Assessment of survival in patients with idiopathic pulmonary fibrosis (IPF) using quantitative HRCT indexes

Background: the predictive role of CT-histogram indexes in assessing survival was poorly studied in IPF. Objective: To investigate the accuracy of CT-histogram derived indexes in the assessment of survival in IPF patients (pts). Methods: 42 IPF pts with a follow-up time of 3 years were retrospectively collected; HRCT and pulmonary function tests (FVC and DLCO) performed at diagnosis were analyzed; the extension of fibrosis was quantified on HRCT using Kurtosis, Skewness, Mean Lung Density (MLD), High Attenuation Areas (HAA%) and Fibrotic Areas (FA%). Univariate Cox-regression analysis assessed Hazard ratios for each variable. A Multivariate model including Skewness, FVC, DLCO and age tested their prognostic value in assessing survival. Through ROC analysis, threshold values demonstrating the best sensitivity and specificity in predicting mortality were identified and used as cut-off points to graph Kaplan-Meier curves specific for the CT-indexes. Results: Kurtosis, Skewness, MLD, HAA% and FA% were good predictors of mortality (HR 0.44, 0.74, 1.01, 1.12, 1.06; p=0.03, p=0.01, p=0.02, p=0.02 and p=0.017 respectively). Skewness, which demonstrated the lowest AIC value (55.52), was considered the best analyzed CT-index in predicting mortality. Significant survival differences considering proposed cut-off points were also demonstrated according to Kurtosis (p=0.02), Skewness (p=0.005), MLD (p=0.003), HAA% (p=0.009) and FA% (p=0.02) – obtained from quantitative HRCT analysis at the diagnosis time. Conclusions: CT-histogram derived indexes may provide an accurate estimation of survival in a population of IPF patients since the time of diagnosis.