Medium-long term prognosis prediction for idiopathic pulmonary fibrosis patients based on quantitative analysis of fibrotic lung volume.

PURPOSE

To investigate the prognostic value of quantitative analysis of CT among patients with idiopathic pulmonary fibrosis (IPF) by quantifying the fibrosis extent and to attempt to provide precise medium-long term prognostic predictions for individual patients.

METHODS

This was a retrospective cohort study that included 95 IPF patients in Zhongshan Hospital, Fudan University. 64 patients firstly diagnosed with IPF from 2009 to 2015 was included as the derivation cohort. Information regarding sex, age, the Gender-Age-Physiology (GAP) index, high-resolution computed tomography (HRCT) images, survival status, and pulmonary function parameters including forced vital capacity (FVC), FVC percent predicted (FVC%pred), diffusing capacity of carbon monoxide (DLCO), DLCO percent predicted (DLCO%pred), carbon monoxide transfer coefficient (KCO), KCO percent predicted (KCO%pred) were collected. 31 patients were included in the validation cohort. The Synapse 3D software was used to quantify the fibrotic lung volume (FLV) and total lung volume (TLV). The ratio of FLV to TLV was calculated and labeled CT, reflecting the extent of fibrosis. All the physiological variants and CT were analyzed for the dimension of survival through both univariate analysis and multivariate analysis. Formulas for predicting the probability of death based on the baseline CT were calculated by logistic regression, and validated by the validation cohort.

RESULTS

The univariate analysis indicated that CT along with DLCO%pred, KCO%pred and GAP index were significantly correlated with survival. However, only CT was meaningful in the multivariate analysis for prognostic prediction (HR 1.114, 95% CI 1.047-1.184, P = 0.0006), and the best cutoff was 11%, based on receiver operating characteristic (ROC) curve analysis. The survival times for the CT ≤ 11% and CT > 11% groups were significantly different. Given the CT data, the death probability of a patient at 1 year, 3 years and 5 years could be calculated by using a particular formula. The formulas were tested by the validation cohort, showed high sensitivity (88.2%), specificity (92.8%) and accuracy (90.3%).

CONCLUSION

Quantitative volume analysis of CT might be useful for evaluating the extent of fibrosis in the lung. The CT could be a valuable biomarker for precisely predicting the medium-long term prognosis of individual patients with IPF.