Artificial intelligence-powered interpretation of lung function in interstitial lung diseases

Abstract

Artificial intelligence (AI) is a branch of computer science developed in the 1950s to imitate the human ability for solving complex problems. Currently, it is widely applied in various fields of medicine for diagnostic support, assistance with medical practice and drug discovery. After entering respiratory medicine two decades ago, AI has been used in this field to support in making diagnoses and predicting outcomes based on clinical data, imaging, pathology and pulmonary function tests (PFTs). Imaging is the field where AI has made the greatest progress in respiratory medicine.1–3 The current application areas of AI in interstitial lung disease (ILD), a complex group of disorders in respiratory medicine, are not few: drug discovery, risk assessment, decision-making for treatment, identifying cohort from databases, epidemiological analysis, medical assistance and support in interpreting imaging and other diagnostic tests.1 In the field of ILD, the increasing adoption of AI techniques owing to the complexities in ILD diagnosis and management has led to research primarily on AI-supported evaluation of imaging but also gene expression, imaging and genomic data, proteomic data and plasma biomarkers, volatile organic compounds and PFTs.4 Diagnosing ILD is challenging; it is often misdiagnosed initially or diagnosed late in the disease course as PFTs are only minimally affected at the onset. …