Introducción: El número creciente de estudios ecocardiográficos y la necesidad de cumplir rigurosamente con las recomendaciones de guías internacionales de cuantificación, ha llevado a que los cardiólogos deban realizar tareas sumamente extensas y repetitivas, como parte de la interpretación y análisis de cantidades de información cada vez más abrumadoras.Novedosas técnicas de machine learning (ML), diseñadas para reconocer imágenes y realizar mediciones en las vistas adecuadas, están siendo cada vez más utilizadas para responder a esta necesidad evidente de automatización de procesos.Objetivo: Nuestro objetivo fue evaluar un modelo alternativo de interpretación y análisis de estudios ecocardiográficos, basado fundamentalmente en la utilización de software de ML, capaz de identificar y clasificar vistas y realizar mediciones estandarizadas de forma automática.Material y métodos: Se utilizaron imágenes obtenidas en 2000 sujetos normales, libres de enfermedad, de los cuales 1800 fueron utilizados para desarrollar los algoritmos de ML y 200 para su validación posterior.Primero, una red neuronal convolucional fue desarrollada para reconocer 18 vistas ecocardiográficas estándar y clasificarlas de acuerdo con 8 grupos (stacks) temáticos.Los resultados de la identificación automática fueron comparados con la clasificación realizada por expertos.Luego, algoritmos de ML fueron desarrollados para medir automáticamente 16 parámetros de eco Doppler de evaluación clínica habitual, los cuales fueron comparados con las mediciones realizadas por un lector experto.Finalmente, comparamos el tiempo necesario para completar el análisis de un estudio ecocardiográfico con la utilización de métodos manuales convencionales, con el tiempo necesario con el empleo del modelo que incorpora ML en la clasificación de imágenes y mediciones ecocardiográficas iniciales.La variabilidad inter e intraobservador también fue analizada.Resultados: La clasificación automática de vistas fue posible en menos de 1 segundo por estudio, con una precisión de 90 % en imágenes 2D y de 94 % en imágenes Doppler.La agrupación de imágenes en stacks tuvo una precisión de 91 %, y fue posible completar dichos grupos con las imágenes necesarias en 99% de los casos.La concordancia con expertos fue excelente, con diferencias similares a las observadas entre dos lectores humanos.La incorporación de ML en la clasificación y medición de imágenes ecocardiográficas redujo un 41 % el tiempo de análisis y demostró menor variabilidad que la metodología de interpretación convencional.Conclusión: La incorporación de técnicas de ML puede mejorar significativamente la reproducibilidad y eficiencia de las interpretaciones y mediciones ecocardiográficas.La implementación

Background: The growing number of echocardiographic tests and the need for strict adherence to international quantification guidelines have forced cardiologists to perform highly extended and repetitive tasks when interpreting and analyzing increasingly overwhelming amounts of data. Novel machine learning (ML) techniques, designed to identify images and perform measurements at relevant visits, are becoming more common to meet this obvious need for process automation. Objective: Our objective was to evaluate an alternative model for the interpretation and analysis of echocardiographic tests mostly based on the use of ML software in order to identify and classify views and perform standardized measurements automatically. Methods: Images came from 2000 healthy subjects, 1800 of whom were used to develop ML algorithms and 200 for subsequent validation. First, a convolutional neural network was developed in order to identify 18 standard echocardiographic views and classify them based on 8 thematic groups (stacks). The results of automatic identification were compared to classification by experts. Later, ML algorithms were developed to automatically measure 16 Doppler scan parameters for regular clinical evaluation, which were compared to measurements by an expert reader. Finally, we compared the time required to complete the analysis of an echocardiographic test using conventional manual methods with the time needed when using the ML model to classify images and perform initial echocardiographic measurements. Inter- and intra-observer variability was also analyzed. Results: Automatic view classification was possible in less than 1 second per test, with a 90% accuracy for 2D images and a 94% accuracy for Doppler scan images. Stacking images had a 91% accuracy, and it was possible to complete the groups with any necessary images in 99% of cases. Expert agreement was outstanding, with discrepancies similar to those found between two human readers. Applying ML to echocardiographic imaging classification and measurement reduced time of analysis by 41% and showed lower variability than conventional reading methods. Conclusion: Application of ML techniques may significantly improve reproducibility and efficiency of echocardiographic interpretations and measurements. Using this type of technologies in clinical practice may lead to reduced costs and increased medical staff satisfaction. How to cite this article: Cotella JI, Adettia K, Miyoshi T, Kebed K, Blitz A, Schreckenberger M, y cols. Optimizing Echocardiograms Interpretation with the Application of Machine Learning in the WASE Study. Rev Argent Cardiol 2024;92:5-14. http://dx.doi.org/10.7775/rac.v92.i1.20723

