Abstract Background The clinical course of Idiopathic pulmonary fibrosis(IPF)can be variable. Nevertheless, there is a paucity of robust data identifying reliable mortality predictors. Objectives To investigate the value of the high-resolution computed tomography (HRCT) patterns and deep learning-based quantitative CT in predicting and assessing prognosis in Idiopathic pulmonary fibrosis. Materials: This retrospective study included idiopathic pulmonary fibrosis patients with CT examinations and clinical data between January 2017 and December 2022. Methods The HRCT patterns were visually classified based on the 2018 criteria for usual interstitial pneumonia (UIP). CT-derived metrics including extent of lung fibrosis and lung volume was objectively quantified using deep learning algorithms. Cox proportional hazards and linear mixed-effects models, were employed to assess the relationships between CT-derived metrics and outcomes. Results Of the 347 IPF patients followed up for 8 years (median: 33.84 months), 204 patients died. The more extent of lung fibrosis in quartile groups was associated with worse disease severity and prognosis. In multivariable models, compared to the UIP pattern, the probable UIP pattern was found to be an independent protective factor against mortality (hazard ratio 0.28, 95% confidence interval (0.17 - 0.46, P 0.001;C statistic = 0.638 ). The fibrosis extent was strongly associated with survival independent of CT pattern (hazard ratio, 1.03; 95% confidence interval, 1.02 - 1.04; P 0.001; C statistic = 0.718). Linear mixed-effect modeling showed that fibrosis extent and CT lung volume of the annual change in decease IPF patients was significantly higher than those in survival patients. Longitudinally, decreasing lung volume (HR, 1.81,95% CI, 1.16 -2.82, P=0.002 ) and increasing fibrosis volume (HR, 2.57,95% CI, 1.61 - 4.11,P 0.001) were significantly associated with differences in survival. Conclusions Baseline and longitudinally increasing extent of lung fibrosis by quantitative CT is of greater prognostic importance for IPF in patients, independent of visually assessed CT pattern. This abstract is funded by: Jiangsu Provincial Key Research and Development (NO. BE2016611)
Zou et al. (Fri,) studied this question.