scholarly journals Using deep learning algorithms to simultaneously identify right and left ventricular dysfunction from the electrocardiogram.

2021 ◽  
Author(s):  
Akhil Vaid ◽  
Kipp W Johnson ◽  
Marcus A Badgeley ◽  
Sulaiman Somani ◽  
Mesude Bicak ◽  
...  
Keyword(s):  
Heart Failure ◽  
Deep Learning ◽  
Right Ventricular ◽  
Systolic Dysfunction ◽  
External Validation ◽  
Left Ventricular ◽  
Left Ventricular Ejection ◽  
Lv Function ◽  
Ventricular Ejection Fraction ◽  
Left And Right

Background Rapid evaluation of left and right ventricular function using deep learning (DL) on electrocardiograms (ECG) can assist diagnostic workflow. However, DL tools to estimate right ventricular (RV) function do not exist, while ones to estimate left ventricular (LV) function are restricted to quantification of very low LV function only. Objectives This study sought to develop deep learning models capable of comprehensively quantifying left and right ventricular dysfunction from ECG data in a large, diverse population. Methods A multi-center study was conducted with data from five New York City hospitals; four for internal testing and one serving as external validation. We created novel DL models to classify Left Ventricular Ejection Fraction (LVEF) into categories derived from the latest universal definition of heart failure, estimate LVEF through regression, and predict a composite outcome of either RV systolic dysfunction or RV dilation. Results We obtained echocardiogram LVEF estimates for 147,636 patients paired to 715,890 ECGs. We used Natural Language Processing (NLP) to extract RV size and systolic function information from 404,502 echocardiogram reports paired to 761,510 ECGs for 148,227 patients. For LVEF classification in internal testing, Area Under Curve (AUC) at detection of LVEF<=40%, 40%<LVEF<=50%, and LVEF>50% was 0.94 (95% CI:0.94-0.94), 0.82 (0.81-0.83), and 0.89 (0.89-0.89) respectively. For external validation, these results were 0.94 (0.94-0.95), 0.73 (0.72-0.74) and 0.87 (0.87-0.88). For regression, the mean absolute error was 5.84% (5.82-5.85) for internal testing, and 6.14% (6.13-6.16) in external validation. For prediction of the composite RV outcome, AUC was 0.84 (0.84-0.84) in both internal testing and external validation. Conclusions DL on ECG data can be utilized to create inexpensive screening, diagnostic, and predictive tools for both LV/RV dysfunction. Such tools may bridge the applicability of ECGs and echocardiography, and enable prioritization of patients for further interventions for either sided failure progressing to biventricular disease. Keywords Artificial Intelligence, Deep Learning, Machine Learning, HFrEF, Right Ventricular Dilation, Right Ventricular Systolic Dysfunction, echocardiography, electrocardiogram, ECG, EKG, LVEF, Left Ventricular Ejection Fraction, Left Heart Failure, Right Heart Failure

Abstract 15164: Impact of Conventional Right Ventricular Pacing in Patients With Heart Failure With Preserved Ejection Fraction: Insights From Pursuit-hfpef Registry

Circulation ◽  
2020 ◽  
Vol 142 (Suppl_3) ◽  
Author(s):  
Tetsuya Watanabe ◽  
Shungo Hikoso ◽  
Daisaku D Nakatani ◽  
Masamichi Yano ◽  
Shunsuke Tamaki ◽  
...  
Keyword(s):  
Heart Failure ◽  
Ejection Fraction ◽  
Right Ventricular ◽  
Pacemaker Implantation ◽  
Left Ventricular ◽  
Left Ventricular Ejection ◽  
Lv Function ◽  
Preserved Ejection Fraction ◽  
Ventricular Ejection Fraction ◽  
Patients With Heart Failure

