Identification of subjects with abnormal lung function using partial flow-volume curves generated during a voluntary cough.
Frazer-DG; Mahmoud-AM; Abaza-AA; Goldsmith-WT; Day-JB; Reynolds-JS; Barkley-JB; Petsonk-EL
Am J Respir Crit Care Med 2010 May; 181(Meeting Abstracts):A5557
Rationale: Measurement of partial expiratory flow-volume (PEFV) curves has become an important technique in diagnosing lung disease, particularly in children and in the elderly. In most instances the generation of this curve is effort independent and repeatable. The objective of this study was to determine if the characteristics of a partial flow-volume curve generated during a cough (CPEFV) could be used to diagnose subjects with abnormal lung function. Methods: Cough airflow curves were acquired and saved using a system developed in our laboratory to measure cough sound and airflow characteristics. The CPEFV relationships of twenty-seven male and twenty-one female subjects with obstructive lung disease were compared with those of twenty-seven males and twenty-five females with normal lung function. A variety of features was extracted from the airflow signal in both the time and frequency domains. Feature selection techniques were used to eliminate irrelevant parameters and improve overall system performance. A classifier was developed based on the reconstruction error of the principle components of the most relevant feature set. The classifier was evaluated by testing its ability to distinguish between the coughs from test subjects with normal or abnormal lung function. This was accomplished by comparing the classifier's decisions with the clinical diagnoses of a pulmonary physician. Results: Findings indicated that it was possible to classify patients with lung disease with an overall accuracy of 98% for males and 97% for females. Conclusions: This study demonstrates that CPEFV curves can be used to accurately identify test subjects with abnormal pulmonary function.
Age-groups; Diseases; Laboratory-techniques; Lung; Lung-function; Pulmonary-clearance; Pulmonary-disorders; Pulmonary-function; Pulmonary-function-tests; Pulmonary-system; Pulmonary-system-disorders; Respiratory-system-disorders; Sex-factors; Statistical-analysis; Surveillance;
Author Keywords: Voluntary Cough; Cough Airflow; Partial Flow-Volume Curves; Pattern Classification
Healthcare and Social Assistance; Services: Public Safety
American Journal of Respiratory and Critical Care Medicine