Semiquantitative Measurement of Murine Bleomycin-Induced Lung Fibrosis in In Vivo and Postmortem Conditions Using Microcomputed Tomography: Correlation With Pathologic Scores???Initial Results

Abstract

To evaluate whether the semiquantification of lung inflammation and fibrosis in murine bleomycin-induced lung fibrosis using micro-computed tomography (micro-CT) in in vivo and postmortem conditions is feasible, and to correlate micro-CT and pathologic scores. Bleomycin-induced lung fibrosis was created by intratracheally instilling 3 mg/kg of bleomycin into C57BL/6 mice. Mice were allocated randomly to 2-week, 4-week, and 8-week follow-up groups. In each group, in vivo and follow-up postmortem micro-CT were performed using a voxel size of 35 x 35 x 35 microm. Ground-glass opacity (GGO), consolidation, parenchymal lines, honeycombing, and peripheral bronchial dilatation were scored on micro-CT images in a semiquantitative fashion, whereas inflammation and fibrosis were scored histopathologically. The confidence levels of micro-CT findings were also scored. Correlations between micro-CT and pathologic findings were examined using Spearman rank correlation analysis, and differences between CT scores and confidence levels for in vivo and postmortem micro-CT were subjected to Wilcoxon signed rank testing. Agreements between in vivo and postmortem micro-CT scores were tested using weighted kappa statistics. Consolidation in vivo (r = 0.46) and at postmortem (r = 0.39) and GGO in vivo (r = 0.31) by micro-CT showed fair to moderate correlation with pathologic inflammation scores (P < 0.001). By in vivo and postmortem micro-CT, parenchymal lines (r = 0.72 vs. 0.83) showed good to excellent and peripheral bronchial dilatation (r = 0.47 vs. 0.68) showed moderate to good correlation with pathologic fibrosis scores (P < 0.001). For GGO, consolidation, peripheral bronchial dilatation, and parenchymal lines, fair to moderate agreement was obtained between in vivo and postmortem micro-CT. However, confidence levels for peripheral bronchial dilatation, parenchymal lines, and honeycombing were significantly higher by postmortem micro-CT (P < 0.001). Micro-CT scores and pathologic scores were found to be well correlated by in vivo and postmortem micro-CT. Although agreements between in vivo and postmortem micro-CT were significant, the confidence levels for fibrosis-related CT findings were significantly higher by postmortem micro-CT.