Introduction: Previous studies suggest that conventional right ventricular (RV) pacing may have a deleterious effect on left ventricular (LV) function. However, there are no reports examining the effects of permanent pacemakers (PM) in patients with heart failure with preserved ejection fraction (HFpEF) admitted with acute decompensated heart failure (ADHF). Methods and Results: Patients' data were extracted from The Prospective mUlticenteR obServational stUdy of patIenTs with Heart Failure with Preserved Ejection Fraction (PURSUIT-HFpEF) study, which is a prospective multicenter observational registry for ADHF patients with LVEF ≥50% in Osaka. Echocardiographic measurements were performed just before and 1 year after discharge. We analyzed 623 patients (median age: 81[30-101] years) after exclusion of patients on dialysis. At the admission, 55 of 623 patients had a previously implanted pacemaker (PM(+)). There were no significant differences in baseline clinical characteristics before discharge such as age, gender, renal function, echocardiographic parameters between PM(+) and PM(-) groups. One year after discharge, left ventricular ejection fraction (LVEF) in PM(+) group was significantly lower (58.3±8.6 vs 62.3±8.4%, p=0.005) than those in PM(-) group. During a mean follow-up period of 1.25±0.84years, 171patients had heart failure hospitalization (HFH). At the multivariate Cox analysis, pacemaker was significantly associated with HFH, independently of age, serum NT-proBNP level, and LVEF after the adjustment with NYHA functional class, sex, and hemoglobin. Kaplan-Meier curve analysis revealed that patients in PM(+) group had a significantly higher risk of HFH than those in PM(-) group. Conclusion: The development of cardiac systolic and diastolic dysfunction occurred in patients with HFpEF with a RV pacemaker implantation, which would be associated with poor outcome.

ACCESSMENT OF RIGHT VENTRICULAR SYSTOLIC FUNCTION USING TAPSE

2011 ◽  
pp. 62-70
Author(s):  
Lien Nhut Nguyen ◽  
Anh Vu Nguyen
Keyword(s):  
Heart Failure ◽  
Left Ventricular Ejection Fraction ◽  
Ejection Fraction ◽  
Right Ventricular ◽  
Systolic Function ◽  
Systolic Heart Failure ◽  
Left Ventricular ◽  
Ventricular Ejection ◽  
Left Ventricular Ejection ◽  
Ventricular Ejection Fraction

Background: The prognostic importance of right ventricular (RV) dysfunction has been suggested in patients with systolic heart failure (due to primary or secondary dilated cardiomyopathy - DCM). Tricuspid annular plane systolic excursion (TAPSE) is a simple, feasible, reality, non-invasive measurement by transthoracic echocardiography for evaluating RV systolic function. Objectives: To evaluate TAPSE in patients with primary or secondary DCM who have left ventricular ejection fraction ≤ 40% and to find the relation between TAPSE and LVEF, LVDd, RVDd, RVDd/LVDd, RA size, severity of TR and PAPs. Materials and Methods: 61 patients (36 males, 59%) mean age 58.6 ± 14.4 years old with clinical signs and symtomps of chronic heart failure which caused by primary or secondary DCM and LVEF ≤ 40% and 30 healthy subject (15 males, 50%) mean age 57.1 ± 16.8 were included in this study. All patients and controls were underwent echocardiographic examination by M-mode, two dimentional, convensional Dopler and TAPSE. Results: TAPSE is significant low in patients compare with the controls (13.93±2.78 mm vs 23.57± 1.60mm, p<0.001). TAPSE is linearly positive correlate with echocardiographic left ventricular ejection fraction (r= 0,43; p<0,001) and linearly negative correlate with RVDd (r= -0.39; p<0.01), RVDd/LVDd (r=-0.33; p<0.01), RA size (r=-0.35; p<0.01), TR (r=-0.26; p<0.05); however, no correlation was found with LVDd and PAPs. Conclusions: 1. Decreased RV systolic function as estimated by TAPSE in patients with systolic heart failure primary and secondary DCM) compare with controls. 2. TAPSE is linearly positive correlate with LVEF (r= 0.43; p<0.001) and linearly negative correlate with RVDd (r= -0.39; p<0.01), RVDd/LVDd (r=-0.33; p<0.01), RA size (r=-0.35; p<0.01), TR (r=-0.26; p<0.05); however, no correlation is found with LVDd and PAPs. 3. TAPSE should be used routinely as a simple, feasible, reality method of estimating RV function in the patients systolic heart failure DCM (primary and secondary).

