Quantitative computed tomography improves prediction of survival across ILD subtypes

Abstract

Background Quantitative computed tomography (CT) methods are recognized as valuable tools for risk stratification in fibrotic interstitial lung disease (ILD). Data-driven textural analysis (DTA) is one such approach that correlates with worse lung function and shorter survival in idiopathic pulmonary fibrosis (IPF) and other fibrotic ILDs. Broader clinical adoption of quantitative CT would be supported by a clearer understanding of DTA's prognostic performance across different ILD entities. Methods DTA was measured from CT images in the Pulmonary Fibrosis Foundation (PFF) Patient Registry (discovery) and University of Chicago ILD research cohort (replication). Cox regression models were used to examine associations of DTA-fibrosis score with 3-year transplant-free survival (TFS). Harrell's C-index was used to assess performance of DTA-fibrosis score at predicting 3-year TFS in IPF versus non-IPF cases compared with clinical risk factors. Results There were 668 (IPF=435, non-IPF=233) and 225 (IPF=102, non-IPF=123) patients with DTA assessments in the discovery and replication cohorts, respectively. In the discovery cohort, a 1% increment in DTA-fibrosis score was associated with a hazard ratio for TFS of 1.03 (95%CI 1.02–1.04) in IPF cases and 1.03 (95% CI 1.01–1.06) in non-IPF cases. The DTA-fibrosis score demonstrated better performance at predicting 3-year TFS than traditional clinical risk factors in IPF cases (0.69 versus 0.62) and non-IPF cases (0.80 versus 0.65). These findings were similarly replicated in the replication cohort. Conclusion Higher DTA-fibrosis scores improved prognostication in IPF and non-IPF fibrotic ILDs, underscoring the importance of CT-based fibrosis extent in clinical risk assessment of various forms of ILD.