Observational Prediction Model for Clinical Outcomes in Idiopathic Pulmonary Fibrosis: a Multicentre, ML-driven Study (OPEN-IPF)
Observational Prediction Model for Clinical Outcomes in Idiopathic Pulmonary Fibrosis: a Multicentre, ML-driven Study (OPEN-IPF)
Idiopathic pulmonary fibrosis (IPF) is a progressive fibrotic lung disease with marked inter-individual heterogeneity in trajectories and outcomes. Despite antifibrotic therapies, reliable risk stratification in routine practice remains suboptimal. OPEN-IPF is a multicentre retrospective observational cohort study designed to build a harmonised real-world dataset across Italian IPF referral centres to enable the development and external validation of machine-learning (ML) models predicting clinically relevant outcomes.
OPEN-IPF addresses the current limitation of AI/ML research in IPF-namely, the lack of large multicentre real-world datasets with harmonised variables and robust external validation. The study will retrospectively include adult patients with IPF followed in routine practice in participating Italian referral centres from 1 January 2015 to 31 December 2025 (data lock). No study-specific procedures will be performed. De-identified/pseudonymised data will be collected using a common data model, including demographics, smoking history, comorbidities, pulmonary function (FVC, DLCO), oxygen requirement, 6-minute walk test (where available), antifibrotic treatment exposure, HRCT features routinely reported, basic laboratory parameters, and clinical outcomes. The primary modelling targets are disease progression, acute exacerbations of IPF (AE-IPF), and real-world response to antifibrotic treatment. Model development will be performed using multicentre data with explicit external validation across centres
Inclusion Criteria:
Exclusion Criteria:
rtonelli@unimore.it0039059425934
stefania.cerri@unimore.it00390594225335