Abstract 3791: Improved Ventricular-Arterial Coupling After 4 Weeks of Exercise Training in Patients with Chronic Heart Failure

Circulation ◽  
2007 ◽  
Vol 116 (suppl_16) ◽  
Author(s):  
Marcus Sandri ◽  
Stephan Gielen ◽  
Norman Mangner ◽  
Volker Adams ◽  
Sandra Erbs ◽  
...  
Keyword(s):  
Heart Failure ◽  
Chronic Heart Failure ◽  
Exercise Training ◽  
Endothelial Function ◽  
Training Group ◽  
Left Ventricular ◽  
Left Ventricular Ejection ◽  
Control Group ◽  
Lv Function ◽  
Ventricular Ejection Fraction

Background: The concept of ventricular-arterial coupling implies that LV-function is determined by the three factors left ventricular diastolic, left ventricular systolic and arterial elastance. We have previously documented an improvement in endothelial function and systolic LV-function in patients with chronic heart failure (CHF) after 6 months of exercise training (ET). It remains, however, unclear, how shorter ET periods may affect endothelial, systolic and diastolic ventricular function as echocardiographic parameters related to ventricular arterial coupling in patients with CHF. METHODS: In this ongoing study we randomised 43 patients with stable CHF (age 60.3 ± 2.9 years, EF 27.4 ± 1.7%, VO 2 max 14.7 ± 4.3ml/kg*min) to a training or a control group (C). Patients in the training group exercised 4 times daily at 70% of the individual heart rate reserve for 4 weeks under supervision. At baseline and after 4 weeks the E/A ratio and septal/lateral E’/A’ velocities were determined by echocardiography with tissue Doppler. Exercise capacity was measured by ergospirometry and flow-mediated dilatation (FMD) was assessed by high-resolution radial ultrasound. RESULTS: After only 4 weeks of ET oxygen uptake at peak exercise increased from 14.9 ± 3.3 to 18.1 ± 4.7 ml/min/kg, (p<0.01 vs. C) in training subjects. Left ventricular ejection fraction improved from 26.8 ± 4.6 to 33.1 ± 5.5% (p<0.05 vs. C) in patients of the training group while it remained unchanged in the control group. E/A-ratio mended from 0.63 ± 0.12 to 0.81 ± 0.22 (p<0.01 vs. C) in training patients. Septal E’ velocities increased from 5.5 ± 0.5 to 7.8 ± 1.4 cm/s in training patients (p<0.05 vs. C). FMD of the radial artery improved from 8.2 ± 2.1 to 15.2 ± 3.8% (p<0.01 vs. C) as a result of ET. CONCLUSIONS: Only 4 weeks of endurance training are highly effective with significantly improved FMD accompanied by an emended systolic and diastolic LV-function. We hypothesise that the improvement in LV-EF in training patients may be caused by a corrected ventricular-arterial coupling: ventricular diastolic relaxation and effective endothelial function are ameliorated resulting in an augmentation of stroke volume.

Atrial function discriminates paroxysmal AF patients with HFpEF from those without HFpEF: subanalysis from AF-RISK study

EP Europace ◽  
2021 ◽  
Vol 23 (Supplement_3) ◽  
Author(s):  
V Artola ◽  
B Santema ◽  
R De With ◽  
B Nguyen ◽  
D Linz ◽  
...  
Keyword(s):  
Heart Failure ◽  
Atrial Fibrillation ◽  
Ejection Fraction ◽  
Left Ventricular ◽  
Atrial Function ◽  
Volume Index ◽  
Left Ventricular Ejection ◽  
Ventricular Ejection Fraction ◽  
Left And Right ◽  
Right Atria

