2DE Images Research Articles

Multi-vendor validation of automated left ventricular analysis from 3D echocardiography using artificial intelligence

Abstract Background Accurate echocardiographic assessment of left ventricular (LV) function is dependent on both image acquisition and analysis. Although 3D echocardiography (3DE) is increasingly recommended for volume quantification, manually derived volumes are limited by operator subjectivity, variable image quality, and the lengthy duration of analysis. Consequently, several automated approaches based on artificial intelligence (AI) have been proposed to segment the LV cavity from 3DE and have demonstrated better accuracy and reproducibility compared to human experts. While the utility of AI models in the clinic depends on their ability to generalise to different populations and ultrasound vendors, the external validation of developed models remains limited. Purpose This study evaluates the ability of an existing AI model trained using data from a single vendor to accurately estimate routine 3DE LV indices in an independent external dataset acquired using equipment from a different vendor. Methods A previously published and internally validated deep learning model for segmentation of the LV cavity and myocardium from 3DE was used to automatically derive LV indices including end-diastolic volume (EDV), end-systolic volume (ESV), LV mass, and ejection fraction (EF) [1]. A second independent dataset consisting of paired 3DE and CMR images from 65 marathon runners from a different centre (using equipment from a different vendor) was subsequently used for external validation. Images from both datasets were subjectively scored from 1 (poor) to 5 (excellent) as a measure of image quality. Agreement with CMR as the reference was assessed using a two-way mixed effects intraclass correlation coefficient (ICC), supplemented by Bland-Altman analysis, to determine the model bias (between 3DE and CMR) and 95% limits of agreement (LOA). To assess whether performance could be improved by inclusion of multi-vendor data during training, the initial model was fine-tuned using a small subset (n=9) of images from the second centre. Results The mean image quality score from the second dataset was 2.5 (± 1.1), which was significantly lower than the mean of 3.4 (± 1.0) from the first dataset. Despite vendor and image quality differences, the initial model produced good agreement with CMR in EDV, ESV, and LV mass but only poor agreement in EF (Fig 1). Fine tuning greatly improved agreement in EF with reduced bias and narrower 95% LOA (Fig 2). Conclusion With the addition of a small number of cases for fine-tuning, AI models for automated LV analysis can achieve good performance analysing 3DE images acquired from a different vendor and population. Advancing model generalisability through more sophisticated data augmentation strategies may enable truly vendor-agnostic automated LV analysis from 3DE.

Automated non-invasive quantification of myocardial work using 3D echocardiography and validation with invasive catheter measurements

Abstract Background Myocardial work is an emerging technique for the assessment of left ventricular (LV) performance. Using blood pressure as an afterload determinate offers further insight into LV mechanics than can be detected with load-sensitive parameters of ejection fraction or global longitudinal strain (GLS). Observational studies have signalled benefit in identifying vulnerable ventricles in cardiomyopathy, valvular, and coronary disease. While the reference standard for calculating myocardial work is through catheterisation, this is impractical for routine assessment. Currently, a vendor specific non-invasive index of myocardial work is commercially available, but requires subjective manual analysis and identification of valvular events. With increasing availability of 3D echocardiography (3DE) and automated analysis techniques, there is potential to derive more reproducible estimates of myocardial work. Purpose We sought to test whether global myocardial work could be accurately, non-invasively, and automatically estimated using 3DE, in combination with standard non-invasive clinical measurements. Methods Patients undergoing left heart catheterisation as part of routine care were prospectively recruited, and 3DE was performed within one hour of invasive LV pressure measurement. A validated deep learning model was used to automatically analyse the 3DE image volumes to produce GLS curves, from which a smaller number of features (principal components) were extracted. The invasive pressure traces were separated into cycles, which were each paired with the strain trace that had the minimal difference in R-R interval (and excluded if the relative R-R interval difference exceeded 25%). For each patient, an invasively-derived estimate of myocardial work was calculated as the area enclosed by the median pressure-strain (PS) loop. A multivariate regression model was used to relate a set of routine clinical parameters to the invasively-derived myocardial work, and subsequently assessed using leave-one-out cross-validation. Results A total of 131 out of 153 patients were included for analysis (2 excluded due to R-R interval mismatch, 20 excluded due to poor image or pressure trace quality), with basic demographics presented in Table I. The final parameter set consisted of 6 variables: systolic and diastolic brachial cuff pressures, peak GLS, R-R interval, and two strain curve principal components. When fit into a leave-one-out model, this resulted in a Pearson r-value of 0.86, demonstrating a very high correlation between predicted and invasively-derived myocardial work estimates, with a bias of 0 mmHg% and 95% limits of agreement (LOA) of ± 423 mmHg% (Figure 1). Conclusion This study demonstrates that fully automated and reliable estimates of myocardial work can be derived using 3DE and deep learning combined with routine clinical parameters. Automated and non-invasive approaches may facilitate increased clinical uptake of myocardial work assessment.

