The multi-modal smart chest patch system integrated with a multi-criteria fusion machine learning model achieved a classification accuracy of 87% (AUC > 0.92) for detecting cardiopulmonary anomalies.
Observational (n=475)
Does a multimodal smart chest patch (SCP) system combined with a machine learning model improve the detection of cardiac and respiratory anomalies compared to conventional methods?
A novel, lightweight multimodal smart chest patch combined with a machine learning model accurately detects cardiopulmonary anomalies and monitors real-time physiological shifts during exercise.
Effect estimate: AUC >0.92
Abstract Cardiopulmonary homeostasis is critical for health, with disruptions often indicating early-stage pathologies. However, current wearable devices, constrained by their limited focus on single or few modalities, fall short in capturing the full spectrum of these critical interactions. In this work, we present a multimodal smart chest patch (SCP) system integrating flexible sensing modules with a multi-criteria, multimodal fusion (MCMF) machine learning model. The patch (5.4 g and 3.6 mm) simultaneously monitors electrocardiogram (ECG), heart sound (HS), and respiratory (Resp) signals, enabling real-time extraction of 12 cardiopulmonary parameters. Compared with commercial devices, our flexible patch maintains stable signal quality and demonstrates broad applicability across diverse individuals. The proposed MCMF model was further validated on an extended dataset of 5,561 recordings from 475 participants and achieved a classification accuracy of 87% for detecting cardiac and respiratory anomalies, surpassing conventional methods. Our results enable real-time exercise monitoring, revealing dynamic physiological shifts (ΔHR = 21 bpm, ΔPEP = −30 ms) with superior signal fidelity, and therefore holding significant promise for scalable, personalized health management.
Qiu et al. (Thu,) conducted a observational in Cardiopulmonary anomalies (n=475). Multi-modal smart chest patch (SCP) system vs. Commercial devices and conventional machine learning methods was evaluated on Classification accuracy for detecting cardiac and respiratory anomalies (AUC >0.92). The multi-modal smart chest patch system integrated with a multi-criteria fusion machine learning model achieved a classification accuracy of 87% (AUC > 0.92) for detecting cardiopulmonary anomalies.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: