Unevenness of lung perfusion images and pulmonary diseases

The purpose of the study was to quantify the unevenness of perfusion distribution in the lungs in conjunction with underlying lung pathology. Twenty-one parameters as described previously were defined out of horizontal radioactive count profiles on perfusion lung image data taken in 64x64 matrixes. Principal component analysis has revealed that the 1st component or Z1 is represented by AREA, the area of the lung, and ANG, the slope of the mean count profile, Z2, by N, the number of peaks, Z3 and Z4, by YG and XG, the barycentric coordinates of count distribution, Z5, by MAC, the maximal count and Z6, by CSD, the degree of scatter in count from the peak count. How those parameters differ in each lung pathology has been determined from 657 lung perfusion image data. In pulmonary emphysema, the lung volumes are larger than those in normal subjects. The AREA and ANG were consequently larger in value and N was also significantly larger, indicating the increased regional alveolar pressure and the compressed or destroyed vascular beds. In diffuse panbronchiolitis (DPB), N was increased probably because the distal airways were either narrowed or obstructed by inflammatory processes inducing regional alveolar hypoxia and/or alveolar hyperinflation. In fibrosis, both AREA and N were significantly smaller. In congestive heart failure with postcapillary pulmonary hypertension, YG was smaller probably because of 'reversal of perfusion'. In pulmonary sarcoidosis, an increase in YG was the only abnormality. (author)