Radiomics for Differentiating Autoimmune-Related Interstitial Lung Disease from Idiopathic Pulmonary Fibrosis: A Narrative Review

Abstract

Interstitial lung disease (ILD) represents a wide variety of lung diseases, commonly resulting in irreversible changes with worsening quality of life and a high index of casualties per year. Idiopathic pulmonary fibrosis (IPF) shows the highest prevalence and the worst survival rates among ILD subtypes. In contrast, autoimmune-related ILD often follows a more indolent course and is frequently secondary to an identifiable underlying condition that can potentially be managed. Early differentiation between them can optimize adequate treatment with favorable disease progression and patient outcomes. While high-resolution computed tomography (HRCT) and pathology studies are the preferred methods for diagnosis and distinguishing between them, the radiological differentiation of IPF and autoimmune-related ILD disease is often subtle, necessitating the introduction of a revolutionary tool with more accurate results. Radiomics utilizes the imaging characteristics of lung tissue for early recognition and differentiation of ILD subtypes, predicting disease severity, and assessing treatment response. Key radiomics features differentiating IPF from autoimmune-related ILDs include texture features, indices of pixel density heterogeneity, volumetric parameters, and reticulation volume. While artificial intelligence-based methods have expanded the analytical capabilities of radiomics, significant challenges persist regarding robustness, interpretability, and generalizability. This article reviews the potential role and application of radiomics as an emerging tool in the recognition and differentiation of the 2 most frequent forms of ILD.