Evaluation of Deep Learning Estimation of Whole Heart Anatomy from Automated Cardiovascular Magnetic Resonance Short- and Long-Axis Analyses in UK Biobank.

AIMS: Standard methods of heart chamber volume estimation in cardiovascular magnetic resonance (CMR) typically utilize simple geometric formulae based on a limited number of slices. We aimed to evaluate whether an automated deep learning neural network prediction of 3D anatomy of all four chambers would show stronger associations with cardiovascular risk factors and disease than standard volume estimation methods in the UK Biobank. METHODS: A deep learning network was adapted to predict 3D segmentations of left and right ventricles (LV, RV) and atria (LA, RA) at ∼1mm isotropic resolution from CMR short and long axis 2D segmentations obtained from a fully automated machine learning pipeline in 4723 individuals with cardiovascular disease (CVD) and 5733 without in the UK Biobank. Relationships between volumes at end-diastole (ED) and end-systole (ES) and risk/disease factors were quantified using univariate, multivariate and logistic regression analyses. Strength of association between deep learning volumes and standard volumes was compared using the area under the receiving operator characteristic curve (AUC). RESULTS: Univariate and multivariate associations between deep learning volumes and most risk and disease factors were stronger than for standard volumes (higher R2 and more significant P values), particularly for sex, age, and body mass index. AUC for all logistic regressions were higher for deep learning volumes than standard volumes (p<0.001 for all four chambers at ED and ES). CONCLUSIONS: Neural network reconstructions of whole heart volumes had significantly stronger associations with cardiovascular disease and risk factors than standard volume estimation methods in an automatic processing pipeline.

Whole-heart electromechanical simulations using Latent Neural Ordinary Differential Equations.

Cardiac digital twins provide a physics and physiology informed framework to deliver personalized medicine. However, high-fidelity multi-scale cardiac models remain a barrier to adoption due to their extensive computational costs. Artificial Intelligence-based methods can make the creation of fast and accurate whole-heart digital twins feasible. We use Latent Neural Ordinary Differential Equations (LNODEs) to learn the pressure-volume dynamics of a heart failure patient. Our surrogate model is trained from 400 simulations while accounting for 43 parameters describing cell-to-organ cardiac electromechanics and cardiovascular hemodynamics. LNODEs provide a compact representation of the 3D-0D model in a latent space by means of an Artificial Neural Network that retains only 3 hidden layers with 13 neurons per layer and allows for numerical simulations of cardiac function on a single processor. We employ LNODEs to perform global sensitivity analysis and parameter estimation with uncertainty quantification in 3 hours of computations, still on a single processor.

Neural network emulation of the human ventricular cardiomyocyte action potential for more efficient computations in pharmacological studies.

Computer models of the human ventricular cardiomyocyte action potential (AP) have reached a level of detail and maturity that has led to an increasing number of applications in the pharmaceutical sector. However, interfacing the models with experimental data can become a significant computational burden. To mitigate the computational burden, the present study introduces a neural network (NN) that emulates the AP for given maximum conductances of selected ion channels, pumps, and exchangers. Its applicability in pharmacological studies was tested on synthetic and experimental data. The NN emulator potentially enables massive speed-ups compared to regular simulations and the forward problem (find drugged AP for pharmacological parameters defined as scaling factors of control maximum conductances) on synthetic data could be solved with average root-mean-square errors (RMSE) of 0.47 mV in normal APs and of 14.5 mV in abnormal APs exhibiting early afterdepolarizations (72.5% of the emulated APs were alining with the abnormality, and the substantial majority of the remaining APs demonstrated pronounced proximity). This demonstrates not only very fast and mostly very accurate AP emulations but also the capability of accounting for discontinuities, a major advantage over existing emulation strategies. Furthermore, the inverse problem (find pharmacological parameters for control and drugged APs through optimization) on synthetic data could be solved with high accuracy shown by a maximum RMSE of 0.22 in the estimated pharmacological parameters. However, notable mismatches were observed between pharmacological parameters estimated from experimental data and distributions obtained from the Comprehensive in vitro Proarrhythmia Assay initiative. This reveals larger inaccuracies which can be attributed particularly to the fact that small tissue preparations were studied while the emulator was trained on single cardiomyocyte data. Overall, our study highlights the potential of NN emulators as powerful tool for an increased efficiency in future quantitative systems pharmacology studies.

