Search

This study aimed to evaluate whether quantitative analysis of wrist photoplethysmography (PPG) could detect atrial fibrillation (AF). Continuous electrocardiograms recorded using an electrophysiology recording system and PPG obtained using a wrist-worn smartwatch were simultaneously collected from patients undergoing catheter ablation or electrical cardioversion. PPG features were extracted from 10, 25, 40, and 80 heartbeats of the split segments. Machine learning with a support vector machine and random forest approach were used to detect AF. A total of 116 patients were evaluated. We annotated > 117 h of PPG. A total of 6475 and 3957 segments of 25-beat pulse-to-pulse intervals (PPIs) were annotated as AF and sinus rhythm, respectively. The accuracy of the 25 PPIs yielded a test area under the receiver operating characteristic curve (AUC) of 0.9676, which was significantly better than the AUC for the 10 PPIs (0.9453; P

BACKGROUND Atrial fibrillation (AF) is associated with an increased risk of stroke, heart failure, and all-cause mortality. The electrocardiogram (ECG)–based strategy of screening for AF has some limitations. Photoplethysmography (PPG) is used in AF detection algorithms and allows passive and continuous monitoring by modern wearable devices. OBJECTIVE The objective of this study was to investigate the following: (1) whether quantitatively analyzing wrist PPG waveforms can clearly distinguish AF from sinus rhythm and (2) to determine the appropriate data length of the PPG for feature extraction to optimize the PPG analytics program for AF detection. METHODS Continuous waveforms of ECG through an electrophysiology recording system and PPG signals through a wrist–worn smartwatch were simultaneously collected from patients undergoing catheter ablation or electrical cardioversion for AF. The PPG features (temporal, spectral, or morphological) were extracted from 10, 25, 40, or 80 heartbeats of split segments. Machine learning with a support vector machine (SVM) approach was used for detecting AF. The receiver operating characteristic (ROC) curves were determined to evaluate the diagnostic accuracy. RESULTS A total of 116 patients were evaluated. The mean age was 59.6±11.4 years and 32.8% were women. We collected and annotated more than 117 hours of PPG waveforms. A total of 6478 and 3957 segments of 25-beat pulse-to-pulse interval (PPI) were annotated as AF and sinus rhythm, respectively. A total of eight features were extracted to distinguish AF, including the PPI standard deviation (SD), PPI root-mean-squared standard deviation (RMSSD), Shannon Entropy with bin = 10, 100, 1000 (SE10, SE100, SE1000), moving average of 3 PPI SD, moving SD of 3 PPI RMSSD, and moving SD of maximum FFT frequency in 3 PPI. The accuracy of all the eight PPG features extracted from the 25 PPI achieved a test AUC (area under the receiver operating characteristic curve) which was significantly better than that from the 10 PPI (the AUC was 0.9676 versus 0.9453, respectively; P CONCLUSIONS This study demonstrated that quantitatively analyzing PPG waveforms can clearly discriminate the signals of AF from those of sinus rhythm. The appropriate data length of the PPG to optimize the PPG analytics program was 25 heartbeats. CLINICALTRIAL N/A