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – EU funding. Main funding source(s): European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie. Grant support from the Dutch Heart Foundation [NHS2010B233] Background. Atrial fibrillation (AF) and heart failure with preserved ejection fraction (HFpEF) are two cardiovascular conditions that often coexist. Overlapping symptoms, biomarker profile, and echocardiographic changes hinder the diagnosis of underlying HFpEF in patients with AF and suggest that both conditions might reflect similar remodelling processes in the heart. Purpose. To assess cardiac remodelling in AF patients with versus without concomitant HFpEF by transthoracic echocardiography, focusing on atrial dimension and strain. Methods. We selected 120 patients included in AF-RISK, a prospective, observational, multicentre study aiming to identify a risk profile to guide atrial fibrillation therapy study. Patients had paroxysmal AF diagnosed within three years before inclusion, had a left ventricular ejection fraction (LVEF) ≥50% and were in sinus rhythm at the moment of performing echocardiography and blood sampling. Patients were matched by nearest neighbour by age and sex with a 1:1 ratio and were classified into two groups: 1) AF with HFpEF (n = 60) and 2) AF without HFpEF (n = 60). The diagnosis of HFpEF was based on the 2016 ESC heart failure guidelines, including symptoms and signs of heart failure, N-terminal pro-B-type natriuretic peptide (NT-proBNP) ≥125pg/ml, and one of the following echocardiographic measures: left atrium volume index (LAVI) &gt;34ml/m2, left ventricular mass index ≥115g/m2 for men and ≥95g/m2 for women, average E/e’ ≥13cm/s and average e’ &lt;9cm/s. Measurements of reservoir, conduit and contraction strain of both atria were performed in apical four-chamber by echocardiography (GE, EchoPac BT12). Associations of clinical and echocardiographic characteristics were tested for collinearity by multivariable logistic regression analyses. LAVI, LV mass index and NT-proBNP were excluded from multivariable analysis since these markers were part of the HFpEF diagnostic criteria. Results. Patients with paroxysmal AF and concomitant HFpEF had more often hypertension (72% vs. 45%, P = 0.005), had more impaired strain phases of both the left and right atria (figure 1), had comparable LVEF and global longitudinal strain (GLS) (P = 0.168 and P = 0.212, respectively). In a model adjusted for the number of comorbidities and sex, LA contraction decrease was associated with presence of HFpEF (odds ratio per 1% LA contraction-percent was 0.94, 95% confidence interval 0.87–0.99, P = 0.042). LA contraction was not explained by LAVI in patients with concomitant HFpEF (Spearman’s rho= -0.07, P = 0.08). Conclusion. Our results show that atrial function may differentiate paroxysmal AF patients with HFpEF from those without HFpEF. In patients with paroxysmal AF, more impaired strain phases of the left and right atria were associated with concomitant HFpEF, whereas ventricular function, reflected by LVEF and GLS, did not differ. Abstract Figure. Strain distribution of both atria

Right ventricular speckle tracking in patients with heart failure – a comparison of right ventricular measures

2020 ◽  
Vol 41 (Supplement_2) ◽  
Author(s):  
I.J Lundorff ◽  
M Sengeloev ◽  
S Pedersen ◽  
D Modin ◽  
N.E Bruun ◽  
...  
Keyword(s):  
Risk Factors ◽  
Heart Failure ◽  
Right Ventricular ◽  
Left Ventricular ◽  
Left Ventricular Ejection ◽  
Funding Source ◽  
Private Company ◽  
Ventricular Ejection Fraction ◽  
Risk Of Mortality ◽  
Patients With Heart Failure