BACKGROUND:Left ventricular (LV) circumferential strain has received less attention than longitudinal deformation, which has recently become part of routine clinical practice. Among other reasons, this is because of the lack of established normal values. Accordingly, the aim of this study was to establish normative values for LV circumferential strain and determine sex-, age-, and race-related differences in a large cohort of healthy adults. METHODS:Complete two-dimensional transthoracic echocardiograms were obtained in 1,572 healthy subjects (51% men), enrolled in the World Alliance Societies of Echocardiography Normal Values Study. Subjects were divided into three age groups (<35, 35-55, and >55 years) and stratified by sex and by race. Vendor-independent semiautomated speckle-tracking software was used to determine LV regional circumferential strain and global circumferential strain (GCS) values. Limits of normal for each measurement were defined as 95% of the corresponding sex and age group falling between the 2.5th and 97.5th percentiles. Intergroup differences were analyzed using unpaired t tests. RESULTS:Circumferential strain showed a gradient, with lower magnitude at the mitral valve level, increasing progressively toward the apex. Compared with men, women had statistically higher magnitudes of regional and global strain. Older age was associated with a stepwise increase in GCS despite an unaffected ejection fraction, a decrease in LV volume, and relatively stable global longitudinal strain in men, with a small gradual decrease in women. Asian subjects demonstrated significantly higher GCS magnitudes than whites of both sexes and blacks among women only. In contrast, no significant differences in GCS were found between white and black subjects of either sex. Importantly, despite statistical significance of these differences across sex, age, and race, circumferential strain values were similar in all groups, with variations of the order of magnitude of 1% to 2%. Notably, no differences in GCS were found among brands of imaging equipment. CONCLUSION:This study established normal values of LV regional circumferential strain and GCS and identified sex-, age-, and race-related differences when present.

Abstract Funding Acknowledgements Type of funding sources: None. Background New or worsening tricuspid regurgitation (TR) is associated with right-sided heart failure and worsened outcomes. Cardiac Implantable Electronic Devices (CIEDs), which are being implanted at growing rates worldwide, are increasingly being recognized as associated with TR occurrence related to interference with the tricuspid valve (TV) apparatus. Purpose: We sought to identify echocardiographic features in the right ventricle and TV that differentiate patients who have anatomically demonstrated interference with the TV on direct pathology inspection. Methods Explanted hearts from 44 consecutive patients undergoing orthotopic heart transplant (55 ±13 yrs, 68% men) with known implanted CIEDs were dissected to assess the presence (n = 18) or absence (n = 26) of CIED interference with the TV (Figure). Echocardiographic measurements performed prior to transplantation, including left and right ventricular (LV, RV) size and performance metrics as well as TR severity, were compared between both groups using non-parametric testing. Results Echocardiographic features of patients with and without anatomically demonstrated TV interference are shown in the Table. Although overall LV dimensions and volumes were not different between the two groups and LV ejection fraction was severely reduced in both groups, patients demonstrating CIED interference trended towards larger right atrial volumes (Table) and also larger RV and tricuspid annular sizes. Importantly, however, they were more than 4 times likely to have abnormal right ventricular function. Lastly, patients with tricuspid apparatus interference tended to have more significant TR, although these differences have not reached statistical significance (Table). Conclusion CIED interreference with tricuspid valve apparatus occurs frequently (41%) among patients with CIEDs, who undergo orthotopic heart transplantation. This may be associated with worsening TV function with subsequent changes in right atrial and ventricular geometry and function. In light of prior data showing poor outcomes with CIED associated TR, this study emphasizes the importance of non-invasive assessment of CIED interference with the tricuspid valve, in order to improve device placement and patient outcomes. Abstract Figure Abstract Table