A three-dimensional left atrial motion estimation from retrospective gated computed tomography: application in heart failure patients with atrial fibrillation.

Background: A reduced left atrial (LA) strain correlates with the presence of atrial fibrillation (AF). Conventional atrial strain analysis uses two-dimensional (2D) imaging, which is, however, limited by atrial foreshortening and an underestimation of through-plane motion. Retrospective gated computed tomography (RGCT) produces high-fidelity three-dimensional (3D) images of the cardiac anatomy throughout the cardiac cycle that can be used for estimating 3D mechanics. Its feasibility for LA strain measurement, however, is understudied. Aim: The aim of this study is to develop and apply a novel workflow to estimate 3D LA motion and calculate the strain from RGCT imaging. The utility of global and regional strains to separate heart failure in patients with reduced ejection fraction (HFrEF) with and without AF is investigated. Methods: A cohort of 30 HFrEF patients with (n = 9) and without (n = 21) AF underwent RGCT prior to cardiac resynchronisation therapy. The temporal sparse free form deformation image registration method was optimised for LA feature tracking in RGCT images and used to estimate 3D LA endocardial motion. The area and fibre reservoir strains were calculated over the LA body. Universal atrial coordinates and a human atrial fibre atlas enabled the regional strain calculation and the fibre strain calculation along the local myofibre orientation, respectively. Results: It was found that global reservoir strains were significantly reduced in the HFrEF + AF group patients compared with the HFrEF-only group patients (area strain: 11.2 ± 4.8% vs. 25.3 ± 12.6%, P = 0.001; fibre strain: 4.5 ± 2.0% vs. 15.2 ± 8.8%, P = 0.001), with HFrEF + AF patients having a greater regional reservoir strain dyssynchrony. All regional reservoir strains were reduced in the HFrEF + AF patient group, in whom the inferior wall strains exhibited the most significant differences. The global reservoir fibre strain and LA volume + posterior wall reservoir fibre strain exceeded LA volume alone and 2D global longitudinal strain (GLS) for AF classification (area-under-the-curve: global reservoir fibre strain: 0.94 ± 0.02, LA volume + posterior wall reservoir fibre strain: 0.95 ± 0.02, LA volume: 0.89 ± 0.03, 2D GLS: 0.90 ± 0.03). Conclusion: RGCT enables 3D LA motion estimation and strain calculation that outperforms 2D strain metrics and LA enlargement for AF classification. Differences in regional LA strain could reflect regional myocardial properties such as atrial fibrosis burden.

Performance of Image-navigated and Diaphragm-navigated 3D Late Gadolinium-enhanced Cardiac MRI for the Assessment of Atrial Fibrosis.

Purpose To perform a qualitative and quantitative evaluation of the novel image-navigated (iNAV) 3D late gadolinium enhancement (LGE) cardiac MRI imaging strategy in comparison with the conventional diaphragm-navigated (dNAV) 3D LGE cardiac MRI strategy for the assessment of left atrial fibrosis in atrial fibrillation (AF). Materials and Methods In this prospective study conducted between April and September 2022, 26 consecutive participants with AF (mean age, 61 ± 11 years; 19 male) underwent both iNAV and dNAV 3D LGE cardiac MRI, with equivalent spatial resolution and timing in the cardiac cycle. Participants were randomized in the acquisition order of iNAV and dNAV. Both, iNAV-LGE and dNAV-LGE images were analyzed qualitatively using a 5-point Likert scale and quantitatively (percentage of atrial fibrosis using image intensity ratio threshold 1.2), including testing for overlap in atrial fibrosis areas by calculating Dice score. Results Acquisition time of iNAV was significantly lower compared with dNAV (4.9 ± 1.1 minutes versus 12 ± 4 minutes, P < .001, respectively). There was no evidence of a difference in image quality for all prespecified criteria between iNAV and dNAV, although dNAV was the preferred image strategy in two-thirds of cases (17/26, 65%). Quantitative assessment demonstrated that mean fibrosis scores were lower for iNAV compared with dNAV (12 ± 8% versus 20 ± 12%, P < .001). Spatial correspondence between the atrial fibrosis maps was modest (Dice similarity coefficient, 0.43 ± 0.15). Conclusion iNAV-LGE acquisition in individuals with AF was more than twice as fast as dNAV acquisition but resulted in a lower atrial fibrosis score. The differences between these two strategies might impact clinical interpretation. ©RSNA, 2024.

