Quantitative diagnosis of connective tissue disease-associated interstitial pneumonia using thoracic computed tomography images

Patients with polymyositis (PM) or dermatomyositis (DM) frequently show interstitial pneumonia (IP), which is sometimes rapidly progressive or resistant to treatment, thereby significantly affecting the prognosis. The diagnosis and response evaluation of IP are commonly performed qualitatively based on imaging findings, which may cause disagreement among rheumatologists in the evaluation of early lesions and atypical interstitial changes. To determine whether IP could be diagnosed in a quantitative manner during the early stage of PM/DM using a workstation that allows quantitative image processing. Thoracic computed tomography (CT) images of 20 PM/DM patients were reconstructed into a three-dimensional (3D) image using an image processing workstation. The CT values of the constituent voxels were arranged in a histogram of -1000 to +1000 Hounsfield units (HU). The most frequent lung field density was -900 to -801 HU, and relative size was as follows: IP (+) group 0.45 and IP (-) group 0.53. Between -1000 and -701 HU, relative size was not significantly different between the IP (+) group and IP (-) group. Between -700 and -1 HU, the relative size of the lung field was significantly larger in the IP (+) than in the IP (-) group, demonstrating its IP-diagnosing ability. Particularly, within the range from -700 to -301 HU, the macroscopically-assessed ground glass opacity was consistent with the CT value, which, in turn, was closely correlated with KL-6, the pre-existing marker for IP diagnosis. The results of this study may lead to the establishment of quantitative methods of evaluating IP and possible elucidation of the pathogenesis of IP.