Normal reference values of left atrioventricular coupling index on two-dimensional echocardiography

Abstract Background From a physiological background, there is close relationship between left atrial (LA) and left ventricular (LV) expressed as coupling. The assessment of LA/LV coupling may reflect LV dysfunction and differentiate between normal and abnormal LV function. Aim of study The study is designed to sample normal subjects to provide the normal reference values of LA/LV coupling index (LACI) according to age and sex using 2-dimensional echocardiography (2DE). Methods 550 healthy subjects (51.9±17.8 years, 56% males) were included with good 2DE image quality enabled measurement of end-diastolic volumes of both LA and LV. LACI was defined as the ratio between LA to LV end-diastolic volumes. Results An off-line analysis of apical 4-chamber and 2-chamber views to obtain the 2DE measurements was performed within short time (1.5± 0.3 min). Measurement of LACI was obtained in all individuals. Subjects were divided into 3 age groups (G I: 18-40 years, G II: 41-65 years, and GIII: >65 years). As shown in table 1, the normal LACI values in the whole population was 0.17±0.08 with significantly higher in females (0.178±0.09) than males (0.15±0.06); P<0.02. The LACI values tends to be higher in females than males in the subdivided 3 age groups but not statistically significant. In both sexes, increasing age was associated with significant reduction of the mean LVEDV (P<0.0001), significant increase of the mean LAEDV (P= 0.004), and significant increase of LACI (P< 0.0001); especially considering the differences between the youngest and oldest age groups. Conclusions Measurement of LACI is feasible and easily obtainable using 2DE. Age and sex should be considered when defining normal reference values for LACI. The standardization of normal LACI values will help the future studies to assess the changes of these normal values in various cardiac abnormalities for prognostic risk stratification.

CardiacField: computational echocardiography for automated heart function estimation using two-dimensional echocardiography probes

Abstract Aims Accurate heart function estimation is vital for detecting and monitoring cardiovascular diseases. While two-dimensional echocardiography (2DE) is widely accessible and used, it requires specialized training, is prone to inter-observer variability, and lacks comprehensive three-dimensional (3D) information. We introduce CardiacField, a computational echocardiography system using a 2DE probe for precise, automated left ventricular (LV) and right ventricular (RV) ejection fraction (EF) estimations, which is especially easy to use for non-cardiovascular healthcare practitioners. We assess the system’s usability among novice users and evaluate its performance against expert interpretations and advanced deep learning (DL) tools. Methods and results We developed an implicit neural representation network to reconstruct a 3D cardiac volume from sequential multi-view 2DE images, followed by automatic segmentation of LV and RV areas to calculate volume sizes and EF values. Our study involved 127 patients to assess EF estimation accuracy against expert readings and two-dimensional (2D) video-based DL models. A subset of 56 patients was utilized to evaluate image quality and 3D accuracy and another 50 to test usability by novice users and across various ultrasound machines. CardiacField generated a 3D heart from 2D echocardiograms with <2 min processing time. The LVEF predicted by our method had a mean absolute error (MAE) of 2.48%, while the RVEF had an MAE of 2.65%. Conclusion Employing a straightforward apical ring scan with a cost-effective 2DE probe, our method achieves a level of EF accuracy for assessing LV and RV function that is comparable to that of three-dimensional echocardiography probes.

Feasibility of Three-Dimensional Echocardiographic Analysis in Newborns by Novice Observers.

Three-dimensional echocardiography (3DE) provides better interobserver agreement than conventional methods. However, more evidence of whether there is good agreement between novice and experienced observers, especially in pediatric 3DE analysis, is required. We conducted 3DE analysis training in novice observers and investigated the agreement when analyzing 3DE images between novice and experienced observers. One experienced and 4 novice observers independently analyzed 60 3DE images obtained from neonates. The left and right ventricular end-diastolic volume (LVEDV and RVEDV), end-systolic volume (LVESV and RVESV), ejection fraction (LVEF and RVEF), left ventricular global longitudinal strain, and global circumferential strain (LVGLS and LVGCS) were calculated. The novices received hands-on instruction in the analysis procedure before the analysis and received further feedback after their first 40 analyses. Agreement between the novices and theexperienced observer was evaluated by the intra-class correlation coefficient (ICC) and percentage difference in 3 groups of 20 images each (images 1-20, 21-40, and 41-60). The ICC of LVEDV was > 0.85 from the first 20 images and increased with experience. The ICCs for RVEDV and RVESV were low from the first 20 images but increased linearly, reaching an ICC > 0.9 in 3 of the 4 novice observers in the last 20 images. The range of the percentage difference was small for LVEDV, RVEDV, LVEF, and RVEF. Novices show an experienced level of 3DE analysis capability in LVEDV and RVEDV after analyzing 40 to 60 images. Training of novices is effective for analyzing LVEDV and RVEDV in 3DE images.

