Generalised Label-free Artefact Cleaning for Real-time Medical Pulsatile Time Series

2025-04-29

AI Summary

Artifacts in hemodynamic time-series signals (e.g., arterial blood pressure [ABP], photoplethysmography [PPG]) severely compromise clinical decision-making; however, existing supervised artifact removal methods suffer from poor generalizability across intra-/inter-patient distribution shifts, heterogeneous monitoring devices, and diverse pathologies. To address this, we propose GenClean—the first label-free, patient-adaptive online artifact cleaning framework. It integrates waveform prior–driven unsupervised anomaly detection, patient-level feature disentanglement via representation learning, and a lightweight real-time inference architecture. GenClean supports multimodal hemodynamic signals (ABP/PPG) and deploys seamlessly on ICM+ clinical monitoring systems with end-to-end latency <10 ms. Evaluated on 180K ABP samples and the MIMIC-III multi-pathology cohort, GenClean significantly outperforms state-of-the-art methods. Clinical evaluation shows a 32% improvement in inter-rater agreement, demonstrating its robust generalizability and clinical utility.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Natural Language Processing: Generation
• Multiagent Systems: Adversarial Agents

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Web Mining and Content Analysis: Web data integration and cleaning
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Artefacts compromise clinical decision-making in the use of medical time series. Pulsatile waveforms offer probabilities for accurate artefact detection, yet most approaches rely on supervised manners and overlook patient-level distribution shifts. To address these issues, we introduce a generalised label-free framework, GenClean, for real-time artefact cleaning and leverage an in-house dataset of 180,000 ten-second arterial blood pressure (ABP) samples for training. We first investigate patient-level generalisation, demonstrating robust performances under both intra- and inter-patient distribution shifts. We further validate its effectiveness through challenging cross-disease cohort experiments on the MIMIC-III database. Additionally, we extend our method to photoplethysmography (PPG), highlighting its applicability to diverse medical pulsatile signals. Finally, its integration into ICM+, a clinical research monitoring software, confirms the real-time feasibility of our framework, emphasising its practical utility in continuous physiological monitoring. This work provides a foundational step toward precision medicine in improving the reliability of high-resolution medical time series analysis

Problem

• Detects and cleans artefacts in medical pulsatile time series without labels
• Addresses patient-level distribution shifts in artefact detection
• Ensures real-time feasibility for diverse medical signals like ABP and PPG

Innovation

• Label-free framework for real-time artefact cleaning
• Leverages large in-house ABP dataset for training
• Validated on cross-disease cohorts and diverse signals