Background: Cardiac amyloidosis (CA) is an infiltrative cardiomyopathy in which abnormally folded proteins deposit within the myocardium and the atrial walls. While left atrial dysfunction has been previously reported, the impact of CA on right atrial (RA) structure and function is unknown. Methods: We retrospectively studied 118 patients (67 immunoglobulin light chain [AL-CA], 51 transthyretin [ATTR-CA]; age, 70 +/- 12 years; 57% men) who underwent transthoracic echocardiogram in sinus rhythm. Right atrial reservoir, conduit, and booster strain were quantified using speckle-tracking and compared between patients with CA and 50 healthy age-, sex-, and race-matched controls using the chi-squared or Mann-Whitney test. The relationship between RA parameters and mortality was assessed using Cox regression. Results: Right atrial volume was significantly larger in cases with CA compared with in controls: 29 (22-37) vs 21 (15-25) mL/m(2), P <.001. Right atrial reservoir (21% [14%-35%] vs 37% [34%-43%], P <.001), conduit 11% [18%-6%] vs 14% [11%-17%], P <.001), and booster (10% [17%-5%] vs 23% [20%-27%], P <.001) strains were all significantly more impaired in the CA group compared with controls. Compared with AL-CA, ATTR-CA patients had significantly larger RA volume (34 [26-44] vs 28 [20-35] mL/m(2), P =.005) and significantly more impaired RA reservoir (17% [10%-30%] vs 27% [17%-37%], P =.007), conduit (8% [13%-6%] vs 13% [20%8%], P =.031), and booster (7% [14%-4%] vs 11% [18%-6%], P =.030) strain. Among CA patients, RA reservoir (hazard ratio = 0.97 per %, P =.006) and RA conduit (hazard ratio = 1.05 per %, P =.004) were significantly associated with mortality, while RA volume (P =.362) and RA booster strain (P =.180) were not. Conclusions: In CA, abnormalities in RA size and strain are highly prevalent and associated with worse prognosis, suggesting the presence of intrinsic RA atriopathy. Right atrial strain appears to be a potentially useful marker in the diagnosis, subtype differentiation, and risk stratification of CA.

Abstract Funding Acknowledgements Type of funding sources: None. Background The mitral valve (MV) is a complex and dynamic structure that can be best evaluated by 3D imaging. Recent advances in mitral percutaneous interventions have escalated the need for a more quantitative and comprehensive assessment of the MV. Understanding normal valve size, structure and function is essential for differentiation of healthy from disease states. Purpose Using data from the World Alliance of Societies of Echocardiography (WASE) normal values study, we aimed to determine how normative values for the MV apparatus vary across ages and sexes. Methods 3D full volume data sets obtained with transthoracic echocardiography in 618 normal subjects (50% male) were analyzed using commercial MV analysis software (TOMTEC) to determine annular and leaflet dimensions and areas (Figure). The subjects were divided into groups by sex (301 men and 311 women) and by age: 18-40 years (N = 211), 41-65 years (N = 201) and >65 years (N = 206) to identify sex- and age-related differences. Inter- and intra-observer variability was assessed in a subset of 30 subjects and expressed as mean absolute difference between pairs of repeated measurements. Results Compared to women, men had larger annular dimensions in both the anterior-posterior and anterolateral-posteromedial planes, as well as larger annular circumference and area, larger tenting size parameters, and larger leaflet length and area (Table, bottom left). Age-related differences in multiple parameters were statistically significant (Table, bottom right). Of note, sex- and age-related differences in most MV parameters were comparable to or smaller than the corresponding measurement variability (Table, top right). Conclusion This analysis of the WASE data provides normative values of the mitral apparatus size and morphology for male and female subjects. While sex- and age-related differences were noted, they need to be interpreted with caution in view of the associated measurement variability. Abstract Figure. Abstract Figure.