Use of two- and three-dimensional echocardiography for assessment of the left ventricular outflow tract and aortic orifice areas in dogs

Introduction/objectivesIn clinical practice, dogs are screened for subaortic stenosis (SAS) using two-dimensional (2DE) and Doppler echocardiography. There is no accepted antemortem diagnostic criterion to distinguish between mild SAS and unaffected, therefore additional means of evaluating the left ventricular outflow tract (LVOT) and aorta may be desirable. This study sought to determine and compare LVOT and aortic orifice areas using 2DE and three-dimensional echocardiography (3DE) in apparently healthy dogs of various breeds and somatotypes. Animals, materials, and methodsSixty-nine healthy, privately-owned dogs. The LVOT and aortic orifice areas were determined using 2DE aortic valve (AV) diameter-derived area; the continuity equation (CE); and 3DE planimetry of the LVOT, AV, sinus of Valsalva, and sinotubular junction. Orifice areas were indexed to body surface area (BSA). ResultsObtaining 3DE images and performing planimetry were feasible in all dogs. The mean indexed area measured using the 2DE AV diameter (2.85 cm2/m2) was significantly lower than that derived from 3DE AV planimetry (3.85 cm2/m2; mean difference, 1.00 cm2/m2; P<0.001). There was poor agreement between the effective area calculated using the CE and the anatomic areas calculated using 2DE AV diameter and 3DE planimetry. The area calculated using the CE was less than all other calculations of area. Interobserver and intraobserver repeatability and reproducibility for 3DE planimetry were excellent. ConclusionsMethods for determining aortic orifice areas in dogs are not interchangeable, and this must be taken into account if these methods are investigated in the evaluation of dogs with SAS in the future.

Role and application of three-dimensional transthoracic echocardiography in the assessment of left and right ventricular volumes and ejection fraction: a UK nationwide survey

Three-dimensional echocardiography (3DE) imaging has permitted advancements in the quantification of left ventricular (LV) and right ventricular (RV) volumes and ejection fraction. We evaluated the availability of 3DE equipment / analysis software, the integration of 3DE assessment of the LV and RV in routine clinical practice, current training provisions in 3DE, and aimed to ascertain barriers preventing the routine use of 3DE for volumetric analysis. Through the British Society of Echocardiography (BSE) regional representatives’ network, echocardiographers were invited to participate in an open online survey. A total of 181 participants from echocardiography departments in the United Kingdom (UK), the majority from tertiary centres (61%), completed the 28-question survey. For 3DE quantification, 3DE-LV was adopted more frequently than 3DE-RV (48% vs 11%, respectively). Imaging feasibility was a recognised factor in 3DE RV and LV adoption. Many respondents had access to 3D probes (93%). The largest observed barriers to 3DE routine use were training deficiencies, with 83% reporting they would benefit from additional training opportunities and the duration of time permitted for the scan, with 68% of responders reporting allowances of less than the BSE standard of 45–60 min per patient (8% < 30-min). Furthermore, of those respondents who had undertaken professional accreditation, competence in 3DE was not formally assessed in 89%. This UK survey also reported good accessibility to magnetic resonance imaging (72%), which was related to overall 3DE adoption. In summary, although 3DE is now readily available, it remains underutilised. Further training opportunities, integrated formal assessment, improved adoption of BSE minimum recommended scanning times, alongside industry and societal support, may increase 3DE utilisation in routine practice.

Three-dimensional printing of mitral valve models using echocardiographic data improves the knowledge of cardiology fellow physicians in training.

High fidelity three-dimensional Mitral valve models (3D MVM) printed from echocardiography are currently being used in preparation for surgical repair. We hypothesize that printed 3DMVM could have relevance to cardiologists in training by improving their understanding of normal anatomy and pathology. Sixteen fellow physicians in pediatric and adult cardiology training were recruited. 3D echocardiography (3DE) video clips of six mitral valves (one normal and five pathological) were displayed and the fellows were asked to name the prolapsing segments in each. Following that, three still images of 3D MVMs in different projections: enface, profile and tilted corresponding to the same MVs seen in the clip were presented on a screen. Participating physicians were presented with a comprehensive questionnaire aimed at assessing whether the 3D MVM has improved their understanding of valvular anatomy. Finally, a printed 3D MVM of each of the valves was handed out, and the same questionnaire was re-administered to identify any further improvement in the participants' perception of the anatomy. The correct diagnosis using the echocardiography video clip of the Mitral valve was attained by 45% of the study participants. Both pediatric and adult trainees, regardless of the year of training demonstrated improved understanding of the anatomy of MV after observing the corresponding model image. Significant improvement in their understanding was noted after participants had seen and physically examined the printed model. Printed 3D MVM has a beneficial impact on the cardiology trainees' understanding of MV anatomy and pathology compared to 3DE images.

