RATIONALE

Idiopathic pulmonary fibrosis (IPF) is a progressive fibrosing interstitial lung disease associated with uncontrolled progressive fibrotic changes. The integration of reliable imaging-based biomarkers is needed to evaluate patients in the most holistic approach. 

METHODS

We used a retrospective cohort of 343 IPF patients from the Open-source Image Consortium (OSIC) dataset. Baseline CTs of all patients and 143 follow-up CTs were used for a longitudinal analysis. The AI-based lung quantification platform LungQ (Thirona, The Netherlands) was used to quantify ILD abnormalities (ILA%, expressed as the % of total lung volume) as well as the percentage of large (diameter ?2mm; AVX_LA) pulmonary artery volume using the artery-vein phenotyping analysis AVX.

RESULTS

We identified that automatized ILD quantification (ILA%) was correlated with FVC (%pred) and DLCO (%pred)(r=-0.47, p<0.001 and r=-0.54, p<0.001 respectively) at baseline and with lung decline over time (r=-0.48, p<0.001 and r=-0.43, p<0.001 respectively). AVX_LA changes were correlated with DLCO (%pred)(r=-0.36, p<0.001) at baseline and with DLCO modification over time (r=-0.32, p<0.001).

CONCLUSIONS

Longitudinal changes of LungQ ILA% and AVX_LA in IPF show to be related to functional changes over time. Combining imaging-based disease markers will help clinicians in the clinical follow-up as clinical decision support tool. Further clinical validation is needed in order to specify its potential clinical use.