segmentation approaches to solve this problem are generally based on two steps that run in sequence: a rigid detection, followed by a non-rigid segmentation. In this paper, we propose a new approach, where both the rigid and non-rigid segmentation are performed in a single shot using a sparse low-dimensional manifold that represents the visual object deformations. Given the multi-modality of these deformations, the manifold partitions the training data into several patches, where each patch provides a segmentation proposal during the inference process. These multiple segmentation proposals are merged using the classification results produced by deep belief networks (DBN) that compute the confidence on each segmentation proposal. Thus, an ensemble of DBN classifiers is used for estimating the final segmentation. Compared to current methods proposed in the field, our proposed approach is advantageous in four aspects: (i) it is a unified framework to produce rigid and non-rigid segmentations; (ii) it uses an ensemble classification process, which can help the segmentation robustness; (iii) it provides a significant reduction in terms of the number of dimensions of the rigid and non-rigid segmentations search spaces, compared to current approaches that divide these two problems; and (iv) this lower dimensionality of the search space can also reduce the need for large annotated training sets to be used for estimating the DBN models. Experiments on the problem of left ventricle endocardial segmentation from ultrasound images, and lip segmentation from frontal facial images using the extended Cohn-Kanade (CK+) database, demonstrate the potential of the methodology through qualitative and quantitative evaluations, and the ability to reduce the search and training complexities without a significant impact on the segmentation accuracy.

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.

Individual tooth detection and identification from dental panoramic X-ray images via point-wise localization and distance regularization

Dental panoramic X-ray imaging is a popular diagnostic method owing to its very small dose of radiation. For an automated computer-aided diagnosis system in dental clinics, automatic detection and identification of individual teeth from panoramic X-ray images are critical prerequisites. In this study, we propose a point-wise tooth localization neural network by introducing a spatial distance regularization loss. The proposed network initially performs center point regression for all the anatomical teeth (i.e., 32 points), which automatically identifies each tooth. A novel distance regularization penalty is employed on the 32 points by considering L₂ regularization loss of Laplacian on spatial distances. Subsequently, teeth boxes are individually localized using a multitask neural network on a patch basis. A multitask offset training is employed on the final output to improve the localization accuracy. Our method successfully localizes not only the existing teeth but also missing teeth; consequently, highly accurate detection and identification are achieved. The experimental results demonstrate that the proposed algorithm outperforms state-of-the-art approaches by increasing the average precision of teeth detection by 15.71 % compared to the best performing method. The accuracy of identification achieved a precision of 0.997 and recall value of 0.972. Moreover, the proposed network does not require any additional identification algorithm owing to the preceding regression of the fixed 32 points regardless of the existence of the teeth.

Automatic estimation of aortic and mitral valve displacements in dynamic CTA with 4D graph-cuts

The location of the mitral and aortic valves in dynamic cardiac imaging is useful for extracting functional derived parameters such as ejection fraction, valve excursions, and global longitudinal strain, and when performing anatomical structures tracking using slice following or valve intervention's planning. Completely automatic segmentation methods are still challenging tasks because of their fast movements and the different positions that prevent good visibility of the leaflets along the full cardiac cycle. In this article, we propose a processing pipeline to track the displacement of the aortic and mitral valve annuli from high-resolution cardiac four-dimensional computed tomographic angiography (4D-CTA). The proposed method is based on the dynamic separation of left ventricle, left atrium and aorta using statistical shape modeling and an energy minimization algorithm based on graph-cuts and has been evaluated on a set of 15 electrocardiography-gated 4D-CTAs. We report a mean agreement distance between manual annotations and our proposed method of 2.52±1.06 mm for the mitral annulus and 2.00±0.69 mm for the aortic valve annulus based on valve locations detected from manual anatomical landmarks. In addition, we show the effect of detecting the valvular planes on derived functional parameters (ejection fraction, global longitudinal strain, and excursions of the mitral and aortic valves).

In-silico pace-mapping using a detailed whole torso model and implanted electronic device electrograms for more efficient ablation planning

BACKGROUND

Pace-mapping is a commonly used electrophysiological (EP) procedure which aims to identify exit sites of ventricular tachycardia (VT) by matching ventricular activation patterns (assessed by QRS morphology) at specific pacing locations with activation during VT. However, long procedure durations and the need for VT induction render this technique non-optimal. To demonstrate the potential of in-silico pace-mapping, using stored electrogram (EGM) recordings of clinical VT from implanted devices to guide pre-procedural ablation planning.

METHOD

Six scar-related VT episodes were simulated in a 3D torso model reconstructed from computed tomography (CT) imaging data, including three different infarct anatomies mapped from infarcted porcine imaging data. In-silico pace-mapping was performed to localise VT exit sites and isthmuses by using 12-lead electrocardiogram (ECG) signals and different combinations of EGM sensing vectors from implanted devices, through the creation of conventional correlation maps and reference-less maps.