Using Gaussian process for velocity reconstruction after coronary stenosis applicable in positron emission particle tracking: An in-silico study.

Accurate velocity reconstruction is essential for assessing coronary artery disease. We propose a Gaussian process method to reconstruct the velocity profile using the sparse data of the positron emission particle tracking (PEPT) in a biological environment, which allows the measurement of tracer particle velocity to infer fluid velocity fields. We investigated the influence of tracer particle quantity and detection time interval on flow reconstruction accuracy. Three models were used to represent different levels of stenosis and anatomical complexity: a narrowed straight tube, an idealized coronary bifurcation with stenosis, and patient-specific coronary arteries with a stenotic left circumflex artery. Computational fluid dynamics (CFD), particle tracking, and the Gaussian process of kriging were employed to simulate and reconstruct the pulsatile flow field. The study examined the error and uncertainty in velocity profile reconstruction after stenosis by comparing particle-derived flow velocity with the CFD solution. Using 600 particles (15 batches of 40 particles) released in the main coronary artery, the time-averaged error in velocity reconstruction ranged from 13.4% (no occlusion) to 161% (70% occlusion) in patient-specific anatomy. The error in maximum cross-sectional velocity at peak flow was consistently below 10% in all cases. PEPT and kriging tended to overestimate area-averaged velocity in higher occlusion cases but accurately predicted maximum cross-sectional velocity, particularly at peak flow. Kriging was shown to be useful to estimate the maximum velocity after the stenosis in the absence of negative near-wall velocity.

Heart size disparity drives sex-specific response to cardiac resynchronization therapy: a post-hoc analysis of the MORE-MPP CRT trial.

Background: Studies have reported that female sex predicts superior cardiac resynchronization therapy (CRT) response. One theory is that this association is related to smaller female heart size, thus increased "relative dyssynchrony" at given QRS durations (QRSd). Objective: To investigate the mechanisms of sex-specific CRT response relating to heart size, relative dyssynchrony, cardiomyopathy type, QRS morphology, and other patient characteristics. Methods: A post-hoc analysis of the MORE-CRT MPP trial (n=3739, 28% female), with a sub-group analysis of patients with non-ischaemic cardiomyopathy (NICM) and left bundle branch block (LBBB) (n=1308, 41% female) to control for confounding characteristics. A multivariable analysis examined predictors of response to 6 months of conventional CRT, including sex and relative dyssynchrony, measured by QRSd/LVEDV (left ventricular end-diastolic volume). Results: Females had a higher CRT response rate than males (70.1% vs. 56.8%, p<0.0001). Subgroup analysis: Regression analysis of the NICM LBBB subgroup identified QRSd/LVEDV, but not sex, as a modifier of CRT response (p<0.0039). QRSd/LVEDV was significantly higher in females (0.919) versus males (0.708, p<0.001). CRT response was 78% for female patients with QRSd/LVEDV>median value, compared to 68% < median value (p=0.012). Association between CRT response and QRSd/LVEDV was strongest at QRSd<150ms. Conclusions: In the NICM LBBB population, increased relative dyssynchrony in females, who have smaller heart sizes than their male counterparts, is a driver of sex-specific CRT response, particularly at QRSd <150ms. Females may benefit from CRT at a QRSd <130ms, opening the debate on whether sex-specific QRSd cut-offs or QRS/LVEDV measurement should be incorporated into clinical guidelines.

Cost-effectiveness analysis of leadless cardiac resynchronization therapy.

BACKGROUND: The Wireless Stimulation Endocardially for CRT (WiSE-CRT) system is a novel technology used to treat patients with dyssynchronous heart failure (HF) by providing leadless cardiac resynchronization therapy (CRT). Observational studies have demonstrated its safety and efficacy profile, however, the treatment cost-effectiveness has not previously been examined. METHODS: A cost-effectiveness evaluation of the WiSE-CRT System was performed using a cohort-based economic model adopting a "proportion in state" structure. In addition to the primary analysis, scenario analyses and sensitivity analyses were performed to test for uncertainty in input parameters. Outcomes were quantified in terms of quality-adjusted life year (QALY) differences. RESULTS: The primary analysis demonstrated that treatment with the WiSE-CRT system is likely to be cost-effective over a lifetime horizon at a QALY reimbursement threshold of £20 000, with a net monetary benefit (NMB) of £3781 per QALY. Cost-effectiveness declines at time horizons shorter than 10 years. Sensitivity analyses demonstrated that average system battery life had the largest impact on potential cost-effectiveness. CONCLUSION: Within the model limitations, these findings support the use of WiSE-CRT in indicated patients from an economic standpoint. However, improving battery technology should be prioritized to maximize cost-effectiveness in times when health services are under significant financial pressures.