Theoretically, artificial intelligence can provide an accurate automatic solution to measure right ventricular (RV) ejection fraction (RVEF) from cardiovascular magnetic resonance (CMR) images, despite the complex RV geometry. However, in our recent study, commercially available deep learning (DL) algorithms for RVEF quantification performed poorly in some patients. The current study was designed to test the hypothesis that quantification of RV function could be improved in these patients by using more diverse CMR datasets in addition to domain-specific quantitative performance evaluation metrics during the cross-validation phase of DL algorithm development. We identified 100 patients from our prior study who had the largest differences between manually measured and automated RVEF values. Automated RVEF measurements were performed using the original version of the algorithm (DL1), an updated version (DL2) developed from a dataset that included a wider range of RV pathology and validated using multiple domain-specific quantitative performance evaluation metrics, and conventional methodology performed by a core laboratory (CORE). Each of the DL-RVEF approaches was compared against CORE-RVEF reference values using linear regression and Bland–Altman analyses. Additionally, RVEF values were classified into 3 categories: ≤ 35%, 35–50%, and ≥ 50%. Agreement between RVEF classifications made by the DL approaches and the CORE measurements was tested. CORE-RVEF and DL-RVEFs were obtained in all patients (feasibility of 100%). DL2-RVEF correlated with CORE-RVEF better than DL1-RVEF (r = 0.87 vs. r = 0.42), with narrower limits of agreement. As a result, DL2 algorithm also showed increasing accuracy from 0.53 to 0.80 for categorizing RV function. The use of a new DL algorithm cross-validated on a dataset with a wide range of RV pathology using multiple domain-specific metrics resulted in a considerable improvement in the accuracy of automated RVEF measurements. This improvement was demonstrated in patients whose images were the most challenging and resulted in the largest RVEF errors. These findings underscore the critical importance of this strategy in the development of DL approaches for automated CMR measurements.

Abstract Funding Acknowledgements Type of funding sources: Private company. Main funding source(s): Caption Health Background Artificial intelligence (AI) has the potential to expedite the acquisition of transthoracic echocardiograms (TTE) and provide automated quantitative data including left ventricular ejection fraction (LVEF). Specifically, AI-based imaging systems may permit less experienced individuals to obtain quantitative measurements of LVEF, with important implications for clinical workflow. Purpose We sought to evaluate the accuracy of an AI-based imaging system for the evaluation of LVEF in a spectrum of novice TTE imagers in a real-world clinical setting, and hypothesized that after appropriate training, it can perform as well as experienced sonographers. Methods Consecutive exams (N = 102; BMI 29 ± 7; most common indications for TTE: heart failure, arrhythmia, valve assessment) were performed by a cohort of physician trainees (MD) and nurses (RN) with no prior TTE experience, using an AI-guided imaging system equipped with real-time prescriptive guidance software that automatically measures LVEF. Guided imaging included 3 views: parasternal long-axis (PLAX), apical 4-and apical 2-chamber (AP4, AP2), with the software recognizing when a satisfactory imaging window is obtained and then auto-capturing and automatically generating an AI based LVEF. AI-EF measurements were compared against the reference standard EF (Ref EF) measured by 2 expert sonographers according to ASE chamber quantification guidelines. Bland-Altman analysis was performed to determine inter-technique agreement. Results Feasibility was 80%. AI-EF and Ref EF demonstrated strong correlation when all 3 views were obtained, with a small bias (Table). In cases when <3 views were available, the combination of PLAX and AP4 views achieved comparable results, with a slight increase in bias and limits of agreement. When results were stratified by cohort (MD vs RN), MD AI-EF attempts showed greater feasibility (100%, n = 51) with stronger correlation (r = 0.93) and smaller bias (-1.9%) compared to RN (feasibility of 61%; n = 31 and r= 0.85, bias -2.1%). Conclusions Use of an AI-assisted imaging system for limited TTE imaging by novices is feasible in a real-world setting, with the AI based EF in good agreement with reference standard. Acquisition of all 3 views provided optimal results, but the combination of AP4/PLAX views performed reasonably well, without the AP2 view that is more difficult to acquire for less skilled users. Untrained MD were more successful when compared to RN, suggesting additional training may be needed for specific user groups. Abstract Figure.

Abstract Funding Acknowledgements Type of funding sources: Private company. Main funding source(s): Philips Healthcare Background. Echocardiographic quantification of left ventricular function (LVF) is mainly based on ejection fraction (EF) measurements, which relies on either manual or automated identification of endocardial boundaries followed by calculation of model-based end-systolic and end-diastolic LV volumes. Recent developments in artificial intelligence resulted in computer algorithms that allow fully automated detection of endocardial boundaries and measurement of LV volumes and EF. However, this methodology is prone to errors and inter-measurement variability. We hypothesized that a fully automated deep learning algorithm could be developed, which would accurately classify LVF while avoiding volume and EF measurements. This study was designed to test the accuracy of this approach. Methods. Deep learning algorithm was developed (Philips Research) based on convolutional neural network (CNN) that uses as input dynamic sequences of apical 2- and 4-chamber echocardiographic views obtained without ultrasound enhancing agents. We used for CNN development a database of clinical DICOM studies: a training set of 14,427 studies with normal LV function and 6,135 abnormal, and a validation set of 2,898 normal and 1,081 abnormal studies, based on Philips IntelliSpace Cardiovascular (ISCV) codes found (defined by cardiologists) in the patients’ reports. The CNN was trained to automatically classify LVF into 3 categories: (1) normal, (2) mildly-to-moderately or moderately reduced, and (3) moderately-to-severely or severely reduced. In the validation set, the automated classifications were compared to those in the patients’ reports as a reference standard. Accuracy of the automated classification was tested using contingency tables, from which sensitivity, specificity, and negative and positive predictive values (NPV, PPV) and overall accuracy were calculated for each category of LVF. Additionally, the area under ROC curve (AUC) was calculated to assess the diagnostic accuracy of the automated classification for each LVF category. Results. Automated classification of LVF showed high levels of diagnostic accuracy in identifying cases with LVF in all 3 categories, reflected by high AUC values: (1) 0.94, (2) 0.87 and (3) 0.97 (Figure), and overall accuracy of 0.84 (Table). Conclusions. Deep learning algorithm based on CNN allowed accurate automated classification of LVF, when tested on ∼4,000 clinical studies and compared to ISCV codes found in the patients’ reports. This novel fully-automated methodology may become a useful aid in the interpretation of echocardiographic images by providing the reader with a preliminary assessment of LVF. Abstract Figure.

Background Procedural experience for residents and fellows is critical for achieving competence, and documentation of procedures performed is required. Procedure logs serve as the record of this experience, but are commonly generated manually, require substantial administrative effort, and cannot be corroborated for accuracy. Objective We developed and implemented a structured clinical-educational report template (CERT), which automatically generates procedure logs directly from the clinical record. Methods Our CERT aimed to replace the post-procedure note template for our cardiac catheterization laboratory and was incorporated into the electronic health record system. Numbers of documented procedures in automated CERT-derived procedure logs over a 1-year post-intervention period (2018-2019) were compared to manual logs and corrected for clinical volume changes. The CERT's impact on fellowship experience was also assessed. Results Automated CERT procedure logs increased weekly procedural documentation over manual procedure logs for total procedures (24.2 ± 6.1 vs 17.1 ± 6.8, P = .007), left heart catheterizations (14.5 ± 3.6 vs 10.8 ± 4.2, P = .039), total procedural elements (40.2 ± 11.4 versus 20.9 ± 12.5, P < .001), and captured procedural details not previously documented. The CERT also reduced self-reported administrative time and improved fellowship experience. Conclusions A novel CERT allows procedure logs to be automatically derived from the clinical record and increased the number of documented procedures, compared to manual logging. This innovation ensures an accurate record of procedural experience and reduces self-reported non-educational administrative time for cardiology fellows.