RESULTS

Our in-silico platform was successful in identifying VT exit sites for a variety of different VT morphologies from both ECG correlation maps and corresponding EGM maps, with the latter dependent upon the number of sensing vectors used. We also showed the added utility of both ECG and EGM reference-less pace-mapping for the identification of slow-conducting isthmuses, uncovering the optimal algorithm parameters. Finally, EGM-based pace-mapping was shown to be more dependent upon the mapped surface (epicardial/endocardial), relative to the VT origin.

CONCLUSIONS

In-silico pace-mapping can be used along with EGMs from implanted devices to localise VT ablation targets in pre-procedural planning.

3D Printing, Computational Modeling, and Artificial Intelligence for Structural Heart Disease

Structural heart disease (SHD) is a new field within cardiovascular medicine. Traditional imaging modalities fall short in supporting the needs of SHD interventions, as they have been constructed around the concept of disease diagnosis. SHD interventions disrupt traditional concepts of imaging in requiring imaging to plan, simulate, and predict intraprocedural outcomes. In transcatheter SHD interventions, the absence of a gold-standard open cavity surgical field deprives physicians of the opportunity for tactile feedback and visual confirmation of cardiac anatomy. Hence, dependency on imaging in periprocedural guidance has led to evolution of a new generation of procedural skillsets, concept of a visual field, and technologies in the periprocedural planning period to accelerate preclinical device development, physician, and patient education. Adaptation of 3-dimensional (3D) printing in clinical care and procedural planning has demonstrated a reduction in early-operator learning curve for transcatheter interventions. Integration of computation modeling to 3D printing has accelerated research and development understanding of fluid mechanics within device testing. Application of 3D printing, computational modeling, and ultimately incorporation of artificial intelligence is changing the landscape of physician training and delivery of patient-centric care. Transcatheter structural heart interventions are requiring in-depth periprocedural understanding of cardiac pathophysiology and device interactions not afforded by traditional imaging metrics.

A computational investigation into rate-dependant vectorcardiogram changes due to specific fibrosis patterns in non-ischæmic dilated cardiomyopathy

Patients with scar-associated fibrotic tissue remodelling are at greater risk of ventricular arrhythmic events, but current methods to detect the presence of such remodelling require invasive procedures. We present here a potential method to detect the presence, location and dimensions of scar using pacing-dependent changes in the vectorcardiogram (VCG). Using a clinically-derived whole-torso computational model, simulations were conducted at both slow and rapid pacing for a variety of scar patterns within the myocardium, with various VCG-derived metrics being calculated, with changes in these metrics being assessed for their ability to discern the presence and size of scar. Our results indicate that differences in the dipole angle at the end of the QRS complex and differences in the QRS area and duration may be used to predict scar properties. Using machine learning techniques, we were also able to predict the location of the scar to high accuracy, using only these VCG-derived rate-dependent changes as input. Such a non-invasive predictive tool for the presence of scar represents a potentially useful clinical tool for identifying patients at arrhythmic risk.

A publicly available virtual cohort of four-chamber heart meshes for cardiac electro-mechanics simulations

Computational models of the heart are increasingly being used in the development of devices, patient diagnosis and therapy guidance. While software techniques have been developed for simulating single hearts, there remain significant challenges in simulating cohorts of virtual hearts from multiple patients. To facilitate the development of new simulation and model analysis techniques by groups without direct access to medical data, image analysis techniques and meshing tools, we have created the first publicly available virtual cohort of twenty-four four-chamber hearts. Our cohort was built from heart failure patients, age 67±14 years. We segmented four-chamber heart geometries from end-diastolic (ED) CT images and generated linear tetrahedral meshes with an average edge length of 1.1±0.2mm. Ventricular fibres were added in the ventricles with a rule-based method with an orientation of -60° and 80° at the epicardium and endocardium, respectively. We additionally refined the meshes to an average edge length of 0.39±0.10mm to show that all given meshes can be resampled to achieve an arbitrary desired resolution. We ran simulations for ventricular electrical activation and free mechanical contraction on all 1.1mm-resolution meshes to ensure that our meshes are suitable for electro-mechanical simulations. Simulations for electrical activation resulted in a total activation time of 149±16ms. Free mechanical contractions gave an average left ventricular (LV) and right ventricular (RV) ejection fraction (EF) of 35±1% and 30±2%, respectively, and a LV and RV stroke volume (SV) of 95±28mL and 65±11mL, respectively. By making the cohort publicly available, we hope to facilitate large cohort computational studies and to promote the development of cardiac computational electro-mechanics for clinical applications.