Unsupervised phenotyping of Severe Asthma Research Program participants using expanded lung data

• Published in: Journal of allergy and clinical immunology Vol. 133; no. 5; pp. 1280 - 1288
• Main Authors: Wu, Wei, Bleecker, Eugene, Moore, Wendy, Busse, William W., Castro, Mario, Chung, Kian Fan, Calhoun, William J., Erzurum, Serpil, Gaston, Benjamin, Israel, Elliot, Curran-Everett, Douglas, Wenzel, Sally E.
• Format: Journal Article
• Language: English
• Published: New York, NY Elsevier Inc 01.05.2014; Elsevier; Elsevier Limited
• Subjects: Adrenal Cortex Hormones - administration & dosage; Adult; Age of Onset; Algorithms; Allergy and Immunology; Asthma; Asthma - complications; Asthma - drug therapy; Asthma - metabolism; Asthma - pathology; Asthma - physiopathology; Asthma phenotyping; Biological and medical sciences; Bronchoalveolar Lavage; Classification; Cluster analysis; Eosinophilia - complications; Eosinophilia - drug therapy; Eosinophilia - metabolism; Eosinophilia - pathology; Eosinophilia - physiopathology; Female; Fundamental and applied biological sciences. Psychology; Fundamental immunology; Humans; Immunopathology; Information theory; Leukemia; Lung - metabolism; Lung - parasitology; Lung - pathology; Male; Medical sciences; Methods; Middle Aged; Nasal Polyps - complications; Nasal Polyps - drug therapy; Nasal Polyps - metabolism; Nasal Polyps - pathology; Nasal Polyps - physiopathology; Patients; Phenotype; Sarcoidosis. Granulomatous diseases of unproved etiology. Connective tissue diseases. Elastic tissue diseases. Vasculitis; Severity of Illness Index; Studies; supervised machine learning approaches; unsupervised approaches; variable analysis; Variables; Feno; INFOGAIN; variable analysis; Asthma phenotyping; unsupervised approaches; SARP; AQLQ; HC; HCU; BAL; supervised machine learning approaches; SA; Information gain; F eno; Fraction of exhaled nitric oxide; Asthma Quality of Life Questionnaire; Severe Asthma Research Program; Bronchoalveolar lavage; Health care use; Healthy control subject; Severe asthma; Immunopathology; Lung; Clinical research; Respiratory system; Learning; Phenotype; Acquisition process; Immunology; Research program; Expansion
• ISSN: 0091-6749; 1097-6825; 1097-6825
• DOI: 10.1016/j.jaci.2013.11.042

Previous studies have identified asthma phenotypes based on small numbers of clinical, physiologic, or inflammatory characteristics. However, no studies have used a wide range of variables using machine learning approaches. We sought to identify subphenotypes of asthma by using blood, bronchoscopic, exhaled nitric oxide, and clinical data from the Severe Asthma Research Program with unsupervised clustering and then characterize them by using supervised learning approaches. Unsupervised clustering approaches were applied to 112 clinical, physiologic, and inflammatory variables from 378 subjects. Variable selection and supervised learning techniques were used to select relevant and nonredundant variables and address their predictive values, as well as the predictive value of the full variable set. Ten variable clusters and 6 subject clusters were identified, which differed and overlapped with previous clusters. Patients with traditionally defined severe asthma were distributed through subject clusters 3 to 6. Cluster 4 identified patients with early-onset allergic asthma with low lung function and eosinophilic inflammation. Patients with later-onset, mostly severe asthma with nasal polyps and eosinophilia characterized cluster 5. Cluster 6 asthmatic patients manifested persistent inflammation in blood and bronchoalveolar lavage fluid and exacerbations despite high systemic corticosteroid use and side effects. Age of asthma onset, quality of life, symptoms, medications, and health care use were some of the 51 nonredundant variables distinguishing subject clusters. These 51 variables classified test cases with 88% accuracy compared with 93% accuracy with all 112 variables. The unsupervised machine learning approaches used here provide unique insights into disease, confirming other approaches while revealing novel additional phenotypes.