Abstract Background RV dysfunction is associated with increased mortality and morbidity in patients with heart failure. Due to the complex shape and position of the RV, assessing RV function from echocardiographic images remains a challenge. Purpose We have previously found that global longitudinal strain from 2DSTE is superior to left ventricular ejection fraction (LVEF) in identifying HFrEF patients with high risk of mortality. In this study we wanted to examine RV 2DSTE in patients with HFrEF and compare its prognostic value to conventional RV measures. Methods and results Echocardiographic examinations were retrieved from 701 patients with HFrEF. RV estimates were analysed offline, and end point was all-cause mortality. During follow-up (median 39 months) 118 patients (16.8%) died. RV GLS and RV FWS remained associated with mortality after multivariable adjustment, independent of TAPSE (RV GLS: HR 1.07, 95% CI 1.02–1.13, p=0.010, per 1% decrease) (RV FWS: HR 1.05, 95% CI 1.01–1.09, p=0.010, per 1% decrease). This seemed to be caused by significant associations in men as TAPSE remained as the only independent prognosticator in women. All RV estimates provided prognostic information incremental to established risk factors and significantly increased C-statistics (TAPSE: 0.74 to 0.75; RVFAC: 0.74 to 0.75; RVFWS: 0.74 to 0.77; RVGLS: 0.74 to 0.77). Conclusions RV strain from 2DSTE was associated with mortality in patients with HFrEF, independent of TAPSE and established risk factors. Our results indicate that RV strain is particularly valuable in male patients, whereas in women TAPSE remains a stronger prognosticator. RV GLS and the risk of mortality Funding Acknowledgement Type of funding source: Private company. Main funding source(s): PGJ reports receiving lecture fee from Novo Nordisk.

P4553Clinical and prognostic value of right ventricular dysfunction in patients with chagas heart disease during acute decompensated heart failure

2019 ◽  
Vol 40 (Supplement_1) ◽  
Author(s):  
S R R Siqueira ◽  
S M Ayub-Ferreira ◽  
P R Chizzola ◽  
V M C Salemi ◽  
S H G Lage ◽  
...  
Keyword(s):  
Heart Failure ◽  
Right Ventricular ◽  
Prognostic Value ◽  
Acute Decompensated Heart Failure ◽  
Decompensated Heart Failure ◽  
Left Ventricular ◽  
Left Ventricular Ejection ◽  
Ventricular Ejection Fraction ◽  
Jugular Venous Distension

Abstract Introduction The occurrence of right ventricular disfunction (RVD) is common in heart failure (HF) patients due to Chagas' disease (ChD). However, its clinical and prognostic value has not been studied during episodes of acute decompensated heart failure (ADHF). Purpose Evaluate the prognostic value of RVD in ADHF patients with ChD during hospitalization and after 180 days of discharge compared to other etiologies. Methods We analysed a prospective cohort of consecutive 768 patients admitted for ADHF between March 2013 and October 2018; 490 (63.7%) patients were male and the median age was 58 (48.3–66.8) years and left ventricular ejection fraction was 26% (median) (IQR 22–35%). We compared the clinical characteristics and the prognosis of ChD patients according to the presence of RVD in the echocardiogram to other etiologies. Results RVD was presented in 289 (37.6%) patients. Among patients with non-chagasic etiologies, those with RVD were younger [53 (41–62) vs 61 (52–70) years, p<0.0001], had high levels of BNP in the moment of hospitalization [1195 (606–2209) vs 886 (366– 555) pg/mL], p<0,0001], received more inotropes (79.2% vs 57.9%, p<0,0001), had longer hospitalization [35 (17–51) vs 21 (10–37) days, p<0.001] and more clinical signs of congestion as hepatomegaly (49% vs 28.6%, p<0.0001); jugular venous distension (68.3% vs 41.2%, p<0.0001) and leg edema (65.4% vs 49.2%, p=0.001). Among patients with ChD, those with RVD were older [61 (48- 66) vs 58 (48 - 67) years, p=0.017], and had more frequently signs of hypoperfusion (56.8% vs 36.5%, p=0.029), jugular venous distension (72.8% vs 52.8%, p=0.01) and hepatomegaly (56.8% vs 31.1%, p=0.011), higher BNP levels [1288 (567–2180) vs 1066 (472–2007) pg/mL, p=0.006] and more frequent use of intravenous inotropes (88.9% vs 67.1%, p=0.003); additionally ChD patients with RVD had a higher rate of death and transplant during hospitalization (51.2% vs 38.3%, p=0.001). When all groups were compared together, ChD patients with RVD had the highest rate of death, transplant and readmissions at 180-days of follow-up (Figure). Figure 1 Conclusion Patients with RVD demonstrated a distinct clinical presentation, biomarkers and worse prognosis in all etiologies. ChD patients with RVD in ADHF had the worst prognosis with the highest rate of death, heart transplant e rehospitalization in follow-up.

