Blood pressure changes during routine transthoracic echocardiography.

Background: Increased afterload affects many of the flow dependent metrics assessed during transthoracic echocardiography (TTE) especially in the evaluation valvular disease. A single timepoint blood pressure (BP) may not accurately reflect the afterload present at the time of flow-dependent imaging and quantification. We assessed the magnitude of change in BP at discrete timepoints during routine TTE. Method: We conducted a prospective study where participants underwent automated BP measurement while undergoing a clinically indicated TTE. The first reading was obtained right after the patient lay supine and subsequent readings were taken at 10-min intervals during image acquisition. Result: We included 50 participants (66% were male, with a mean age of 64 years). After 10 min, 40 (80%) participants had a drop in systolic BP of >10 mmHg. Compared to the baseline, there was a significant drop in systolic BP (mean decrease 20.0 ± 12.8 mmHg; P < 0.05), and diastolic BP (mean decrease 15.7 ± 13.2 mmHg; P < 0.05) at 10 min. The systolic BP remained different from the baseline value throughout the duration of the study (average decrease from baseline to study end was 12.4 ± 16.0 mmHg, p < 0.05). Conclusion: BP recorded just prior to TTE does not accurately reflect the afterload present during most of the study. This finding has important implications for valvular heart disease imaging protocols that incorporate flow dependent metrics, where the presence or absence of hypertension may lead to under- or over-estimation of disease severity.

Automated Detection of Aortic Stenosis Using Machine Learning.

BACKGROUND: Aortic stenosis (AS) is a degenerative valve condition that is underdiagnosed and undertreated. Detection of AS using limited two-dimensional echocardiography could enable screening and improve appropriate referral and treatment of this condition. The aim of this study was to develop methods for automated detection of AS from limited imaging data sets. METHODS: Convolutional neural networks were trained, validated, and tested using limited two-dimensional transthoracic echocardiographic data sets. Networks were developed to accomplish two sequential tasks: (1) view identification and (2) study-level grade of AS. Balanced accuracy and area under the receiver operator curve (AUROC) were the performance metrics used. RESULTS: Annotated images from 577 patients were included. Neural networks were trained on data from 338 patients (average n = 10,253 labeled images), validated on 119 patients (average n = 3,505 labeled images), and performance was assessed on a test set of 120 patients (average n = 3,511 labeled images). Fully automated screening for AS was achieved with an AUROC of 0.96. Networks can distinguish no significant (no, mild, mild to moderate) AS from significant (moderate or severe) AS with an AUROC of 0.86 and between early (mild or mild to moderate AS) and significant (moderate or severe) AS with an AUROC of 0.75. External validation of these networks in a cohort of 8,502 outpatient transthoracic echocardiograms showed that screening for AS can be achieved using parasternal long-axis imaging only with an AUROC of 0.91. CONCLUSION: Fully automated detection of AS using limited two-dimensional data sets is achievable using modern neural networks. These methods lay the groundwork for a novel method for screening for AS.

Accuracy of echocardiographic estimations of right heart pressures in adult heart transplant recipients.

BACKGROUND: Accurate assessment of right atrial pressure (RAP) and pulmonary artery systolic pressure (PASP) is critical in the management of heart transplant recipients. The accuracy of echocardiography in estimating these pressures has been debated. OBJECTIVE: To assess the correlation and agreement between echocardiographic estimations of right heart pressures with those of respective invasive hemodynamic measurements by right heart catheterization (RHC) in adult heart transplant recipients. METHODS: This is a prospective evaluation of 84 unique measurements from heart transplant recipients who underwent RHC followed by standard echocardiographic evaluation within 159 ± 64 min with no intervening medication changes. The relationship between noninvasive pressure estimations and invasive hemodynamic measurements was examined. RESULTS: Mean RAP was 7 ± 5 mmHg and mean PASP was 33 ± 8 mmHg by RHC. There was no significant correlation between echocardiographic estimation of RAP and invasive RAP (Spearman's rho = -0.05, p = .7), and no significant agreement between these two variables (weighted kappa = -0.1). There was a modest correlation between echocardiographic estimation of PASP and invasive PASP (r = .39, p = .002). Bland-Altman analysis showed a mean bias of 2.1 ± 9 mmHg (limits of agreement = -15 to 20 mmHg). CONCLUSION: In heart transplant recipients, there is no significant correlation or agreement between echocardiographic RAP estimation and invasively determined RAP. Noninvasive PASP estimation correlates significantly but modestly with invasively measured PASP. Further refinement of echocardiographic methods for assessment of RAP is warranted in this unique patient population.