One-Year Change in Quantitative Computed Tomography Is Associated with Meaningful Outcomes in Fibrotic Lung Disease

Abstract

Rationale: Whether change in fibrosis on high-resolution computed tomography is associated with near- and longer-term outcomes in patients with fibrotic interstitial lung disease (fILD) remains unclear. Objectives: We evaluated the association between 1-year change in quantitative fibrosis scores (data-driven textural analysis [DTA]) and subsequent FVC and survival in patients with fILD. Methods: The primary cohort included patients with fILD evaluated from 2017 to 2020 with baseline and 1-year follow-up high-resolution computed tomography and FVC. Associations between DTA change and subsequent FVC were assessed using linear mixed models. Transplant-free survival was assessed using Cox proportional hazards models. The Pulmonary Fibrosis Foundation Patient Registry served as the validation cohort. Measurements and Main Results: The primary cohort included 407 patients (median [interquartile range] age, 70.5 [64.8, 75.9] yr; 214 male). One-year increase in DTA was associated with subsequent FVC decline and transplant-free survival. The largest effect on FVC was observed in patients with low baseline DTA scores, in whom a 5% increase in DTA over 1 year was associated with a change in FVC of -91 ml/yr (95% confidence interval [CI], -117, -65 vs. stable DTA, -49 ml/yr; 95% CI, -69, -29; P = 0.0002). The hazard ratio for transplant-free survival for a 5% increase in DTA over 1 year was 1.45 (95% CI, 1.25, 1.68). The findings were confirmed in the validation cohort. Conclusions: One-year change in DTA score is associated with future disease trajectory and transplant-free survival in patients with fILD. DTA could be a useful trial endpoint, cohort enrichment tool, and metric to incorporate into clinical care.

Keywords: computed tomography; deep learning; interstitial lung disease; pulmonary fibrosis.