Screening for Left Ventricular Hypertrophy Using Artificial Intelligence Algorithms Based on 12 Leads of the Electrocardiogram—Applicable in Clinical Practice?—Critical Literature Review with Meta-Analysis

Figures

• 

  Figure 1. The structure of a search strategy (MEDLINE).

• 

  Figure 2. Diagnostic accuracy of AI algorithms and classical ECGs criteria. CNN—Convolutional Neural Network, DNN—Deep Neural Network, ENN—Extreme Learning Machine Neural Network; CNN-LSTM—Convolutional Neural Network—Long Short-Term Memory; SL—Sokolow-Lyon; Kwon, 2019 [34]; Kokubo, 2022 [33]; Cai, 2024 [36]; Ryu, 2023 [37]; Zhao, 2022 [12].

• 

  Figure 3. Risk of bias. (a) Summary panel; (b) Analysis of the studies: Cai, 2024 [36], Kokubo, 2022 [33], Kwon, 2019 [34], Liu, 2022 [35], Ryu, 2023 [37], Salazar, 2021 [38], Zhao, 2022 [12].

• 

  Figure 4. Results of meta-analysis. Cai, 2024 [36]; Kokubo, 2022 [33]; Kwon, 2019 [34]; Liu, 2022 [35]; Ryu, 2023 [37]; Salazar, 2021 [38]; Zhao, 2022 [12].

• 

  Figure 5. ROC Curve Overview. CNN — Convolutional Neural Network, DNN—Deep Neural Network, ENN—Extreme Learning Machine Neural Network; CNN-LSTM—Convolutional Neural Net-work—Long Short-Term Memory; SL—Sokolow-Lyon.