B23-11 Extent of Lung Fibrosis by Deep Learning-Based Quantitative CT Is the Strongest Predictor of Mortality in Idiopathic Pulmonary Fibrosis: A Retrospective, Multicenter Study
Retrospective study investigates lung fibrosis extent as a mortality predictor in idiopathic pulmonary fibrosis, highlighting its significance.
Key Points
The aim is to assess how deep learning-based quantitative CT and HRCT patterns predict mortality in idiopathic pulmonary fibrosis.
Retrospective analysis of 347 idiopathic pulmonary fibrosis patients with CT exams from January 2017 to December 2022.
Quantified lung fibrosis and volume using deep learning algorithms alongside visual classification of HRCT patterns.
Analyzed relationships between CT metrics and outcomes using Cox proportional hazards and linear mixed-effects models.
204 out of 347 patients died, with greater extent of lung fibrosis linked to worse prognosis.
In multivariable models, the probable UIP pattern was independently protective against mortality (HR 0.28, 95% CI 0.17 - 0.46, P < 0.001).
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 associated with survival differences.