Real-time Three-dimensional Echocardiography: A Feasible and Accurate Method for Measuring Right Ventricular Function in Pulmonary Hypertension

Right ventricular function is independently associated with poor clinical outcomes in patients with various cardiovascular diseases including pulmonary hypertension (PH). The complex geometry and mode of contractility means the right ventricle (RV) inlet and outlet are not in line; hence, 2-dimensional echocardiography fails to accurately quantify RV volumes and function in such patients. Three-dimensional echocardiography (3DE) allows for reliable and reproducible quantification of RV volumes and function by overcoming these limitations of conventional echocardiography. This review focuses on the 3DE assessment of RV function in patients with PH and discusses the following points: (1) acquiring an RV data set for 3DE imaging, including details of all available 3DE systems and software utilized in daily practice; (2) the reliability and feasibility of RV remodeling measured with 3DE with different modalities in patients with PH; and (3) the prognostic value of 3DE-derived RV function in such patients.

Right ventricular strain by speckle tracking echocardiography in patients with systemic scleroderma and absence of pulmonary hypertension

Abstract Funding Acknowledgements Type of funding sources: None. Introduction In patients with scleroderma the morphological remodeling and alteration of right ventricular (RV) function begins even before the setting of pulmonary artery hypertension (PAH). These alterations are associated with worse outcomes and high mortality rate as 30% of deaths are attributed to cardiovascular causes followed by respiratory causes (17%). Purpose Early detection of RV dysfunction is possible using advanced imaging techniques and can lead to specific therapeutic management before installation of PAH and clinical signs and symptoms of heart failure . Purpose of this study is to evaluate new indices, based on speckle tracking echocardiography, for identification of subclinical RV systolic failure. Methods Thirty (30) scleroderma patients without any signs or symptoms of PAH and forty (40) healthy subjects were included in the present study. Two-dimensional echocardiography (2DE) and two-dimensional speckle-tracking echocardiography (2DSTE) were performed. All the usual 2DE, Doppler and tissue Doppler imaging measurements were obtained for the evaluation of RV function. Moreover, TAPSE, maximum systolic myocardial velocity from TDI of lateral tricuspid valve annulus (SRV), auto RV global longitudinal strain (aRVGLS) and auto RV free wall longitudinal strain (aRVFWLS) by 2DSTE were measured. aRVGLS and aRVFWLS are new indices, with appropriate RV function echo logistics in new technology echo systems, independent from known manual measurements. Results In scleroderma patients significantly higher systolic pulmonary artery pressure was recorded, although it remained within normal limits in both groups (24.6+-5.4mmHg vs 19.2+-5.9 mmHg) . Scleroderma patients had significant decreased RV systolic function with significantly lower TAPSE and SRV. Both aRVGLS and aRVFWLS were also lower in scleroderma patients compared to the control group (−18.6+-2.7 vs −23.1+-2.3 % and −25.4+-3.1 vs −26.4+-2.2 % respectively). aRVGLS was the best significant predictor of early RV systolic impairment in patients with scleroderma.The difference in aRVFWLS was not significant. Conclusions Scleroderma is a severe systematic disease associated with high mortality rates. RV remodeling, dysfunction, and progressive pulmonary artery hypertension are among the most important internal organ manifestations of this disease. Early detection of RV dysfunction before installation of pulmonary hypertension is possible by 2DSTE, and may be helpful to guide further therapy in these patients.

Machine Learning the Right Heart: Leveraging CMR for RV Volume Quantification From 3D Echocardiography

Right ventricular (RV) volume indices are important prognostic markers in several cardiac diseases. While cardiac magnetic resonance (CMR) imaging remains the gold-standard for volume quantification, echocardiography is more accessible. Unlike two-dimensional echocardiography (2DE), three-dimensional echocardiography (3DE) enables volume quantification without geometric assumptions. However, manual RV segmentation from 3DE is slow and hindered by speckle noise and poor spatial resolution. Machine learning can be utilised to overcome these challenges for efficient and accurate 3DE RV assessment. Thirty participants (18 with cardiac disease, 12 healthy controls) were imaged with transthoracic 3DE and cine CMR <1 hour apart. The RV segmentations from CMR were spatiotemporally aligned to the corresponding 3DE for each participant across one cardiac cycle. Paired 3DE images and CMR labels were used to train a deep-learning model (80/20 training/testing split) that predicts segmentations from 3DE only. Resulting end-diastolic volume (EDV), end-systolic volume (ESV), and ejection fraction (EF) were compared with CMR for each case. Machine learning segmentations predicted lower EDV (–17±21 mL), ESV (–8±24 mL), and EF (–3±9%) compared with CMR, but there were no statistically significant differences (p>0.05, n=6). These biases were smaller than those previously found for manual 3DE RV segmentations (–33±25 mL, –30±20 mL, and 6±10%). Machine learning enabled accurate RV volume quantification from 3DE, which was not possible with 2DE due to complex RV geometry. Automated segmentations of 3DE produced comparable volumes to gold-standard CMR, which could facilitate the use of 3DE for clinical RV assessment.

