A Machine Learning Model to Detect Small Airway Narrowing Due to Exercise in Overweight and Obese Adults with Asthma

Regular physical exercise can improve lung function and asthma control, especially in individuals with asthma who are overweight or obese. Nevertheless, being active or exercising could be a risk factor for asthma exacerbations. Monitoring physiological signals during exercise may help to detect potential worsening in asthma and could be used to help persons with asthma to adjust their exercise. Our goal was to determine whether small airway narrowing due to exercise could be predicted using conveniently recorded data including participant’s demographics and respiratory-related signals. In this study, 10 adults with asthma and body mass index (BMI) over 25 kg/m
²
were asked to cycle for 10 minutes in a room at 20°C. During exercise, the respiratory signals were recorded using inductance plethysmography on the chest and abdomen. Before and after exercise, small airway narrowing was assessed based on respiratory impedance measured by forced oscillation technique. From the chest respiratory signal and demographic data, a total of four features have been obtained in the time domain. To detect airway narrowing, five different machine learning classifiers were fine-tuned and evaluated using a leave-one-subject-out cross-validation approach. The best model had an accuracy of 77.01%, a recall of 82.69%, a specificity of 68.57%, a precision of 79.63%, and an F1 score of 81.13%. These results provide proof of concept that technologies with embedded respiratory signal monitoring may be able to predict airway narrowing during exercise in individuals with asthma, particularly in the overweight or obese population.