An Integrated Software Package to Classify Human Respiratory Diseases

摘要

The Impulse Oscillometry System (IOS) is a commercially available instrument for performing forced oscillation, to diagnose and monitor human respiratory diseases. IOS is a patient-friendly and objective pulmonary function testing device that measures pressure/flow at the mouth to calculate the spectral components of respiratory system impedance (Z). Our research group has demonstrated that the augmented RIC (aRIC) model, a modification of our extended RIC (eRIC) model, provides physiologically meaningful estimates of small and large airway resistive, inertive, and elastic components. Due to the high dimensionality of the IOS data and the wide range of model parameters that may distinguish between health and disease states, neuro-fuzzy approaches have been proven to be very helpful in classification of pulmonary diseases. This paper introduces an integrated software package that classifies asthma, small airways impairment (SAT), mild SAI, and normal airway function based on aRIC model parameters and a high performance neuro-fuzzy clustering algorithm; and once fully implemented will present classification results in a clinician-friendly 3-D graphical user interface. IOS data from 112 children asthmatic and non-asthmatic children 7-15 years of age were used for this study.