Single-lead arrhythmia detection through machine learning: cross-sectional evaluation of a novel algorithm using real-world data

Abstract

Background: Computer-assisted interpretation of single-lead ECG is the preliminary method for clinicians to flag and further evaluate an arrhythmia of clinical importance for acutely ill patients. Critical scrutiny of novel detection algorithms is lacking, particularly in external real-world data sets. This study's objective was to evaluate a hybrid machine learning model's ability to classify eight arrhythmias from a single-lead ECG signal from acutely ill patients.

Methods: This cross-sectional external retrospective evaluation of a previously trained hybrid machine learning model against an ECG reading team in the setting of home hospital care (acute care delivered at home substituting for traditional hospital care) draws from patients admitted at two hospitals in Boston, Massachusetts, USA between 12 June 2017 and 23 November 2019. We calculated classifier statistics for each arrhythmia, all arrhythmias and strips where the model identified normal sinus rhythm.

Results: The model analysed 2 680 162 min of single-lead ECG data from 423 patients and identified 691 478 arrhythmias. Patients had a mean age of 70 years (SD, 18), 60% were female and 45% were white. For any arrhythmia, the model had a sensitivity of 98%, a specificity of 100%, an accuracy of 98%, a positive predictive value of 100%, a negative predictive value of 93% and an F₁ Score of 99%. Performance was best for pause (F₁ Score, 99%) and worst for paroxysmal supraventricular tachycardia (F₁ Score, 92%). The model's false positive rate for any arrhythmia was 0.2%, ranging from 0.4% for pause to 7.2% for paroxysmal supraventricular tachycardia. The false negative rate for any arrhythmia was 1.9%.

Conclusions: A hybrid machine learning model was effective at classifying common cardiac arrhythmias from a single-lead ECG in real-world data.

Keywords: ARRHYTHMIAS; Arrhythmias, Cardiac; Electrocardiography.

Figure 1

An acutely ill patient generates continuous single-lead ECG interpreted by traditional computer algorithms during usual care with only fair arrhythmia recognition, generating significant alarm fatigue due to false positives. The hybrid model evaluated in this study demonstrated performance comparable to gold standard annotators when challenged with real-world single-lead ECG data from acutely ill patients. Refer to online supplemental figure S1 for a more detailed schematic of the hybrid model. AFIB, atrial fibrillation; Brady, sinus bradycardia; DNN, deep neural network; NSR, normal sinus rhythm; PSVT, paroxysmal supraventricular tachycardia; PVC, premature ventricular contraction; Tachy, sinus tachycardia; V. Big, ventricular bigeminy; V. Trig, ventricular trigeminy.

Figure 2

Example ECGs demonstrating arrhythmias. (A) Atrial fibrillation. (B) Sinus bradycardia. (C) Normal sinus rhythm. (D) Pause. (E) Paroxysmal supraventricular tachycardia. (F) Sinus tachycardia. (G) Ventricular bigeminy. (H) Ventricular trigeminy.

Figure 3

Confusion matrices for each arrhythmia and any arrhythmia. Colours indicate proportion of patients in each category. (A) Any arrhythmia. (B) Atrial fibrillation. (C) Pause. (D) Paroxysmal supraventricular tachycardia. (E) Premature ventricular contraction. (F) Sinus bradycardia. (G) Sinus tachycardia. (H) Ventricular bigeminy. (I) Ventricular trigeminy. (J) Mean of specific arrhythmias. Note: Some arrhythmias show fewer than 3000 strips total due to some strips being duplicated between both the arrhythmia-containing and normal sinus rhythm categories.