Usefulness of chest perfusion computed tomography in the diagnosis of diabetic pulmonary microangiopathy

This paper presents the usefulness of perfusion computed tomography (pCT) in the diagnosis of diabetic pulmonary microangiopathy. Our previous works have shown that perfusion parameters are useful in the diagnosis of diabetic pulmonary microangiopathy. We are looking for such measurements and perfusion parameters that provide the most accurate diagnosis. Two types of comparison were made based on the results of clinical trials: non-diabetic vs. diabetic and diabetes without microangiopathy vs. diabetes with microangiopathy. Our studies have shown that PS (permeability surface) is only perfusion parameter statistically significant. In certain regions of interest logistic regression as a classifier produces very good results in diagnosing lung microangiopathy: sensitivity Sens = 89% and excellent specificity Spec = 100%. The results were obtained on the base of measurements taken from 23 subjects. These results were compared with results reported in the literature and based on diffusion capacity and spirometry measurements and modeling. None of the previous results was as good as those obtained using the PS and logistic regression for binary classification.