Searching for the Best Machine Learning Algorithm for the Detection of Left Ventricular Hypertrophy from the ECG: A Review

Figures

• 

  Figure 1. The sensitivity of the ML algorithms for LVH. Lin & Lui 2020 [31], Sparapani et al., 2019 [32], De la Gar-za-Salazar et al., 2020 [33], Kwon et al., 2020 [34], De la Gar-za-Salazar et al., 2021 [35], Khurshid et al., 2021 [36], Sabovčik et al., 2021 [37], Angelaki et al., 2021 [38], Lim et al., 2021 [39], Zhao et al., 2022 [40], Sammani et al., 2022 [41], Kokubo et al., 2022 [42], Naderi et al., 2023 [43], Liu et al., 2023 [44].

• 

  Figure 2. The specificity of the ML algorithms for LVH. Lin & Lui 2020 [31], Sparapani et al., 2019 [32], De la Gar-za-Salazar et al., 2020 [33], Kwon et al., 2020 [34], De la Gar-za-Salazar et al., 2021 [35], Khurshid et al., 2021 [36], Sabovčik et al., 2021 [37], Angelaki et al., 2021 [38], Lim et al., 2021 [39], Zhao et al., 2022 [40], Sammani et al., 2022 [41], Kokubo et al., 2022 [42], Naderi et al., 2023 [43], Liu et al., 2023 [44].

• 

  Figure 3. The AUC of the ML algorithms for LVH in those studies that reported such data. Lin & Lui 2020 [31], Sparapani et al., 2019 [32], De la Gar-za-Salazar et al., 2020 [33], Kwon et al., 2020 [34], De la Gar-za-Salazar et al., 2021 [35], Khurshid et al., 2021 [36], Sabovčik et al., 2021 [37], Angelaki et al., 2021 [38], Lim et al., 2021 [39], Zhao et al., 2022 [40], Sammani et al., 2022 [41], Kokubo et al., 2022 [42], Naderi et al., 2023 [43].

• 

  Figure 4. The positive predictive value (PPV), negative predictive value (NPV) and accuracy of the ML algorithms for LVH in those studies that reported such data. Lin & Lui 2020 [31], Sparapani et al., 2019 [32], De la Gar-za-Salazar et al., 2020 [33], Kwon et al., 2020 [34], De la Gar-za-Salazar et al., 2021 [35], Khurshid et al., 2021 [36], Sabovčik et al., 2021 [37], Angelaki et al., 2021 [38], Lim et al., 2021 [39], Zhao et al., 2022 [40], Sammani et al., 2022 [41], Kokubo et al., 2022 [42], Naderi et al., 2023 [43], Liu et al., 2023 [44].

• 

  Figure 5. The sensitivity of the ML algorithms for LVH compared to standard ECG LVH criteria in studies that reported such data. Lin & Lui 2020 [31], Sparapani et al., 2019 [32], De la Gar-za-Salazar et al., 2020 [33], Kwon et al., 2020 [34], De la Gar-za-Salazar et al., 2021 [35], Khurshid et al., 2021 [36], Zhao et al., 2022 [40], Kokubo et al., 2022 [42], Liu et al., 2023 [44].

• 

  Figure 6. The specificity of the ML algorithms for LVH compared to standard ECG LVH criteria in studies that reported such data. Lin & Lui 2020 [31], Sparapani et al., 2019 [32], De la Gar-za-Salazar et al., 2020 [33], De la Gar-za-Salazar et al., 2021 [35], Khurshid et al., 2021 [36], Zhao et al., 2022 [40], Kokubo et al., 2022 [42], Liu et al., 2023 [44].

• 

  Figure 7. The accuracy of the ML algorithms for LVH compared to standard ECG LVH criteria in studies that reported such data. Lin & Lui 2020 [31], De la Gar-za-Salazar et al., 2020 [33], Kwon et al., 2020 [34], De la Gar-za-Salazar et al., 2021 [35], Kokubo et al., 2022 [42], Liu et al., 2023 [44].