A comprehensive cardiovascular magnetic resonance (CMR) exam includes assessment of regional wall motion, which typically involves qualitative interpretation by an experienced reader, although time-consuming quantitative measurements of myocardial wall thickening can be performed. Myocardial strain imaging offers the potential to more accurately quantify the extent and severity of regional wall motion abnormalities and has even been proposed as a mechanism to identify changes in regional contractility before they are visually evident1. Strain imaging may be particularly valuable to assess the impact of coronary artery disease on the myocardium because ischemia is known to first alter the longitudinal deformation of the subendocardial myocardial fibers prior to impacting the regional thickening in the radial direction that is required to create a visually appreciable wall motion abnormality1. In patients with chest pain, alterations in strain following the infusion of an inotrope such as dobutamine can be used to improve the detection of underlying coronary artery disease2. Myocardial strain can be measured using several different CMR-based tagging techniques, such as complementary spatial modulation of magnetization (C-SPAMM) 3 with harmonic phase encoding (HARP) post-processing 4,5, displacement encoding with stimulating echoes (DENSE) 6, cine phase contrast imaging, and strain-encoded CMR (SENC) 7–9. Additionally, it is also possible to use feature tracking algorithms to derive strain measurements from standard cine-CMR images 10,11. Each of these approaches has its advantages and limitations. Recent improvements in SENC-CMR allow for image acquisition to occur during a single heart beat in real time without sacrificing temporal resolution. Unlike the other strain imaging approaches, with SENC imaging being performed during a single heart beat, image quality is not significantly impacted by respiratory motion and arrhythmia. Previous single center studies have suggested that SENC-CMR can be used to improve the detection of coronary artery disease when compared to cine-image wall motion interpretation 12. In this multi-center study, we hypothesized that in the presence of coronary artery disease, changes in regional longitudinal and circumferential strain detected using SENC under intermediate-dose dobutamine stress CMR would precede changes in wall thickening visible on cine-CMR images.

Same-day cancellations of hospital-based surgical procedures are a known source of operational inefficiency and lost revenue. 1 Xue W. Yan Z. Barnett R. Fleisher L. Liu R. Dynamics of elective case cancellation for inpatient and outpatient in an academic center. J Anesth Clin Res. 2013; 4: 314 PubMed Google Scholar , 2 Smith M.M. Mauermann W.J. Cook D.J. Hyder J.A. Dearani J.A. Barbara D.W. Same-day cancellation of cardiac surgery: a retrospective review at a large academic tertiary referral center. J Thorac Cardiovasc Surg. 2014; 148: 721-725 Abstract Full Text Full Text PDF PubMed Scopus (17) Google Scholar , 3 Smith B.B. Smith M.M. Hyder J.A. Mauermann W.J. Warner M.E. Licatino L.K. et al. Same-day cancellation in ambulatory surgery: a retrospective review at a large academic tertiary referral center. J Ambul Care Manage. 2018; 41: 118-127 Crossref PubMed Scopus (13) Google Scholar , 4 Matzek L.J. Smith B.B. Mauermann W.J. Bower T.C. Smith M.M. Same-day cancellation in vascular surgery: 10-year review at a large tertiary care center. Ann Vasc Surg. 2020; 62: 349-355 Abstract Full Text Full Text PDF PubMed Scopus (6) Google Scholar Transesophageal echocardiography (TEE) is a widely used, resource-intense, invasive cardiovascular imaging procedure, but same-day TEE cancellations have not been studied. We sought to determine the frequency and causes of same-day TEE cancellations at a large academic medical center and test a targeted remediation strategy. Transesophageal echocardiograms performed in our echocardiography laboratory over a 7-week period were assessed for completion and reasons for cancellation or postponement. Based on identified causes, a two-part targeted remediation strategy was developed and prospectively tested over the ensuing 4-week period.

