Feature extraction tool using temporal landmarks in arterial blood pressure and photoplethysmography waveforms

This study presents an automatic feature extraction tool that first detects temporal location of landmarks within each cardiac cycle of ABP and PPG waveforms, including the systolic phase onset, systolic phase peak, dicrotic notch, and diastolic phase peak. Then, based on these landmarks, extracts 852 features per cardiac cycle, encompassing time-, statistical-, and frequency-domains. The tool's ability to detect landmarks was evaluated on the perioperative MLORD dataset comprising 17,327 patients and on real-time data collected from a patient monitor (retrospective analysis). When compared with markings by an experienced researcher, the tool demonstrated robust performance across both datasets, waveform types, and all four landmarks, achieving average F1-scores above 97% and error rates below 4%. This tool has significant potential for supporting clinical utilization of ABP and PPG waveform features and for facilitating feature-based machine learning models for various clinical applications where features derived from these waveforms play a critical role.