Automated left ventricular dimension assessment using artificial intelligence

Background and purpose Artificial intelligence (AI) has the potential to greatly improve efficiency and reproducibility of quantification in echocardiography, but to gain widespread use it must both meet expert standards of excellence and have a transparent methodology. We developed an online platform to enable multiple collaborators to annotate medical images for training and validating neural networks. Methods Using our online collaborative platform 9 expert echocardiographers labelled 2056 images that comprised the training dataset. They labelled the four points from where the standard parasternal long axis (PLAX) measurements (interventricular septum, posterior wall, left ventricular dimension) would be made. Using these labelled images we trained a 2d convolutional neural network to replicate these labels. Separately, we curated an external validation dataset of the systolic and diastolic frames of 100 PLAX acquisitions. Each of these images were labelled twice by 13 different experts, and the average of the 26 measurements was taken as the consensus standard. We then compared the individual experts and the AI measurements on the external validation dataset to the consensus standard, and calculated the precision standard deviation (SD) of the signed differences from the consensus standard. Results For diastolic septum thickness, the AI had a precision SD of 1.8 mm (ICC 0.81; 95% CI 0.73 to 0.97), compared with 2.0 mm for the individual experts (ICC 0.64; 95% CI 0.57 to 0.72). For diastolic posterior wall thickness, the AI had a precision SD 1.4 mm (ICC 0.54; 95% CI 0.38 to 0.66), and the individual experts 2.2 mm (ICC 0.37; 95% CI 0.29 to 0.46). The AI's precision SD for left ventricular internal dimension was 3.5 mm (ICC 0.93, 95% CI 0.90 to 0.94), and for individual experts was 4.4mm (ICC 0.82, 95% CI 0.78 to 0.95). Both the experts and AI performed better in diastole than systole (precision SD AI 2.5mm vs 4.3mm, p<0.0001; experts 3.3mm vs 5.3mm, p<0.0001). Conclusions AI trained by a group of echocardiography experts was able to perform PLAX measurements which matched the reference standard more closely than any individual expert's own measurements. This open, collaborative approach may be a model for the development of AI that is explainable to, and trusted by clinicians. FUNDunding Acknowledgement Type of funding sources: Public grant(s) – National budget only. Main funding source(s): NIHR Imperil BRC ITMATDr Howard was additionally funded by Wellcome. Online collaborative platform Results of AI and experts

Echocardiographic prognostication of new onset heart failure in a cohort of ischemic heart disease patients

Background Echocardiographic predictors for new onset heart failure in patients with ischemic heart disease with reduced left ventricular ejection fraction (HFrEF) or with preserved left ventricular ejection fraction (HFpEF) in Ethiopian and Sub-Saharan African is not well-known. Methods Two hundred twenty-eight patients with ischemic heart disease were retrospectively recruited and followed. Analysis on baseline clinical and echocardiographic characteristics of patients, and risk factors for new onset HFpEF and new onset HFrEF were done. The exclusion criteria were known heart failure at baseline and those who did not have echocardiography data. Results During the follow up period, heart failure developed in 62.2% (61/98) of ischemic heart disease patients with preserved left ventricular ejection fraction and in 70.1% (92/130) of ischemic heart disease patients with reduced left ventricular ejection fraction. We did not find significant difference between HFrEF and HFpEF in time to new onset heart failure. Systolic blood pressure, diastolic blood pressure, diabetes, left atrium and diastolic left ventricular dimension had significant association with new onset HFrEF on univariate regression analysis. Whereas new onset HFpEF was significantly associated with age, sex, presence of hypertension, Systolic blood pressure and diastolic left ventricular dimension. On cox regression analysis diastolic left ventricular dimension was associated with both new onset HFpEF and HFrEF. Age, diabetes, and dimension of left atrium were also associated with HFrEF. Conclusion This cohort study in ischemic heart disease patients suggests a key role for the diastolic left ventricular dimension, left atrium size, diabetes, sex and age as predictors of new onset HFrEF and HFpEF. Strategies directed to prevention and early treatment of diabetes, dilatation of left ventricle and left atrium may prevent a considerable proportion of HFrEF or HFpEF.

Proportionate or disproportionate secondary mitral regurgitation: how to untangle the Gordian knot?

Recent randomised percutaneous mitral intervention trials in patients with heart failure with secondary mitral regurgitation (SMR) have yielded contrasting results. A ‘relative load’ or ‘proportionality’ conceptual framework for SMR has been proposed to partly explain the disparate results. The rationale behind the framework is that SMR depends on the left ventricular dimension and not vice versa. In this review, we provide an in-depth analysis of the proportionality parameters used in this framework and also discuss the regurgitant fraction. We also consider haemodynamic observations in SMR that may affect the interpretation and comparisons among proportionality parameters. The conclusion is that the proportionality concept remains hypothetical and requires prospective validation before envisaging its use at individual patient level for risk stratification or therapeutic decision-making.

Pulmonary Vein Enlargement as an Independent Predictor for New-Onset Atrial Fibrillation

Pulmonary vein (PV) enlargement is associated with atrial fibrillation (AF). However, the predictive value of PV volume for new-onset AF has not been determined. We retrospectively assessed and enrolled non-AF subjects who underwent echocardiography and cardiac CT angiography (CCTA) around the same time and evaluated the development of AF longitudinally. PV volume was assessed by estimating the three-dimensional CCTA-derived mid-diastolic PV volume from the ostium to tertiary branches. Overall, 1105 subjects were enrolled. Among them, 29 developed AF during a mean follow-up of 4.28 ± 3.08 years after baseline CCTA and echocardiography. The AF group had a higher proportion of older aged subjects, a higher ratio of early mitral flow velocity (E) to early mitral annular tissue velocity (Em), higher Em, and larger left atrial (LAVI) and PV (PVVI) volume indices. PVVI was independently associated with male sex, left ventricular dimension, E/Em and LAVI. AF incidence increased markedly across each baseline PVVI tertile (2.2%, 5.1%, and 10.8%). In the multivariate Cox model, increased PVVI was independently associated with new-onset AF (hazard ratio (HR) = 5.401, 4.931–6.193, p = 0.007). Based on the analysis of multimodal cardiac imaging, our results provide mechanistic insights into PV remodeling and its potential role as a link between diastolic dysfunction and developing AF.