3D echocardiography in mitral valve prolapse.

Mitral valve prolapse (MVP) is the leading cause of mitral valve surgery. Echocardiography is the principal imaging modality used to diagnose MVP, assess the mitral valve morphology and mitral annulus dynamics, and quantify mitral regurgitation. Three-dimensional (3D) echocardiographic (3DE) imaging represents a consistent innovation in cardiovascular ultrasound in the last decades, and it has been implemented in routine clinical practice for the evaluation of mitral valve diseases. The focus of this review is the role and the advantages of 3DE in the comprehensive evaluation of MVP, intraoperative and intraprocedural monitoring.

MITEA: A dataset for machine learning segmentation of the left ventricle in 3D echocardiography using subject-specific labels from cardiac magnetic resonance imaging.

Segmentation of the left ventricle (LV) in echocardiography is an important task for the quantification of volume and mass in heart disease. Continuing advances in echocardiography have extended imaging capabilities into the 3D domain, subsequently overcoming the geometric assumptions associated with conventional 2D acquisitions. Nevertheless, the analysis of 3D echocardiography (3DE) poses several challenges associated with limited spatial resolution, poor contrast-to-noise ratio, complex noise characteristics, and image anisotropy. To develop automated methods for 3DE analysis, a sufficiently large, labeled dataset is typically required. However, ground truth segmentations have historically been difficult to obtain due to the high inter-observer variability associated with manual analysis. We address this lack of expert consensus by registering labels derived from higher-resolution subject-specific cardiac magnetic resonance (CMR) images, producing 536 annotated 3DE images from 143 human subjects (10 of which were excluded). This heterogeneous population consists of healthy controls and patients with cardiac disease, across a range of demographics. To demonstrate the utility of such a dataset, a state-of-the-art, self-configuring deep learning network for semantic segmentation was employed for automated 3DE analysis. Using the proposed dataset for training, the network produced measurement biases of -9 ± 16 ml, -1 ± 10 ml, -2 ± 5 %, and 5 ± 23 g, for end-diastolic volume, end-systolic volume, ejection fraction, and mass, respectively, outperforming an expert human observer in terms of accuracy as well as scan-rescan reproducibility. As part of the Cardiac Atlas Project, we present here a large, publicly available 3DE dataset with ground truth labels that leverage the higher resolution and contrast of CMR, to provide a new benchmark for automated 3DE analysis. Such an approach not only reduces the effect of observer-specific bias present in manual 3DE annotations, but also enables the development of analysis techniques which exhibit better agreement with CMR compared to conventional methods. This represents an important step for enabling more efficient and accurate diagnostic and prognostic information to be obtained from echocardiography.

An automated method for BRISQUE quantification of image quality in echocardiography

Abstract Background Echocardiography (echo) remains the most widely used imaging modality for the assessment, monitoring, and prognostication of the heart. Despite its prevalence, standardisation efforts for echo chamber quantification are ongoing, with challenges owing to subjectivity during acquisition and analysis. Furthermore, the confidence in derived functional indices is often dependent on the quality of the acquired images. However, few studies have investigated the accuracy of echo measurements compared to a reference modality such as cardiac magnetic resonance (CMR) imaging, when stratified by image quality. Purpose To develop an objective and automated method to quantify echo image quality, and subsequently to investigate the relationship between image quality and patient demographics, as well as the magnitude of bias in left ventricular (LV) functional indices compared with CMR. Methods Transthoracic apical 2D echo (2DE) and 3D echo (3DE) data from 128 participants (72 healthy controls and 56 patients with acquired heart disease) were used to train a BRISQUE (Blind/Referenceless Image Spatial Quality Evaluator) algorithm [1]. Briefly, feature extraction was performed by fitting pixel luminances to a generalised Gaussian distribution (GGD), followed by support vector regression to correlate features (i.e., shape, variance, and mean parameters of the GGD) to quality scores (Fig. 1). Independent BRISQUE models were trained on 580 2DE images (consisting of 2-, 3-, and 4-chamber views) and 128 targeted LV 3DE acquisitions at end-diastole, each assigned a subjective perceived quality score between 1 (poor) and 9 (excellent) by a single observer. LV indices including end-diastolic volume (EDV), end-systolic volume (ESV), ejection fraction (EF), and global longitudinal strain (GLS), were assessed according to standard guidelines. Resultant BRISQUE scores were plotted against patient demographics (age, height, weight) and the measurement bias by comparison to CMR (acquired within 1 hour of echo). Results Several linear relationships (where P-value of slope &amp;lt;0.05) were observed between demographics, cardiac indices, and BRISQUE scores. Increasing patient weight (and height in 3DE) were found to be associated with poorer image quality. There was no apparent relationship between image quality and age. Of interest, EF exhibited a relationship with image quality in both 2DE and 3DE (Fig. 2), whereby higher quality images tended to overestimate EF, while lower quality images underestimated EF. For 3DE, image quality dependency was also observed for ESV and GLS biases. Conclusions BRISQUE can objectively quantify image quality to produce scores which correlate to those of an expert observer, with potential utility for the standardised quantification of echo image quality. Using this method, it may be possible to predict patient characteristics which adversely impact echo quality, as well as the magnitude of measurement biases for certain functional indices. Funding Acknowledgement Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Health Research Council (HRC) of New Zealand; National Heart Foundation (NHF) of New Zealand

Abnormal echocardiographic findings after COVID-19 infection: a multicenter registry

Abstract Background The Coronavirus Disease 2019 (COVID-19) pandemic has transformed health systems worldwide. There is conflicting data regarding the degree of cardiovascular involvement following infection, generating uncertainty in patients and an additional healthcare burden with increased diagnostic testing. A registry was designed to evaluate the prevalence of echocardiographic abnormalities in Latin American adults recovered from COVID-19. Methods We prospectively evaluated 595 participants (mean age 45.5±14.9 years; 50.8% female) from 10 institutions in Argentina and Brazil. Echocardiographic studies were conducted with General Electric equipment; 2DE imaging and global longitudinal strain (GLS) of both ventricles were performed. Comparisons between groups were made with Chi-square, Fisher and Student's t-test. Logistic regression was performed to determine variables associated with abnormal echocardiogram findings. Results A total of 61.7% of the participants denied any relevant cardiovascular medical history. Table 1 summarizes the comorbidities of the included patients. The majority of patients (82.5%) had the disease at home or in an out-of-hospital center. Of the patients who required hospitalization, 15.3% were in a general ward, 1.9% in intensive care and 0.3% required mechanical ventilation during the disease. The median time between infection and performance of the echocardiographic study was two months (IQR 1–3 months). Among patients who reported symptoms following COVID-19 recovery (41.8%), the most frequently reported was dyspnea (47.4%), followed by mild symptoms such as asthenia, arterial hypertension or palpitations (32.9%), 12.9% referred chest pain, 6% of patients reported dyspnea and chest pain, and 0.8% reported various other symptoms. The mean left ventricular ejection fraction (LVEF) was 61.0±5.5% and the mean left atrial volume was 33.1±13.2 ml/m2. In patients without prior comorbidities, 8.2% had some echocardiographic abnormality (Figure 1). We found no significant differences in LVEF between symptomatic and asymptomatic patients (61.4% versus 60.6% respectively, p=0.104). Symptomatic patients showed slightly reduced GLS (−20.3% versus −20.9%, p=0.012) with a trend in the same direction in the RV free wall GLS (−25.6% versus −26.3%, p=0.103). Male patients were more likely to have any new echocardiographic abnormalities (OR 2.82, p=0.002). Time elapsed since infection resolution (p=0.245), the presence of symptoms (p=0.927), or history of hospitalization during infection (p=0.671) did not have any correlation with echocardiographic abnormalities. The difference between sexes remains unchanged after adjusting for left atrial volume, wall thicknesses, diastolic function and abnormal wall motion. Conclusion Our results suggest that cardiovascular abnormalities after COVID-19 infection are rare and usually mild, especially in cases of mild disease. These abnormalities may be more frequent among males Funding Acknowledgement Type of funding sources: None.

Right ventricular quantification using 3D echocardiography: a comparison with CMR

Abstract Background Volumetric and functional right ventricular (RV) indices such as ejection fraction (EF) and global strains are known independent predictors of adverse cardiovascular events. While cardiac magnetic resonance (CMR) imaging remains the reference standard for volume quantification, echocardiography is more accessible and allows for rapid ventricular assessment. Compared to conventional 2D echocardiography, 3D echocardiography (3DE) enables full volume acquisitions and the ability to circumvent geometric assumptions. Given the complexity of RV geometry and sensitivity to image plane positioning, this advantage offers the potential to obtain more accurate diagnostic measurements. Purpose Tools for RV analysis in 3DE have been less extensively studied compared to those for the left ventricle (LV). We sought to quantify discrepancies in RV indices derived from 3DE and CMR. Methods Transthoracic real-time 3DE and cine CMR imaging were performed in 20 prospectively recruited participants (12 patients with acquired cardiac disease and 8 healthy controls), &amp;lt;1 hour apart. Dynamic 3D biventricular models were constructed semi-automatically from CMR by identifying fiducial landmarks, correcting in-plane breath-hold mis-registrations, and interactively fitting contours to the endocardial and epicardial borders on long- and short-axis slices. For 3DE, right ventricular endocardial models were created by fitting contours on 2D image planes resampled from the 3D volume at end-diastole and end-systole, which were subsequently tracked over one cardiac cycle (Figure 1). RV indices including end-diastolic volume (EDV), end-systolic volume (ESV), EF, global longitudinal strain (GLS), and global circumferential strain (GCS) were calculated from the 3DE- and CMR-derived 3D geometric models and compared. Paired-sample t-tests were performed to identify statistically significant differences (where P&amp;lt;0.05), and intraclass correlation coefficients (ICC) for absolute agreement were computed to assess the reliability for each measurement. Results Differences (mean ± SD) in RV indices between 3DE and CMR, with corresponding ICCs are presented in Table 1. Statistically significant differences in RV EDV, ESV, EF, and GLS were observed, with 3DE consistently underestimating volumes and overestimating function when compared to CMR. Although a statistically significant difference in RV GCS was not observed, a low ICC score indicated poor reliability. Conclusions Volume underestimation in RV indices between 3DE and CMR were found to be larger than those previously reported for the LV, which is likely due to the increased geometric complexity and surface area to volume ratio for the RV. Moreover, 3DE tends to overestimate RV function in terms of EF and GLS, which may impact treatment pathways if used in a clinical setting. Recognising systematic differences between modalities reinforces the need to further develop 3DE technologies for more accurate RV quantification. Funding Acknowledgement Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Health Research Council (HRC) of New Zealand;National Heart Foundation (NHF) of New Zealand