scholarly journals External validation of a claims-based model to predict left ventricular ejection fraction class in patients with heart failure

PLoS ONE ◽  
2021 ◽  
Vol 16 (6) ◽  
pp. e0252903
Author(s):  
Mufaddal Mahesri ◽  
Kristyn Chin ◽  
Abheenava Kumar ◽  
Aditya Barve ◽  
Rachel Studer ◽  
...  
Keyword(s):  
Heart Failure ◽  
Ejection Fraction ◽  
External Validation ◽  
Model Performance ◽  
Left Ventricular ◽  
Left Ventricular Ejection ◽  
Administrative Claims ◽  
Validation Sample ◽  
Ventricular Ejection Fraction ◽  
Medicare Claims

Background Ejection fraction (EF) is an important prognostic factor in heart failure (HF), but administrative claims databases lack information on EF. We previously developed a model to predict EF class from Medicare claims. Here, we evaluated the performance of this model in an external validation sample of commercial insurance enrollees. Methods Truven MarketScan claims linked to electronic medical records (EMR) data (IBM Explorys) containing EF measurements were used to identify a cohort of US patients with HF between 01-01-2012 and 10-31-2019. By applying the previously developed model, patients were classified into HF with reduced EF (HFrEF) or preserved EF (HFpEF). EF values recorded in EMR data were used to define gold-standard HFpEF (LVEF ≥45%) and HFrEF (LVEF<45%). Model performance was reported in terms of overall accuracy, positive predicted values (PPV), and sensitivity for HFrEF and HFpEF. Results A total of 7,001 HF patients with an average age of 71 years were identified, 1,700 (24.3%) of whom had HFrEF. An overall accuracy of 0.81 (95% CI: 0.80–0.82) was seen in this external validation sample. For HFpEF, the model had sensitivity of 0.96 (95%CI, 0.95–0.97) and PPV of 0.81 (95% CI, 0.81–0.82); while for HFrEF, the sensitivity was 0.32 (95%CI, 0.30–0.34) and PPV was 0.73 (95%CI, 0.69–0.76). These results were consistent with what was previously published in US Medicare claims data. Conclusions The successful validation of the Medicare claims-based model provides evidence that this model may be used to identify patient subgroups with specific EF class in commercial claims databases as well.

scholarly journals Analysis of cardiac involvement in patients recovered from Covid-19 without troponin elevation, evaluated by cardiovascular magnetic resonance.

2021 ◽  
Author(s):  
Adrian Carlessi ◽  
Leonel Perello ◽  
Cristian Pantaley ◽  
Armando Borsini ◽  
Lucia Rossi ◽  
...  
Keyword(s):  
Ejection Fraction ◽  
Right Ventricular ◽  
Cardiac Involvement ◽  
Systolic Dysfunction ◽  
Left Ventricular ◽  
Ventricular Ejection ◽  
Left Ventricular Ejection ◽  
Ventricular Systolic Dysfunction ◽  
Ventricular Ejection Fraction ◽  
Troponin Elevation

Abstract Background The disease caused by coronavirus (COVID-19) affects the cardiovascular system, whether by direct viral aggression or indirectly through systemic inflammation and multiple organ compromise. A widely used method to determine cardiac injury is troponin measurement. The aim of this study is to evaluate the prevalence of cardiac involvement (CINV) in a population recovered from COVID-19, referred to cardiac MRI (CMR), who did not present troponin elevation. Methods There were 156 patients that recovered from COVID-19 and who did not present troponin elevation referred to CMR. CINV was considered to be the presence of: late gadolinium enhancement (LGE), edema, myocarditis, pericarditis, left ventricular systolic dysfunction (LVSD) and/or depressed right ventricular systolic dysfunction (RVSD). Results Prevalence of CINV was 28.8%, being more frequent in men (p = 0.002), in patients who required hospitalization (p = 0.04) and in those who experienced non-mild cases of infection (p = 0.007). RVSD (17.9%) and LVSD (13.4%) were the most frequent findings. The rate of myocarditis was 0.6%. LGE manifested in 7.1% of patients and its presence was related to less left ventricular ejection fraction (LVEF) (p = 0.0001) and right ventricular ejection fraction (RVEF) (p = 0.04). Conclusion In patients who recovered from COVID-19, 28.8% of CINV was found. It was more frequent in men, in patients who required admission and in patients with cases of non-mild infection. The patients that presented LGE had less LVEF and RVSF.

External validation of an electrocardiography artificial intelligence-generated algorithm to detect left ventricular systolic function in a general cardiac clinic in Uganda

2020 ◽  
Vol 41 (Supplement_2) ◽  
Author(s):  
C.K Mondo ◽  
Z.I Attia ◽  
E.D Benavente ◽  
P Friedman ◽  
P Noseworthy ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Heart Failure ◽  
Negative Predictive Value ◽  
Predictive Value ◽  
Left Ventricular Systolic Dysfunction ◽  
External Validation ◽  
Left Ventricular ◽  
Receiver Operating Curve ◽  
Left Ventricular Ejection ◽  
Ventricular Ejection Fraction

Abstract Background Left ventricular systolic dysfunction (LVSD) is associated with increased morbidity and mortality. Although there are effective treatments for patients with LVSD to prevent mortality, heart failure and to improve symptoms, many patients remain undetected and untreated. We have recently derived a deep learning algorithm to detect LVSD using the electrocardiogram (ECG) which could have an important screening role, particularly in limited resources settings. We evaluated the accuracy of this algorithm for the first time in Africa in a sample of subjects attending a cardiology clinic. Methods We conducted a retrospective study in a general cardiac clinic in Uganda. Consecutive patients ≥18 years who had a digital ECG and echocardiogram done within two days of each other were included. We excluded patients with pacemakers or missing information regarding left ventricular ejection fraction (LVEF). Routine 10-second, twelve-lead surface rest ECG were performed using an Edan PC ECG Model SE-1515, Hamburg, Germany. The probability of LVSD was estimated with the Mayo Clinic artificial intelligence (AI) ECG algorithm. LVEF was calculated by the MMode (Teichholz method) using a Philips Ultrasound system, HD7XE, Bothel, Washington, USA. LVSD was defined as a LVEF≤35%. We assessed the overall diagnostic performance of the algorithm to identify LVSD in this population with the area under the receiver operating curve (AUC), and estimated sensitivity, specificity and accuracy using a pre-specified cut-off based on the probability for LVSD generated by the algorithm. We conducted secondary analyses using different LVEF cutoff values. Results We included 634 subjects, 32% (200) of whom had hypertension and 12% (77) clinical heart failure. Mean age was 57±18.8 years, 58% were women and the overall prevalence of LVSD was 4%. The AI-ECG had an AUC of 0.866 (see figure below), sensitivity 73.08%, specificity 91.10%, negative predictive value 98.75%, positive predictive value 26.03% and an accuracy of 90.96% using the original threshold. Using the optimal cutoff based on the AUCs, the sensitivity was 80.77% and specificity was 81.05% with a negative predictive value of 98.99%. The ROC for the detection of LVEF of 40% or below was 0.821. Conclusion The Mayo AI-ECG algorithm demonstrated good accuracy, sensitivity and specificity to detect LVSD in patients seen in a clinical setting in Uganda. This tool may facilitate the identification of people at a high risk for LVSD in settings with low resources. ROC Funding Acknowledgement Type of funding source: None

Sign in / Sign up

Export Citation Format

Share Document