Advancing telemedicine in pulmonology: acoustic-waveform respiratory evaluation (AWARE) via sensing and machine learning on smartphones

ABSTRACT Asthma and COPD respectively affect 25 million and 16 million people in the US. Chronic lower respiratory diseases represented the fourth leading cause of death in the country before the pandemic. For these and other pulmonary diseases like cystic fibrosis (CF), monitoring disease remotely but objectively via telemedicine could lead to marked improvements in disease control, quality of life, and overall prognosis. However, current disease monitoring and management often rely on subjective symptom report, and objective pulmonary function tests (PFTs) are often only done once or twice a year at subspecialty referral centers. There are very few solutions to objectively evaluate lung health out of clinic, and existing portable PFTs have several key limitations: they require additional equipment that is expensive, cumbersome, or needs frequent calibration; and/or they rely on forced maximal respiratory maneuvers, which require professional coaching and are particularly difficult for younger children, older adults, and those with advanced disease. Our highly innovative solution, AWARE (Acoustic WAveform Respiratory Evaluation), aims to fundamentally advance pulmonary telemedicine by inventing new acoustic sensing and machine learning techniques that transform everyday smartphones from a video chat client to a fully functional pulmonary telemedicine examination device. We hypothesize that AWARE can accurately identify specific airway diseases, closely estimate lung function, and adaptively detect deviations from normal that are associated with disease exacerbations. We will test these hypotheses by recruiting a cohort of 750 subjects and completing three specific aims. In Aim 1, we will improve AWARE’s sensing approach to accurately differentiate participants without chronic airway disease, patients with asthma, COPD, CF, and other airway diseases. In Aim 2, we will improve AWARE’s accuracy in estimating lung function indices from spirometry and airway oscillometry, both in healthy participants and in those with respiratory diseases. In Aim 3, we will develop new techniques to evaluate within-patient changes in AWARE between stable vs acute exacerbations and provide explainability to the correlation between these cases by leveraging explainable AI techniques. AWARE will provide an innovative, portable, low-cost, non- invasive, accurate, reliable, and easy-to-use approach to aid in pulmonary disease diagnosis, lung function monitoring, and detection of disease exacerbations, providing personalized reports for clinician’s review and interpretation. Of direct relevance to crises like COVID19, tele-health monitoring with AWARE would reduce the risk for vulnerable populations by avoiding in-person clinic visits and use of shared PFT equipment.