The majority of patients with severe aortic stenosis (AS) had right precordial U-waves on the 12-lead ECG and in most of these patients there was diminution of the U-wave amplitude after TAVR, according to results of a study presented April 28 at Heart Rhythm 2022 and simultaneously published in the Journal of the American College of Cardiology.

Previous work using a convolutional neural network (CNN), a common deep learning model in artificial intelligence algorithms, showed that using just the 12-lead ECG it was possible to identify patients with moderate or severe AS. U-Waves on the 12-lead ECG were identified as the key feature to identify AS.

In the present study, Joshua Lampert, MD; Valentin Fuster, MD, PhD, MACC; Vivek Y. Reddy, MD, et al., assessed the real-world prevalence of U-waves in severe AS and the effect of TAVR on dynamic U-wave changes.

A total of 168 consecutive patients undergoing TAVR at Mount Sinai Hospital in New York between August and December 2021 were screened and those with echocardiographically severe AS were included and those with low-flow, low-gradient AS were excluded. For this analysis, 50 patients were identified using clinical, laboratory, electrocardiographic and echocardiographic data extracted from the electronic medical record. The presence of U-waves was also assessed in 50 patients from the internal medicine service matched for age, gender and history of hypertension as a control group.

U-waves were defined as a positive deflection in leads V1-V3 after the T-wave within a distinct T-P segment not due to artifact. U-wave diminution was defined as a reduction in amplitude of visible U-waves between the ECGs obtained before and after TAVR.

Results showed U-waves were present in 31 of 50 patients with severe AS (62%) before TAVR. Of these patients, the U-waves were diminished in 87% after TAVR; in 7 of 31 (22.5%) the amplitude was diminished but still measurable, and in 20 of 31 (64.5%) the U-Wave resolved completely. The mean U-wave amplitude was 0.48 mm before TAVR and 0.12 mm after TAVR. None of the patients without U-waves before TAVR developed it after TAVR.

No significant differences were seen in baseline clinical variables. No significant difference was seen in mean systolic blood pressure before and after TAVR, however a significant decrease was seen in mean diastolic blood pressure after TAVR (70 mm Hg vs. 65 mm Hg before; p=0.03). A decrease was seen in the mean aortic valve gradient from 53 mm Hg to 12 mm Hg (p<0.01).

In the control population, U-waves were significantly less prevalent, seen in only three of 50 patients and all three exhibited diastolic dysfunction on echocardiography.

Noting the retrospective design and lack of supplementary invasive electrophysiologic or hemodynamic assessment as limitations to the analysis, they also note the manual data extraction buttresses data quality. Moreover, no conclusions can be drawn about the effect of the U-wave diminution on patient outcomes because of the lack of long-term follow-up.

Lampert, et al, write their finding of diminished U-wave amplitude after TAVR is consistent with the hypothesis that U-waves reflect ventricular wall stress in AS. "This is one of the first examples of clinical translation of machine learning-derived feature importance into a real-world clinical setting," they write.

View all of the JACC and JACC: Clinical Electrophysiology research simultaneously publishing at Heart Rhythm 2022 here.

Clinical Topics: Cardiac Surgery, Invasive Cardiovascular Angiography and Intervention, Noninvasive Imaging, Valvular Heart Disease, Aortic Surgery, Cardiac Surgery and VHD, Interventions and Imaging, Interventions and Structural Heart Disease, Echocardiography/Ultrasound, Arrhythmias and Clinical EP

Keywords: Transcatheter Aortic Valve Replacement, Artificial Intelligence, Data Accuracy, Deep Learning, Electrocardiography, Echocardiography, Aortic Valve Stenosis

< Back to Listings