Standardized quantification of pulmonary fibrosis in histological samples.

ABSTRACT

The Ashcroft scale for the evaluation of bleomycin-induced lung fibrosis is the analysis of stained histological samples by visual assessment. Based on the knowledge that this procedure is not standardized in animals and results are highly variable, we hypothesized that modification of this method may improve quantification of lung fibrosis in small animals. To prove our hypothesis, we evaluated pulmonary fibrosis in Lewis rats induced by a single intratracheal injection of 0.3 mg/kg body weight bleomycin (n = 13) compared with the same amount of saline in a control group (n = 4). We modified the Ashcroft scale by precisely defining the assignment of grades from 0 to 8 for the increasing extent of fibrosis in lung histological samples. Thirty-two observers were randomly assigned to evaluate 108 photographs of slides using either the Ashcroft scale or the modified scale. Consistent with our hypothesis, there was a significant reduction in the variability of standard deviations with the modified scale compared with the Ashcroft scale (mean of variability 0.25 versus 0.62, P < 0.0001). Applying the kappa index, the Ashcroft scale showed only a fair to moderate agreement (0.23-0.59) between the observers and a low intra-observer agreement (0.51-0.74) in contrast to the modified scale, which demonstrated a moderate to good agreement between the observers (0.65-0.93, P < 0.0001) and a high intra-observer agreement (0.87-0.91, P < 0.05). To test the modified scale in vivo, we compared both scales with the results of computed tomography (CT) of the lungs obtained from the same mice. In agreement, the modified scale demonstrated a better correlation to CT scans (R = 0.58) compared with the Ashcroft scale (R = 0.33). In summary, quantification of lung fibrosis in histological lung sections using the modified scale is reliable and reproducible.