Modeling Heterogeneity and Critical Care Support for Idiopathic Pulmonary Fibrosis

Introduction: Idiopathic Pulmonary Fibrosis (IPF) is a non-uniform pathology with variation in risks, timing, treatments, and outcomes, making it challenging for healthcare systems to optimize care. Precision medicine for IPF has identified genetic risk (MUC5B, TERT, and RTEL1), environmental factors, and socioeconomic factors (job, housing conditions, access to healthcare, and geographical location). The current models do not capture the complex heterogeneity of IPF.


Methods: We integrated the UK Biobank Open Targets Platform gene networks of IPF relative to other pulmonary conditions (sarcoidosis, lung cancer, emphysema, bronchiectasis, bronchitis, pneumonia, pulmonary hypertension, asthma, COPD, cystic fibrosis, lung disease, and sleep apnea), the CZI single cell atlas of lung, AI-enabled gene extractions from literature, and GWAS colocalization dynamics of IPF loci relative to all other traits. We also generated 100 shared patient, caregiver, and clinician AI-simulated journeys for pre-IPF, diagnosis, treatment, and outcomes.


Results: IPF gene signatures cluster between sarcoidosis and emphysema with enrichment of biological pathways including phospholipase activity (FDR 2.37e-07), vascular endothelial growth factor signaling (1.78e-06), transmembrane receptor protein tyrosine kinase activity (6.52e-11), growth factor binding (2.46e-10), Rap1 signaling (7.90e-08), and Hippo-Merlin signaling dysregulation (1.60e-16). GWAS colocalization identified convergent biological pathways including lung function, plateletcrit, smoking initiation, blood metabolite ratios, pulse pressure, telomere length, COVID-19, and aspartate aminotransferase. The 100 journeys identified disease-modifying factors (dust, mold, chemical exposure, fumes, smoke, solvents) with early intervention suggestions (PPE, spirometry, oxygen). Convergence occurred in education, mechanical, healthcare, maritime, and arts/crafts sectors.


Conclusions: We aim to integrate the new IPF and pulmonary tools into a future AI-enabled ecosystem to merge patient, caregiver, and clinician journeys in IPF with our molecular insights. These future tools hold the promise of guiding omic-generated signature profiles of patients into personalized treatment plans.