Prognostic assessment of acute exacerbation of interstitial lung disease: From conventional models to artificial intelligence-driven models

Acute exacerbation of interstitial lung disease (AE-ILD) is a life-threatening clinical event associated with poor prognosis. Accurate outcome prediction is important for treatment planning and resource allocation. This review evaluates evidence on prognostic risk factors and clinical scoring systems and discusses the application of artificial intelligence (AI) into the key stages of model development for AE-ILD prognostication. Conventional prognostic approaches enable rapid risk stratification but have limited capacity to integrate heterogeneous data and account for clinical complexity. AI-driven approaches have been preliminarily applied to clinical and imaging data for prognostic assessment. However, direct AE-ILD-specific evidence is limited, and the incremental prognostic value of existing AI models over conventional approaches remains uncertain. Multimodal integration, independent external validation, and model interpretability remain major research gaps. Addressing these gaps through representative multicentre studies may enable AI to reinforce conventional prognostic assessment and promote more individualised risk stratification in AE-ILD.