Summary

Background: Lung microangiopathy is a little known negative influence of diabetes mellitus on the functioning of the lungs. In current medical practice lung microangiopathy is diagnosed by comparing two measurements of lung diffusing capacity – once with the subject standing and once with the subject lying down. The necessity to take two measurements is inconvenient. Objectives: The aim of this study is to design a supportive method for diagnosing lung microangiopathy. This will be based on routinely performed pulmonary measurements as well as on investigation of process modelling and data processing. Methods: A model of the diffusion of oxygen from the alveoli to the blood has been described with a set of differential equations. The idea of the proposed model is based on the physiological analysis of the oxygen flow (caused by a concentration gradient) and on general knowledge regarding the kinetics of associating oxygen with haemoglobin. The model parameters are estimated using diffusing capacity and alveolar volume measurements – routinely performed in pulmonary tests. Results: The model parameter estimates proved good candidates for the binary classification of the presence or absence of microangiopathy. The proposed classification procedure, based on parameter values and established diagnostic thresholds, gives sensitivity Sens = 79.34% and specificity Spec = 87.08%. The results of classification with the use of diffusing capacity measurement are worse: Sens = 62.12% and Spec = 79.89%. Conclusions: The proposed classification procedure is based on the model parameters. These have proved to be sensitive indicators of lung microangiopathy. Close to 80% of microangiopathy cases have been classified as such. Less than 20% were false alarms. The oxygen pathway model allows for simulations. Blood saturation and oxygen partial pressure have been simulated for the organism’s various needs for oxygen, both for the normal and the impaired alveoli-capillary barrier.