BACKGROUND:Although cardiac magnetic resonance (CMR) can accurately quantify global left ventricular strain using feature tracking (FT), it has been suggested that FT cannot reliably quantify regional strain. We aimed to determine whether abnormalities in regional strain measured using FT can be detected within areas of myocardial scar and to determine the extent to which the regional strain measurement is impacted by LV ejection fraction (EF). METHODS:We retrospectively studied 96 patients (46 with LVEF ≤ 40%, 50 with LVEF > 40%) with coronary artery disease and a late gadolinium enhancement (LGE) pattern consistent with myocardial infarction, who underwent CMR imaging (1.5T). Regional peak systolic longitudinal and circumferential strains (RLS, RCS) were measured within LGE and non-LGE areas. Linear regression analysis was performed for strain in both areas against LVEF to determine whether the relationship between strain and LGE holds across the LV function spectrum. Receiver-operating curve (ROC) analysis was performed in 33 patients (derivation cohort) to optimize strain cutoff, which was tested in the remaining 63 patients (validation cohort) for its ability to differentiate LGE from non-LGE areas. RESULTS:Both RLS and RCS magnitudes were reduced in LGE areas: RLS = -10.4 ± 6.2% versus -21.0 ± 8.5% (p < 0.001); RCS = -10.4 ± 6.0% versus -18.9 ± 8.6%, respectively (p < 0.001), but there was considerable overlap between LGE and non-LGE areas. Linear regression revealed that it was partially driven by the natural dependence between strain and EF, suggesting that EF-corrected strain cutoff is needed to detect LGE. ROC analysis showed the ability of both RLS and RCS to differentiate LGE from non-LGE areas: area under curve 0.95 and 0.89, respectively. In the validation cohort, optimal cutoffs of RLS/EF = 0.36 and RCS/EF = 0.37 yielded sensitivity, specificity and accuracy 0.74-0.78. CONCLUSION:Abnormalities in RLS and RCS within areas of myocardial scar can be detected using CMR-FT; however, LVEF must be accounted for.

Background: Echocardiographic quantification of left ventricular (LV) ejection fraction (EF) relies on either manual or automated identification of endocardial boundaries followed by model-based calculation of end-systolic and end-diastolic LV volumes. Recent developments in artificial intelligence resulted in computer algorithms that allow near automated detection of endocardial boundaries and measurement of LV volumes and function. However, boundary identification is still prone to errors limiting accuracy in certain patients. We hypothesized that a fully automated machine learning algorithm could circumvent border detection and instead would estimate the degree of ventricular contraction, similar to a human expert trained on tens of thousands of images. Methods: Machine learning algorithm was developed and trained to automatically estimate LVEF on a database of >50 000 echocardiographic studies, including multiple apical 2- and 4-chamber views (AutoEF, BayLabs). Testing was performed on an independent group of 99 patients, whose automated EF values were compared with reference values obtained by averaging measurements by 3 experts using conventional volume-based technique. Inter-technique agreement was assessed using linear regression and Bland-Altman analysis. Consistency was assessed by mean absolute deviation among automated estimates from different combinations of apical views. Finally, sensitivity and specificity of detecting of EF ≤35% were calculated. These metrics were compared side-by-side against the same reference standard to those obtained from conventional EF measurements by clinical readers. Results: Automated estimation of LVEF was feasible in all 99 patients. AutoEF values showed high consistency (mean absolute deviation =2.9%) and excellent agreement with the reference values: r =0.95, bias=1.0%, limits of agreement =±11.8%, with sensitivity 0.90 and specificity 0.92 for detection of EF ≤35%. This was similar to clinicians’ measurements: r =0.94, bias=1.4%, limits of agreement =±13.4%, sensitivity 0.93, specificity 0.87. Conclusions: Machine learning algorithm for volume-independent LVEF estimation is highly feasible and similar in accuracy to conventional volume-based measurements, when compared with reference values provided by an expert panel.

Implantable cardioverter defibrillators (ICDs) have been a relative contraindication to cardiovascular magnetic resonance imaging. Although cardiovascular magnetic resonance provides valuable information regarding scar in patients with ventricular arrhythmias or cardiomyopathy, ICDs in these patients frequently cause artifacts hindering accurate interpretation of both cine and late gadolinium enhancement (LGE) images. We sought to quantify the frequency and severity of artifact on LGE images and assess whether a modified wideband LGE protocol could improve the diagnostic yield of scar identification in agreement with invasive electroanatomic mapping (EAM). Forty-nine patients with ICDs and ventricular tachycardia (VT) or cardiomyopathy underwent CMR (Philips 1.5T), including standard and wideband LGE imaging. A safety algorithm was followed throughout the protocol. Standard and wideband LGE short-axis images were graded using an artifact score on a per-slice basis. LGE on wideband images was compared with EAM in 27 of 49 patients who underwent VT ablation. There were no adverse patient- or device related events. With standard LGE imaging, 84% of patients demonstrated some degree of hyperenhancement artifact, which persisted in 22% on wideband LGE but with much less extent. Wideband LGE imaging resulted in an increase from 48% to 94% diagnostic quality slices, with a significant reduction in artifact score, and correlated with EAM in 21 of 27 patients (78%). In conclusion, assessment of standard LGE is markedly limited by artifact in patients with ICD. The use of wideband LGE significantly improves image quality and can accurately localize myocardial scar before VT ablation. (C) 2019 Elsevier Inc. All rights reserved.