Computational Modelling Enabling In Silico Trials for Cardiac Physiologic Pacing.

Conduction system pacing (CSP) has the potential to achieve physiological-paced activation by pacing the ventricular conduction system. Before CSP is adopted in standard clinical practice, large, randomised, and multi-centre trials are required to investigate CSP safety and efficacy compared to standard biventricular pacing (BVP). Furthermore, there are unanswered questions about pacing thresholds required to achieve optimal pacing delivery while preventing device battery draining, and about which patient groups are more likely to benefit from CSP rather than BVP. In silico studies have been increasingly used to investigate mechanisms underlying changes in cardiac function in response to pathologies and treatment. In the context of CSP, they have been used to improve our understanding of conduction system capture to optimise CSP delivery and battery life, and noninvasively compare different pacing methods on different patient groups. In this review, we discuss the in silico studies published to date investigating different aspects of CSP delivery.

Managing arrhythmia in cardiac resynchronisation therapy.

Arrhythmia is an extremely common finding in patients receiving cardiac resynchronisation therapy (CRT). Despite this, in the majority of randomised trials testing CRT efficacy, patients with a recent history of arrhythmia were excluded. Most of our knowledge into the management of arrhythmia in CRT is therefore based on arrhythmia trials in the heart failure (HF) population, rather than from trials dedicated to the CRT population. However, unique to CRT patients is the aim to reach as close to 100% biventricular pacing (BVP) as possible, with HF outcomes greatly influenced by relatively small changes in pacing percentage. Thus, in comparison to the average HF patient, there is an even greater incentive for controlling arrhythmia, to achieve minimal interference with the effective delivery of BVP. In this review, we examine both atrial and ventricular arrhythmias, addressing their impact on CRT, and discuss the available evidence regarding optimal arrhythmia management in this patient group. We review pharmacological and procedural-based approaches, and lastly explore novel ways of harnessing device data to guide treatment of arrhythmia in CRT.

Guided implantation of a leadless left ventricular endocardial electrode and acoustic transmitter using computed tomography anatomy, dynamic perfusion and mechanics, and predicted activation pattern.

BACKGROUND: The WiSE-CRT System (EBR systems, Sunnyvale, CA) permits leadless left ventricular pacing. Currently, no intraprocedural guidance is used to target optimal electrode placement while simultaneously guiding acoustic transmitter placement in close proximity to the electrode to ensure adequate power delivery. OBJECTIVE: The purpose of this study was to assess the use of computed tomography (CT) anatomy, dynamic perfusion and mechanics, and predicted activation pattern to identify both the optimal electrode and transmitter locations. METHODS: A novel CT protocol was developed using preprocedural imaging and simulation to identify target segments (TSs) for electrode implantation, with late electrical and mechanical activation, with ≥5 mm wall thickness without perfusion defects. Modeling of the acoustic intensity from different transmitter implantation sites to the TSs was used to identify the optimal transmitter location. During implantation, TSs were overlaid on fluoroscopy to guide optimal electrode location that were evaluated by acute hemodynamic response (AHR) by measuring the maximal rate of left ventricular pressure rise with biventricular pacing. RESULTS: Ten patients underwent the implantation procedure. The transmitter could be implanted within the recommended site on the basis of preprocedural analysis in all patients. CT identified a mean of 4.8 ± 3.5 segments per patient with wall thickness < 5 mm. During electrode implantation, biventricular pacing within TSs resulted in a significant improvement in AHR vs non-TSs (25.5% ± 8.8% vs 12.9% ± 8.6%; P < .001). Pacing in CT-identified scar resulted in either failure to capture or minimal AHR improvement. The electrode was targeted to the TSs in all patients and was implanted in the TSs in 80%. CONCLUSION: Preprocedural imaging and modeling data with intraprocedural guidance can successfully guide WiSE-CRT electrode and transmitter implantation to allow optimal AHR and adequate power delivery.

The role of conduction system pacing in patients with atrial fibrillation.

