Quantitative metrics on computed tomography can predict the rapid progression of idiopathic inflammatory myopathies-related interstitial lung disease

<bold>Purpose:</bold> This study aims to utilize quantitative CT based on deep learning to predict the rapid progression of interstitial lung disease (RP-ILD) in patients with Anti-synthase syndrome (ASS) and anti-MDA5-positive dermatomyositis (MDA-5). <bold>Material and Methods:</bold> 511 interstitial lung disease (ILD) consisting of 356 inpatients with ASS and 155 cases with MDA-5 were included in a retrospective study from January 2016 to December 2021. HRCT (High-resolution CT) scans were analyzed using a deep-learning algorithm to quantify GGO (ground glass opacities) and consolidation. Logistic regression and Spearman's correlation were used to evaluate the relationship between these scores and RP-ILD status, which was defined as a decline in symptoms, imaging, or physiology (pulmonary function and blood gas) within 3 months. <bold>Results:</bold> In 356 patients with ASS-ILD, 124 had RP-ILD, while in 155 MDA5-ILD patients, 69 had RP-ILD. GGO and consolidation on HRCT before treatment were associated with RP-ILD in both diseases. Quantitative GGO and consolidation were negatively correlated with pulmonary function tests and blood gas variables, specifically vital capacity and carbon monoxide diffusing capacity. A dose-response relationship was found between the percentage of consolidation in both lungs and RP-ILD, with around 3.0% being identified as the potential threshold for change. <bold>Conclusion:</bold> Quantitative features of GGO and consolidation on HRCT are linked to RP-ILD in patients with MDA5 or ASS. Additionally, there is a negative relationship between these HRCT features and lung function.