Measurement of mechanical input impedance of the respiratory system, Zrs, is noninvasive and requires minimal subject cooporation. Hence, a pulmonary function test (PFT) based on Zrs data would be highly desirable, especially in children. Studies have attempted to estimate specific mechanical properties of the respiration system by fitting inverse models to Zrs data. Most efforts were unsuccessful, even in healthy subjects, because only 2-32 Hz data were used. Recent work shows that such data allows reliable distinction of only total respiratory resistance, inertance, and compliance. To obtain statistically reliable estimates of parameters in more complicated models requires data over a much wider frequency range. With 4-200 Hz data reliable distinctions between airway and tissue properties have recently been obtained in a model applied to healthy animals. For adult humans the same model results in statistically reliable, but physiologically inappropriate parameters. Also, use of Zrs data in disease diagnosis, will require models with parallel inhomogeneities and hence, data betwen 0.2-4 Hz as well as 4-200 Hz. The overall goals of this proposal are to: 1) evaluate and compensate for a model successful in animals but unsuccessful in adult humans; 2) refine inverse model fitting for animals and humans so that parallel and serial inhomogeneities can be incorporated thus allowing application to healthy children and disease patients; and 3) assess, especially in children; the ability of Zrs measurements from 0.2-200 Hz to detect lung disease earlier, to resolve more specific location of abnormalities then current PFTs, and to quantify disease extent (mild to severe). In evaluating previous modeling anomolies, we must determine how the upper airways and cheeks as well as airway acoustic properties can distort Zrs data. We first will apply a technique which minimize upper airway influence on Zrs on humans. Then, using dogs we will determine the importance of tracheal compliance. Data generated with anatomically consistent forward models and modern sensitivity analysis techniques will be used to evaluate candidate inverse models with and without airway acoustic properties, and with both parallel and serial inhomogeneities. Finally, we will acquire Z data from 0.2 -200 Hz in healthy and sick children ages 42 months to 18 years. The statistical reliability and physiological appropriateness of parameters in inverse models applied to such data will be determined.
Showing the most recent 10 out of 35 publications