Conduction system pacing (CSP) has emerged as a promising novel delivery method for Cardiac Resynchronisation Therapy (CRT), providing an alternative to conventional biventricular epicardial (BiV) pacing in indicated patients. Despite increasing popularity and widespread uptake, CSP has rarely been specifically examined in patients with atrial fibrillation (AF), a cohort which forms a significant proportion of the heart failure (HF) population. In this review, we first examine the mechanistic evidence for the importance of sinus rhythm (SR) in CSP by allowing adjustment of atrioventricular delays (AVD) to achieve the optimal electrical response, and thus, whether the efficacy of CSP may be significantly attenuated compared to conventional BiV pacing in the presence of AF. We next evaluate the largest clinical body of evidence in this field, related to patients receiving CSP following atrioventricular nodal ablation (AVNA) for AF. Finally, we discuss how future research may be designed to address the vital question of how effective CSP in AF patients is, and the potential hurdles we may face in delivering such studies.

Cell to whole organ global sensitivity analysis on a four-chamber heart electromechanics model using Gaussian processes emulators.

Cardiac pump function arises from a series of highly orchestrated events across multiple scales. Computational electromechanics can encode these events in physics-constrained models. However, the large number of parameters in these models has made the systematic study of the link between cellular, tissue, and organ scale parameters to whole heart physiology challenging. A patient-specific anatomical heart model, or digital twin, was created. Cellular ionic dynamics and contraction were simulated with the Courtemanche-Land and the ToR-ORd-Land models for the atria and the ventricles, respectively. Whole heart contraction was coupled with the circulatory system, simulated with CircAdapt, while accounting for the effect of the pericardium on cardiac motion. The four-chamber electromechanics framework resulted in 117 parameters of interest. The model was broken into five hierarchical sub-models: tissue electrophysiology, ToR-ORd-Land model, Courtemanche-Land model, passive mechanics and CircAdapt. For each sub-model, we trained Gaussian processes emulators (GPEs) that were then used to perform a global sensitivity analysis (GSA) to retain parameters explaining 90% of the total sensitivity for subsequent analysis. We identified 45 out of 117 parameters that were important for whole heart function. We performed a GSA over these 45 parameters and identified the systemic and pulmonary peripheral resistance as being critical parameters for a wide range of volumetric and hemodynamic cardiac indexes across all four chambers. We have shown that GPEs provide a robust method for mapping between cellular properties and clinical measurements. This could be applied to identify parameters that can be calibrated in patient-specific models or digital twins, and to link cellular function to clinical indexes.

A systematic review of cardiac in-silico clinical trials.

Computational models of the heart are now being used to assess the effectiveness and feasibility of interventions through in-silico clinical trials (ISCTs). As the adoption and acceptance of ISCTs increases, best practices for reporting the methodology and analysing the results will emerge. Focusing in the area of cardiology, we aim to evaluate the types of ISCTs, their analysis methods and their reporting standards. To this end, we conducted a systematic review of cardiac ISCTs over the period of 1 January 2012-1 January 2022, following the preferred reporting items for systematic reviews and meta-analysis (PRISMA). We considered cardiac ISCTs of human patient cohorts, and excluded studies of single individuals and those in which models were used to guide a procedure without comparing against a control group. We identified 36 publications that described cardiac ISCTs, with most of the studies coming from the US and the UK. In 75% of the studies, a validation step was performed, although the specific type of validation varied between the studies. ANSYS FLUENT was the most commonly used software in 19% of ISCTs. The specific software used was not reported in 14% of the studies. Unlike clinical trials, we found a lack of consistent reporting of patient demographics, with 28% of the studies not reporting them. Uncertainty quantification was limited, with sensitivity analysis performed in only 19% of the studies. In 97% of the ISCTs, no link was provided to provide easy access to the data or models used in the study. There was no consistent naming of study types with a wide range of studies that could potentially be considered ISCTs. There is a clear need for community agreement on minimal reporting standards on patient demographics, accepted standards for ISCT cohort quality control, uncertainty quantification, and increased model and data sharing.

Mechanoelectric effects in healthy cardiac function and under Left Bundle Branch Block pathology.

