Table 3.
Characteristics of the included studies
| Study ID | Country and Year of Study | Study Design | Inclusion Criteria | Type of AI model | Participants | Diagnostic Accuracy | |||
|---|---|---|---|---|---|---|---|---|---|
|
|
|
||||||||
| Groups (n) | Gender (Male:Female) | Sensitivity (%) | Specificity (%) | AUC | |||||
| Ko, et al. | USA, 2020 | Retrospective cohort study | Patients aged >18 years old with a diagnosis of HCM based on standard diagnostic criteria who had at least 1 digital, standard, 10-s, 12-lead ECG acquired in the supine position (HCM group); Non-HCM patients who had an ECG and an echocardiogram within the same period of time (Non-HCM/control group) | Deep-learning (Convolutional neural network) | H: 612 N: 12788 |
H: (1703:1357) N: (36.385:27.556) |
87 | 90 | 0.96 |
| Siontis, et al. | USA, 2021 | Retrospective cohort study | Patients aged <18 years old with a diagnosis of HCM based on standard diagnostic criteria (HCM group); Non-HCM patients who had an ECG and an echocardiogram within the same period of time (Non-HCM/control group) | Deep-learning (Convolutional neural network) | H: 300 N: 18439 |
H: (205:95) N: (11844:6595) |
92 | 95 | 0.98 |
| Siontis, et al. | USA, 2023 | Retrospective cohort study | Adult patients with a confirmed diagnosis of Hypertrophic Cardiomyopathy (HCM), according to established diagnostic guidelines, and who have at least one digitally recorded 10-s, 12-lead ECG (HCM group); Non-HCM patients who had an ECG and an echocardiogram within the same period of time (Non-HCM/ control group) | Deep-learning (Convolutional neural network) | H: 100 N: 13294 |
H: (1706:1341) N: (36437:27489) |
80 | 84 | 0.90 |
| Siontis, et al. | USA, 2024 | Retrospective cohort study | Patients with a definite HCM diagnosis by standard European Society of Cardiology (ESC) and Americal College of Cardiology (ACC)/ American Heart Association (AHA) criteria and had at least one 12-lead ECG available in digital format (HCM group); Non-HCM patients who had an ECG and an echocardiogram within the same period of time (Non-HCM/control group) | Deep-learning (Convolutional neural network) | H: 773 N: 3867 |
H: (536:237) N: (36.385:27.556) | 83 | 88 | 0.92 |
| Hirota, et al. | Japan, 2024 | Retrospective cohort study | Adult patients with a confirmed diagnosis of HCM based on one of the following criteria: (1) interventricular septal thickness (IVST) ≥15 mm with no other causes of left ventricular hypertrophy; (2) IVST ≥13 mm with a family history of HCM; or (3) hypertrophy in the apex of the left ventricle (HCM group); Non-HCM patients who had an ECG and an echocardiogram within the same period of time (Non-HCM/control group) | Deep-learning (Convolutional neural network) | H: 140 N: 19030 |
H: (97:43) N: (11388:7642) |
76.4 | 81.4 | 0.854 |
HCM group (H) = Patients with both echocardiographic and clinical diagnosis of HCM; Non-HCM/control group (N) = Non-HCM patients who had an ECG and an echocardiogram within the same period of time. Abbreviations: ECG = Electrocardiogram, HCM = Hypertrophic cardiomyopathy, IVST = Interventricular septal thickness.