Automated right ventricular strain by speckle tracking echocardiography in patients with systemic scleroderma

Abstract Introduction In patients with scleroderma the morphological remodeling and alteration of right ventricular (RV) function begins even before the setting of pulmonary artery hypertension (PAH). These alterations are associated with worse outcomes and high mortality. Purpose Early detection of RV dysfunction is possible using advanced imaging techniques and can lead to specific therapeutic measures before installation of PAH. Purpose of this study is to evaluate new indices, based on speckle tracking technology, for identification of subclinical RV systolic failure. Methods Twenty-five (25) scleroderma patients without any signs or symptoms of PAH and twenty-five (25) healthy subjects were included in the present study. Two-dimensional echocardiography (2DE) and two-dimensional speckle-tracking echocardiography (2DSTE) were performed. All the usual 2DE, Doppler and tissue Doppler imaging measurements were obtained for the evaluation of RV function. Moreover, TAPSE, maximum systolic myocardial velocity from TDI of lateral tricuspid valve annulus (SRV), auto RV global longitudinal strain (aRVGLS) and auto RV free wall longitudinal strain (aRVFWLS) by 2DSTE were measured. aRVGLS is a new index, with appropriate RV function echo logistics in new technology echo systems, independent from known manual measurements. Results In scleroderma patients significantly higher systolic pulmonary artery pressure was recorded, although it remained within normal limits in both groups. Scleroderma patients had significant decreased RV systolic function with significantly lower TAPSE and SRV. Both aRVGLS and aRVFWLS were also significantly lower in scleroderma patients compared to the control group. aRVGLS was the best predictor of early RV systolic impairment in patients with scleroderma. Conclusions Scleroderma is a severe systematic disease associated with high mortality rates. RV remodeling, dysfunction, and progressive pulmonary artery hypertension are among the most important internal organ manifestations of this disease. Early detection of RV dysfunction is possible by 2DSTE, and may be helpful to guide further therapy in these patients. Funding Acknowledgement Type of funding sources: None.

An automated method for BRISQUE quantification of image quality in echocardiography

Abstract Background Echocardiography (echo) remains the most widely used imaging modality for the assessment, monitoring, and prognostication of the heart. Despite its prevalence, standardisation efforts for echo chamber quantification are ongoing, with challenges owing to subjectivity during acquisition and analysis. Furthermore, the confidence in derived functional indices is often dependent on the quality of the acquired images. However, few studies have investigated the accuracy of echo measurements compared to a reference modality such as cardiac magnetic resonance (CMR) imaging, when stratified by image quality. Purpose To develop an objective and automated method to quantify echo image quality, and subsequently to investigate the relationship between image quality and patient demographics, as well as the magnitude of bias in left ventricular (LV) functional indices compared with CMR. Methods Transthoracic apical 2D echo (2DE) and 3D echo (3DE) data from 128 participants (72 healthy controls and 56 patients with acquired heart disease) were used to train a BRISQUE (Blind/Referenceless Image Spatial Quality Evaluator) algorithm [1]. Briefly, feature extraction was performed by fitting pixel luminances to a generalised Gaussian distribution (GGD), followed by support vector regression to correlate features (i.e., shape, variance, and mean parameters of the GGD) to quality scores (Fig. 1). Independent BRISQUE models were trained on 580 2DE images (consisting of 2-, 3-, and 4-chamber views) and 128 targeted LV 3DE acquisitions at end-diastole, each assigned a subjective perceived quality score between 1 (poor) and 9 (excellent) by a single observer. LV indices including end-diastolic volume (EDV), end-systolic volume (ESV), ejection fraction (EF), and global longitudinal strain (GLS), were assessed according to standard guidelines. Resultant BRISQUE scores were plotted against patient demographics (age, height, weight) and the measurement bias by comparison to CMR (acquired within 1 hour of echo). Results Several linear relationships (where P-value of slope &amp;lt;0.05) were observed between demographics, cardiac indices, and BRISQUE scores. Increasing patient weight (and height in 3DE) were found to be associated with poorer image quality. There was no apparent relationship between image quality and age. Of interest, EF exhibited a relationship with image quality in both 2DE and 3DE (Fig. 2), whereby higher quality images tended to overestimate EF, while lower quality images underestimated EF. For 3DE, image quality dependency was also observed for ESV and GLS biases. Conclusions BRISQUE can objectively quantify image quality to produce scores which correlate to those of an expert observer, with potential utility for the standardised quantification of echo image quality. Using this method, it may be possible to predict patient characteristics which adversely impact echo quality, as well as the magnitude of measurement biases for certain functional indices. Funding Acknowledgement Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Health Research Council (HRC) of New Zealand; National Heart Foundation (NHF) of New Zealand