Tricuspid annular (TA) size, assessed by 2D transthoracic echocardiography (TTE), has a well-established prognostic value in patients undergoing mitral valve surgery, with TA dilatation triggering simultaneous tricuspid annuloplasty. While TA dilatation is common in patients with dilated atria secondary to atrial fibrillation, little is known about the mechanisms of TA dilatation in patients with sinus rhythm (SR). This study aimed to identify echocardiographic parameters most closely related to the TA size as a potential tool for identification of patients prone to developing TA enlargement. 120 patients with SR underwent clinically indicated TTE, including 30 patients with normal hearts and 90 patients diagnosed with at least one right heart abnormality, defined as: right ventricular (RV) or right atrial (RA) dilatation, ≥ moderate tricuspid regurgitation (TR) and elevated systolic pulmonary artery pressure (sPAP). RA and RV end-diastolic and end-systolic volumes (EDV, ESV) and function were measured using commercial 3D software (TomTec). 3D RV long and short axes were used as surrogate indices of RV shape. Degrees of TR and sPAP were estimated by 2D TTE. 3D TA sizing was performed at end-diastole using 3D custom software. Linear regression analysis was used to identify variables best correlated with TA size, followed by multivariate analysis to identify independent associations. The highest correlations were found between TA area and: RA ESV (r = 0.73; p < 0.01), RV EDV (r = 0.58; p < 0.01), RV end-diastolic long and short axes (r = 0.53, 0.42; both p < 0.01), TR degree (r = 0.40; p < 0.01) and sPAP (r = 0.32; p < 0.01). Multivariate analysis revealed that RA ESV was the only parameter independently associated with TA area (p < 0.05, r = 0.85). In conclusion, RA volume plays an important role in TA dilatation even in patients with normal SR. Understanding of annular remodeling mechanisms could aid in identifying patients at higher risk for TA dilatation, especially those scheduled for mitral valve surgery.

Severely reduced left ventricular (LV) ejection fraction (EF) derived from 2D echocardiographic (2DE) images is associated with increased mortality and used to guide therapeutic choices. Global longitudinal strain (GLS) is more sensitive than LVEF to detect abnormal LV function, and accordingly may help identify patients with mildly-to-moderately reduced LVEF who are at a similarly high cardiovascular (CV) risk. We hypothesized that 3D echocardiographic (3DE) measurements of EF and GLS, which are more reliable and reproducible, may have even better predictive value than the 2DE indices, and compared their ability to identify such patients. We retrospectively studied 104 inpatients with 2DE-derived LVEF of 30–50% who underwent transthoracic echocardiography during 2006–2010 period, had good quality images, and were followed-up through 2016. Both 2DE and 3DE images were analyzed to measure LVEF and GLS. Kaplan–Meier survival curves were generated for two subgroups defined by the median of each parameter as the cutoff. Of the 104 patients, 32 died of CV related causes. Cox regression revealed that 3D GLS was the only variable associated with CV mortality. Kaplan–Meier curves showed that 2D LVEF, 2D GLS and 3D EF were unable to differentiate patients at higher CV mortality risk, but 3D GLS was the only parameter to do so. Because 3D GLS is able to identify patients with mildly-to-moderately reduced LVEF who are at higher CV mortality risk, its incorporation into clinical decisions may improve survival of those who would benefit from therapeutic interventions not indicated according to the current guidelines.