Mechanoelectric feedback (MEF) in the heart operates through several mechanisms which serve to regulate cardiac function. Stretch activated channels (SACs) in the myocyte membrane open in response to cell lengthening, while tension generation depends on stretch, shortening velocity, and calcium concentration. How all of these mechanisms interact and their effect on cardiac output is still not fully understood. We sought to gauge the acute importance of the different MEF mechanisms on heart function. An electromechanical computer model of a dog heart was constructed, using a biventricular geometry of 500K tetrahedral elements. To describe cellular behavior, we used a detailed ionic model to which a SAC model and an active tension model, dependent on stretch and shortening velocity and with calcium sensitivity, were added. Ventricular inflow and outflow were connected to the CircAdapt model of cardiovascular circulation. Pressure-volume loops and activation times were used for model validation. Simulations showed that SACs did not affect acute mechanical response, although if their trigger level was decreased sufficiently, they could cause premature excitations. The stretch dependence of tension had a modest effect in reducing the maximum stretch, and stroke volume, while shortening velocity had a much bigger effect on both. MEF served to reduce the heterogeneity in stretch while increasing tension heterogeneity. In the context of left bundle branch block, a decreased SAC trigger level could restore cardiac output by reducing the maximal stretch when compared to cardiac resynchronization therapy. MEF is an important aspect of cardiac function and could potentially mitigate activation problems.

Fiber Organization has Little Effect on Electrical Activation Patterns during Focal Arrhythmias in the Left Atrium.

Over the past two decades there has been a steady trend towards the development of realistic models of cardiac conduction with increasing levels of detail. However, making models more realistic complicates their personalization and use in clinical practice due to limited availability of tissue and cellular scale data. One such limitation is obtaining information about myocardial fiber organization in the clinical setting. In this study, we investigated a chimeric model of the left atrium utilizing clinically derived patient-specific atrial geometry and a realistic, yet foreign for a given patient fiber organization. We discovered that even significant variability of fiber organization had a relatively small effect on the spatio-temporal activation pattern during regular pacing. For a given pacing site, the activation maps were very similar across all fiber organizations tested.

Pacing interventions in non-responders to cardiac resynchronization therapy.

Non-responders to Cardiac Resynchronization Therapy (CRT) represent a high-risk, and difficult to treat population of heart failure patients. Studies have shown that these patients have a lower quality of life and reduced life expectancy compared to those who respond to CRT. Whilst the first-line treatment for dyssynchronous heart failure is "conventional" biventricular epicardial CRT, a range of novel pacing interventions have emerged as potential alternatives. This has raised the question whether these new treatments may be useful as a second-line pacing intervention for treating non-responders, or indeed, whether some patients may benefit from these as a first-line option. In this review, we will examine the current evidence for four pacing interventions in the context of treatment of conventional CRT non-responders: CRT optimization; multisite left ventricular pacing; left ventricular endocardial pacing and conduction system pacing.

Deep learning estimation of three-dimensional left atrial shape from two-chamber and four-chamber cardiac long axis views.

AIMS: Left atrial volume is commonly estimated using the bi-plane area-length method from two-chamber (2CH) and four-chamber (4CH) long axes views. However, this can be inaccurate due to a violation of geometric assumptions. We aimed to develop a deep learning neural network to infer 3D left atrial shape, volume and surface area from 2CH and 4CH views. METHODS AND RESULTS: A 3D UNet was trained and tested using 2CH and 4CH segmentations generated from 3D coronary computed tomography angiography (CCTA) segmentations (n = 1700, with 1400/100/200 cases for training/validating/testing). An independent test dataset from another institution was also evaluated, using cardiac magnetic resonance (CMR) 2CH and 4CH segmentations as input and 3D CCTA segmentations as the ground truth (n = 20). For the 200 test cases generated from CCTA, the network achieved a mean Dice score value of 93.7%, showing excellent 3D shape reconstruction from two views compared with the 3D segmentation Dice of 97.4%. The network also showed significantly lower mean absolute error values of 3.5 mL/4.9 cm2 for LA volume/surface area respectively compared to the area-length method errors of 13.0 mL/34.1 cm2 respectively (P < 0.05 for both). For the independent CMR test set, the network achieved accurate 3D shape estimation (mean Dice score value of 87.4%), and a mean absolute error values of 6.0 mL/5.7 cm2 for left atrial volume/surface area respectively, significantly less than the area-length method errors of 14.2 mL/19.3 cm2 respectively (P < 0.05 for both). CONCLUSIONS: Compared to the bi-plane area-length method, the network showed higher accuracy and robustness for both volume and surface area.