Vocal biomarkers correlate with FEV1 variations during methacholine challenge

• Published in: Clinical and translational allergy Vol. 15; no. 4; pp. e70055 - n/a
• Main Authors: Paoletti, Giovanni, Costanzo, Giovanni, Merigo, Morena, Puggioni, Francesca, Ferri, Sebastian, Messina, Maria Rita, Cordella, Fulvio, Ranieri, Giuseppe, Arienzo, Arianna, Savevski, Victor, Canonica, Giorgio Walter, Brito Martins, Ayana, Heffler, Enrico
• Format: Journal Article
• Language: English
• Published: England John Wiley & Sons, Inc 01.04.2025; John Wiley and Sons Inc; Wiley
• Subjects: artificial intelligence; Asthma; Biomarkers; digital health; Machine learning; mobile health technology; Original; vocal biomarkers; asthma; digital health; vocal biomarkers; mobile health technology; artificial intelligence
• ISSN: 2045-7022; 2045-7022
• DOI: 10.1002/clt2.70055

Background Mobile health applications are increasingly valued for their role in asthma management and the opportunity for large dataset collection. Our study aimed to investigate the feasibility of applying signal‐processing and machine‐learning technologies to detect alterations in the lower airway caliber and develop a machine‐learning algorithm to identify changes in vocal biomarkers and detect bronchoconstriction in patients with airway hyperreactivity. Methods This is an explorative observational prospective longitudinal study focused on vocal biomarkers and their association with bronchial constriction and respiratory function. Non‐smoker adults with clinical suspicion of asthma were consecutively enrolled from May 2023 to September 2023. At each step of a Methacholine Challenge Test (MCT) performed on these patients, the respiratory sounds were recorded via a smartphone through an app specifically developed. Several biomarkers were extracted and their relationship with the change in Forced Expiratory Volume in the first second (FEV1) was measured. Results Forty‐two subjects were enrolled. The highest correlation with FEV1 came from exhalation vocal events. No single feature exhibited robust behavior across different subjects, while each subject showed “personal” highly correlated features. All values were strongly statistically significant irrespectively of the result of MCT. Conclusion The app’s algorithm is sensitive in correlating specific vocal biomarkers to FEV1 variations during MCT. This feature may assist physicians in diagnosing asthma and its exacerbation and in assessing therapy response and adherence. The socio‐economic implications might be significant, and the simplicity of use makes it an ideal tool for research.