Summary:
UCLA researchers in the Department of Radiological Sciences have developed a quantitative imaging tool designed to standardize the diagnostic scoring of myositis lung disease (MLD) for improved patient intervention and monitoring.
Background:
Myositis lung disease (MLD) is a complex autoimmune condition characterized by muscle inflammation and progressive lung fibrosis, resulting in severe respiratory decline. Since MLD is a major driver of mortality in its patient population, the proper management and development of therapeutics rely on accurately tracking lung tissue changes over time for early intervention. Specifically, lung disease occurs in 33-65% of idiopathic inflammatory myopathies (IIM), a diverse group of diseases characterized by autoimmune-mediated inflammation of skeletal muscle. Approximately 80% of the mortality in IIM is associated with some form of lung disease, including MLD, and patients may often be misclassified with idiopathic pulmonary fibrosis (IPF). Currently, high-resolution chest computed tomography (HRCT) remains the clinical standard for the diagnosis and monitoring of MLD, where scans are manually evaluated by expert thoracic radiologists. Due to the reliance on subjective visual assessments and confounding traditional endpoints of MLD, there is an increasing demand for objective, quantitative HRCT. Current first-generation quantitative HRCT technologies rely on taxonomies designed for general fibrotic lung diseases, failing to capture complex patterns specific to myositis. To address this critical limitation, there is a need for a diagnostic and quantitative imaging tool that is specifically designed to quantify MLD for improved patient outcome.
Innovation:
Researchers at UCLA have developed a diagnostic and quantitative imaging biomarker tool, known as quantitative MLD scoring (QMS), designed for early and objective MLD identification. QMS utilizes a hybrid artificial intelligence framework that integrates deep machine learning with cognitive AI-based reasoning to establish a novel, MLD-specific taxonomy. This framework aligns computational data with human expertise on MLD disease patterns and clinical outcomes. QMS is capable of MLD quantification, measurement of longitudinal lung tissue change, early and accurate diagnosis of MLD, and the prediction of outcomes such as acute respiratory failure or time to death. This novel framework can significantly improve MLD and lung disease identification in IIM cases, improving patient outcome. Furthermore, QMS establishes an objective reproducible measure that is unconfounded by disease-related factors. By standardizing HRCT scoring across clinical sites, QMS enables reliable pooling of data, overcoming the recruitment and data limitations associated with rare diseases such as IIM. QMS establishes a scalable and disease-specific standard that can accelerate clinical trial timelines, improve risk stratification, and drive the development of novel therapeutics for MLD.
Potential Applications:
● Clinical Trial Endpoints
○ Drug efficacy in MLD
● Companion Diagnostic (CDx)
● Clinical Decision Support
● Predictive Prognostic Modeling
● Rare Disease Data Standardization
Advantages:
● Disease Specificity
○ Proprietary taxonomy
● Unconfounded and Objective Data
○ Eliminates human subjectivity
● Accelerated Trial Timelines
● Precision
○ Capable of detecting microscopic longitudinal changes not possible by human experts
Development-To-Date: First description of complete invention (oral or written)
Related Papers:
● Brown, Matthew, et al. "Quantitative CT and Artificial Intelligence in Myositis-associated Interstitial Lung Disease: A Review." Journal of Thoracic Imaging, 2024, journals.lww.com/thoracicimaging/fulltext/9900/quantitative_ct_and_artificial_intelligence_in.206.aspx.
Reference:
UCLA Case No. 2026-189
Lead Inventors:
Jonathan Goldin, Grace Hyun Kim, Sangmee Bae, Matthew Brown