Prediction of therapeutic effect of antifibrotic therapy for idiopathic pulmonary fibrosis using quantitative CT...

To determine whether CT texture analysis can predict therapeutic response after antifibrotic therapy for idiopathic pulmonary fibrosis (IPF).

Forty-one IPF patients treated with antifibrotic drugs for more than 1 year were enrolled. Ratios of 7 lesion types (air space; consolidation; consolidation with bronchiectasis, i.e., fibrosis; ground-glass opacity; honeycombing; reticular opacity; and traction bronchiectasis), normal lungs, emphysema, and vascular-related structures were compared between responders and non-responders. Clinical and spirometric parameters were also compared. Univariate and multivariate logistic regression analyses were used to compare all parameters. Predictive capability was evaluated using a receiver operating characteristic (ROC) curve. Kaplan-Meier curves of overall survival were compared between the high ratio group and the low ratio group of the significant parameter. In a validation study, 25 IPF patients were examined.

Fibrosis and traction bronchiectasis ratios showed significance in the univariate logistic regression analysis. Fibrosis ratio alone showed significance and revealed the best odds ratio (2.09) in the multivariate logistic regression analysis. Diagnostic capability of fibrosis ratio in upper lobes (AUC, 0.85) was superior to that of entire lung or lower lobes. Median overall survival for a high fibrosis ratio in upper lobes group was 41.6 months, and that for a low fibrosis ratio group was 69.2 months (P = 0.026). The validation study also yielded high diagnostic capability of F ratio in upper lobes (AUC, 0.83).

Fibrosis ratio in upper lobes seems to be a promising tool for predicting therapeutic response in patients with IPF treated with antifibrotic drugs.