SlicerHeart: An open-source computing platform for cardiac image analysis and modeling.

Cardiovascular disease is a significant cause of morbidity and mortality in the developed world. 3D imaging of the heart's structure is critical to the understanding and treatment of cardiovascular disease. However, open-source tools for image analysis of cardiac images, particularly 3D echocardiographic (3DE) data, are limited. We describe the rationale, development, implementation, and application of SlicerHeart, a cardiac-focused toolkit for image analysis built upon 3D Slicer, an open-source image computing platform. We designed and implemented multiple Python scripted modules within 3D Slicer to import, register, and view 3DE data, including new code to volume render and crop 3DE. In addition, we developed dedicated workflows for the modeling and quantitative analysis of multi-modality image-derived heart models, including heart valves. Finally, we created and integrated new functionality to facilitate the planning of cardiac interventions and surgery. We demonstrate application of SlicerHeart to a diverse range of cardiovascular modeling and simulation including volume rendering of 3DE images, mitral valve modeling, transcatheter device modeling, and planning of complex surgical intervention such as cardiac baffle creation. SlicerHeart is an evolving open-source image processing platform based on 3D Slicer initiated to support the investigation and treatment of congenital heart disease. The technology in SlicerHeart provides a robust foundation for 3D image-based investigation in cardiovascular medicine.

Leveraging CMR for 3D echocardiography: an annotated multimodality dataset for AI

Abstract Funding Acknowledgements: Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Health Research Council of New Zealand (HRC) National Heart Foundation of New Zealand (NHF) Segmentation of the left ventricular myocardium and cavity in 3D echocardiography (3DE) is a critical task for the quantification of systolic function in heart disease. Continuing advances in 3DE have considerably improved image quality, prompting increased clinical uptake in recent years, particularly for volumetric measurements. Nevertheless, analysis of 3DE remains a difficult problem due to inherently complex noise characteristics, anisotropic image resolution, and regions of acoustic dropout. One of the primary challenges associated with the development of automated methods for 3DE analysis is the requirement of a sufficiently large training dataset. Historically, ground truth annotations have been difficult to obtain due to the high degree of inter- and intra-observer variability associated with manual 3DE segmentation, thus, limiting the scope of AI-based solutions. To address the lack of expert consensus, we instead used labels derived from cardiac magnetic resonance (CMR) images of the same subjects. By spatiotemporally registering CMR labels to corresponding 3DE image data on a per subject basis (Figure 1), we collated 520 annotated 3DE images from a mixed cohort of 130 human subjects (2 independent single-beat acquisitions per subject at end-diastole and end-systole) consisting of healthy controls and patients with acquired cardiac disease. Comprising images acquired across a range of patient demographics, this curated dataset exhibits variation in image quality, 3DE acquisition parameters, as well as left ventricular shape and pose within the 3D image volume. To demonstrate the utility of such a dataset, nn-UNet, a self-configuring deep learning method for semantic segmentation was employed. An 80/20 split of the dataset was used for training and testing, respectively, and data augmentations were applied in the form of scaling, rotation, and reflection. The trained network was capable of reproducing measurements derived from CMR for end-diastolic volume, end-systolic volume, ejection fraction, and mass, while outperforming an expert human observer in terms of accuracy as well as scan-rescan reproducibility (Table I). As part of ongoing efforts to improve the accuracy and efficiency of 3DE analysis, we have leveraged the high resolution and signal-to-noise-ratio of CMR (relative to 3DE), to create a novel, publicly available benchmark dataset for developing and evaluating 3DE labelling methods. This approach not only significantly reduces the effects of observer-specific bias and variability in training data arising from conventional manual 3DE analysis methods, but also improves the agreement between cardiac indices derived from 3DE and CMR. Figure 1. Data annotation workflow Table I. Results