CT quantification of interstitial fibrosis in patients with asbestosis: a comparison of two methods.

Abstract

The likelihood of interstitial lung disease being detected on high-resolution CT scans and having functional significance is often related to the severity of the disease. The extent and severity of the abnormalities seen on high-resolution CT are usually assessed subjectively. This study was undertaken to investigate whether a subjective semiquantitative scoring method or a method using a cumulation of the different high-resolution CT features of asbestosis were comparable in suggesting asbestosis in a group of patients with histopathologic confirmation of disease. A secondary objective was to compare the results of these two high-resolution CT methods with chest radiographs in the same population. This study group consisted of 24 patients and six lungs obtained at autopsy. Histopathologic asbestosis was present in 25 of the 30 patients or lungs. The patients or lungs were imaged using selected high-resolution CT scans. The high-resolution CT scans were assessed in two ways. One used a subjective semiquantitative extent and severity score consisting of four levels of severity, while the other was a cumulative score adding the different types of high-resolution abnormalities in asbestosis. The commonest high-resolution CT abnormalities in the cases with confirmed asbestosis were interstitial lines (84%), parenchymal bands (76%), and architectural distortion of secondary pulmonary lobules (56%). Subpleural lines and honeycombing were less frequent. The histopathologic severity of asbestosis was independently graded on a four-point scale. Chest radiographs, when available, were classified according to the International Labor Organization (ILO) classification of pneumoconioses. With the subjective semiquantitative high-resolution CT severity score, asbestosis was suggested in 16 (64%) instances, all with disease. With the cumulative method, any one type of abnormality was present in 88% of cases with asbestosis, two types in 78%, and three in 56%. However, to include only cases with asbestosis, three different abnormalities had to be present. The high-resolution CT scans were normal or near normal in five instances of asbestosis. Chest radiographs using the ILO classification predicted asbestosis with a lesser frequency than high-resolution CT in this selected population. We conclude that a subjective semiquantitative grading system of the extent and severity of asbestosis and a method using a cumulative addition of the different findings in asbestosis give similar results in suggesting the presence of disease. Thus, for the high-resolution CT detection of asbestosis, a combination of the cumulative number of different findings and an assessment of the extent and severity of the abnormalities could be complimentary. We also conclude that asbestosis can be present histopathologically with a normal or near normal high-resolution CT scan.