Journal: Physiological Measurement

Article Title: Integrating multimodal information in machine learning for classifying acute myocardial infarction

doi: 10.1088/1361-6579/acc77f

Figure Lengend Snippet: Breakdown of numbers of ECG recordings across different diagnostic statements.

Article Snippet: Approximately 50%–55% of patients with acute myocardial ischemia and 40% with unstable angina do not have ST-segment elevation or depression present in their initial ‘snap-shot’ ten-second 12-lead ECG recording (Forberg et al , Rouan et al ).

Techniques: Diagnostic Assay

Journal: Physiological Measurement

Article Title: Integrating multimodal information in machine learning for classifying acute myocardial infarction

doi: 10.1088/1361-6579/acc77f

Figure Lengend Snippet: Schematic diagram of the ECG-only model (yellow) and the Multimodal model (yellow + blue).

Article Snippet: Approximately 50%–55% of patients with acute myocardial ischemia and 40% with unstable angina do not have ST-segment elevation or depression present in their initial ‘snap-shot’ ten-second 12-lead ECG recording (Forberg et al , Rouan et al ).

Techniques:

Journal: Physiological Measurement

Article Title: Integrating multimodal information in machine learning for classifying acute myocardial infarction

doi: 10.1088/1361-6579/acc77f

Figure Lengend Snippet: Performance comparison (in AUROC) between ECG-only models trained with different duration lengths of ECG waveforms (*** indicates p < 0.001 ).

Article Snippet: Approximately 50%–55% of patients with acute myocardial ischemia and 40% with unstable angina do not have ST-segment elevation or depression present in their initial ‘snap-shot’ ten-second 12-lead ECG recording (Forberg et al , Rouan et al ).

Techniques: Comparison

Journal: Physiological Measurement

Article Title: Integrating multimodal information in machine learning for classifying acute myocardial infarction

doi: 10.1088/1361-6579/acc77f

Figure Lengend Snippet: Comparison of the receiver operating characteristic (ROC) curves from the Multimodal and ECG-only models. Dashed lines represent ROC curves from each cross-validation step and solid lines represent the mean ROC curves.

Article Snippet: Approximately 50%–55% of patients with acute myocardial ischemia and 40% with unstable angina do not have ST-segment elevation or depression present in their initial ‘snap-shot’ ten-second 12-lead ECG recording (Forberg et al , Rouan et al ).

Techniques: Comparison, Biomarker Discovery

Journal: Physiological Measurement

Article Title: Integrating multimodal information in machine learning for classifying acute myocardial infarction

doi: 10.1088/1361-6579/acc77f

Figure Lengend Snippet: Breakdown of false positives across four subclasses in the non-MI condition. Abbreviations: FPR = false positive rate; NORM = normal ECG; STTC = ST/T change; HYP = hypertrophy; CD = conduction disturbance.

Article Snippet: Approximately 50%–55% of patients with acute myocardial ischemia and 40% with unstable angina do not have ST-segment elevation or depression present in their initial ‘snap-shot’ ten-second 12-lead ECG recording (Forberg et al , Rouan et al ).

Techniques:

Journal: Physiological Measurement

Article Title: Integrating multimodal information in machine learning for classifying acute myocardial infarction

doi: 10.1088/1361-6579/acc77f

Figure Lengend Snippet: Feature importance (in SHAP value) in the Multimodal model. (a) Comparison of feature importance among three feature categories. (b) Feature importance in ECG tracings. The ECG in (b) is from a MI recording and the plot contains the first 5 s on Lead aVF.

Article Snippet: Approximately 50%–55% of patients with acute myocardial ischemia and 40% with unstable angina do not have ST-segment elevation or depression present in their initial ‘snap-shot’ ten-second 12-lead ECG recording (Forberg et al , Rouan et al ).

Techniques: Comparison

Journal: Cardiovascular Digital Health Journal

Article Title: HRS White Paper on Clinical Utilization of Digital Health Technology

doi: 10.1016/j.cvdhj.2021.07.001

Figure Lengend Snippet: Available wearable technology

Article Snippet: CardioComm Solutions HeartCheck , ECG , 30 seconds , FDA cleared , AA battery , Patient triggered , Handheld , , .

Techniques: Activity Assay, Battery