Enhanced detection of chronic obstructive pulmonary disease via exhaled breath analysis: internet of things and electronic nose system

Abstract

Chronic obstructive pulmonary disease (COPD) remains a major global health burden, highlighting the need for accessible, non-invasive screening tools. This study aims to develop a portable, real-time internet of things (IoT)-integrated electronic nose (e-nose) system for COPD detection using exhaled volatile organic compounds (VOCs). Breath samples from 44 participants (healthy, smokers, and COPD) were analyzed using a MOS based e-nose, and four machine-learning classifiers were evaluated. Data were processed through cloud-based pipelines enabling real-time acquisition and automated analysis. The random forest (RF) model achieved the highest performance (accuracy 86%) in distinguishing COPD-related VOC patterns. This approach overcomes limitations of earlier offline Tedlar-bag methods by enabling direct, real-time breath analysis. The prototype dashboard provides immediate visualization for potential remote monitoring. Key limitations include the small sample size and non-standardized breath sampling, which may affect VOC variability. Overall, this work contributes a cost-effective, portable, IoT-enabled framework demonstrating the feasibility of real-time VOC analysis for early COPD screening and future integration into telehealth and community-based diagnostics.