Population-Based Reference Values for 3D Echocardiographic LV Volumes and Ejection Fraction

Abstract Background The left ventricle (LV) in obese patients undergoes different patterns of remodeling in order to normalize wall stress. However, little is known about how LV volume indices, LV global longitudinal strain and right ventricular free wall strain (GLS) vary according to the pattern of LV remodeling. Aim To define the echocardiographic reference values of LV volumes and biventricular GLS across the different LV remodeling patterns in obese patients with a preserved ejection fraction. Methods 2393 adult obese patients (1428 females, 965 males) with a normal ejection fraction who underwent echocardiography from January 2008 to December 2018 were selected. They were categorized according to 4 cardiac remodeling groups defined by LV mass index (102g/m2 in males, 88g/m2 in females) and relative ventricular wall thickness (0.42): normal geometry (NG), eccentric hypertrophy (EH), concentric remodeling (CR) and concentric hypertrophy (CH). Obese subjects were further categorized by BMI class (30–35, 35–40, >40 kg/m2). Obese subjects were gender matched to controls with a normal BMI (18.5–25 kg/m2) and normal cardiac geometry. Mean ± SD, One-way Anova and Tukey- Kramer HSD were applied. P<0.05 is considered significant. Results The mean age of controls and obese patients' were 50±16 and 57±13.6 years respectively (P<0.0001). LV GLS for controls compared to obese subjects with NG, EH, CR and CH was −21.1±2 vs. −20.2±1.9, −19.6±2.8, −18.5±2.9, −17.5±3.4 respectively (p<0.0001 for all), and for RV GLS it was −27.9±4 vs −26.7±3.9, −25.1±5, −23.5±5.5, −24.1±5.2 respectively (p<0.01 for all, except for NG where p=0.2). The distribution of LV indices according to cardiac remodeling subtypes is shown in the figure. Indexed end diastolic and end systolic volumes were smaller in NG, CH and CR compared to controls (p<0.001 for each respectively). LV GLS and ejection fraction were higher in females, while indexed LV volumes were higher in males within each remodeling category (P<0.0001). No significant difference in LV GLS or indexed LV volume was seen across BMI categories within each remodeling pattern (P>0.05). Obese subjects with CH had the highest incidence of the cardiovascular risk factors hyperlipidemia, hypertension and history of myocardial infarction or stroke, compared to those with other remodeling patterns (p<0.0001 for each, vs. NG, EH and CR). Conclusion To our knowledge, this is the largest study to define LV volumes and left and right ventricular GLS according to LV remodeling pattern and BMI category. The Lowest GLS was noted in CH. Ejection fraction was similar across the LV remodeling patterns. There were no differences in GLS and LV indexed volumes across BMI categories within each remodeling group. These results can be applied as a reference values for the obese population with a normal LV ejection fraction. Funding Acknowledgement Type of funding source: None

Download Full-text

Automated Echocardiographic Quantification of Left Ventricular Ejection Fraction Without Volume Measurements Using a Machine Learning Algorithm Mimicking a Human Expert

Circulation Cardiovascular Imaging ◽

10.1161/circimaging.119.009303 ◽

2019 ◽

Vol 12 (9) ◽

Cited By ~ 12

Author(s):

Federico M. Asch ◽

Nicolas Poilvert ◽

Theodore Abraham ◽

Madeline Jankowski ◽

Jayne Cleve ◽

...

Keyword(s):

Machine Learning ◽

Ejection Fraction ◽

Reference Values ◽

Learning Algorithm ◽

Left Ventricular ◽

Machine Learning Algorithm ◽

Mean Absolute Deviation ◽

Limits Of Agreement ◽

Absolute Deviation ◽

Lv Volumes

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.

Download Full-text

P4347Automated echocardiographic quantification of left ventricular ejection fraction without volume measurements using a machine learning algorithm mimicking a human expert

European Heart Journal ◽

10.1093/eurheartj/ehz745.0755 ◽

2019 ◽

Vol 40 (Supplement_1) ◽

Author(s):

F M Asch ◽

N Poilvert ◽

T Abraham ◽

M Jankowski ◽

J Cleve ◽

...

Keyword(s):

Machine Learning ◽

Ejection Fraction ◽

Reference Values ◽

Learning Algorithm ◽

Left Ventricular ◽

Left Ventricular Ejection ◽

Machine Learning Algorithm ◽

Computer Algorithms ◽

Human Expert ◽

Lv Volumes

Abstract Background Echocardiographic quantification of left ventricular (LV) ejection fraction (EF) relies on either manual or automated identification of endocardial boundaries followed by standard calculation of model-based 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 be developed, which circumvents border detection and instead estimates the degree of ventricular contraction, similar to a human expert trained on tens of thousands of images. Purpose This study was designed to test the feasibility and accuracy of this approach. Methods Machine learning algorithm was developed and trained on a database of >50,000 echocardiographic studies, including multiple apical 2- and 4-chamber views, to automatically estimate LVEF (AutoEF, BayLabs). Testing was performed on an independent group of 99 unselected patients, whose automated EF values were compared to 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 of bias and limits of agreement (LOA). Consistency was assessed by mean absolute deviation (MAD) among automated estimates based on different combinations of apical views. Finally, sensitivity and specificity of detecting of EF≤35% was 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 (MAD=2.9%) and excellent agreement with the reference values: r=0.95, bias=1.0%, LOA=±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%, LOA=±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 to reference values provided by an expert panel. Acknowledgement/Funding Bay Labs, Inc.

Download Full-text