Quantitative CT powered by machine learning could help doctors better predict death and disease severity in patients with idiopathic pulmonary fibrosis, international researchers say.

Results from their registry-based study suggest that imaging using data-driven texture analysis (DTA) would also enable earlier initiation of anti-fibrotic therapy and better timing of lung transplant evaluation before “irreversible progression has occurred”.

Unlike current prediction models, which do not help to stratify patients for disease progression in idiopathic pulmonary fibrosis (IPF), the technique provides “clinically meaningful” morphological data, say the Australian and US authors.

The team, including respiratory specialists from Brisbane, Sydney and Melbourne, used CT scans from some 390 adults (median age 70) enrolled in the Australian IPF Registry to assess the prognostic value of lung fibrosis extent quantified by DTA.