Clinical Report: Multimodal Model for Early Detection of Respiratory Failure in DM-ILD

Overview

This study presents a multimodal model developed to predict 90-day respiratory failure in patients with dermatomyositis-associated interstitial lung disease (DM-ILD). The model demonstrated high discriminative performance.

Background

Dermatomyositis-associated interstitial lung disease (DM-ILD) is a severe condition often leading to respiratory failure and high mortality rates, especially in patients with anti-MDA5 antibody positivity. