Key result
A CNN-LSTM deep learning model predicted left ventricular hypertrophy from 12-lead ECGs with an AUC of 0.62 and 68% sensitivity, offering higher sensitivity than Cornell and Sokolow-Lyon criteria.
Why the study?
Current ECG criteria for left ventricular hypertrophy have low sensitivity, and deep learning techniques have rarely been applied to LVH diagnosis.
Does a CNN-LSTM deep learning model improve the detection of left ventricular hypertrophy on 12-lead ECG compared to standard voltage criteria in hospitalized patients?
Observational (n=2,316)
No
Does a CNN-LSTM deep learning model improve the detection of left ventricular hypertrophy on 12-lead ECG compared to standard voltage criteria in hospitalized patients?
Absolute Event Rate: 0.62% vs 0.57%
p-value: p=0.075
No takes yet. Share an insight, caveat, or question.
DL ECG models may boost LVH detection sensitivity in practice; leaves open prospective validation before clinical use.
Zhao et al. (2022) conducted an observational in Left ventricular hypertrophy (n=2,316). CNN-LSTM deep learning model vs. Cornell voltage criteria and Sokolow-Lyon voltage criteria was evaluated on Area under the curve (AUC) for predicting left ventricular hypertrophy (p=0.075). A CNN-LSTM deep learning model predicted left ventricular hypertrophy from 12-lead ECGs with an AUC of 0.62 and 68% sensitivity, offering higher sensitivity than Cornell and Sokolow-Lyon criteria.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: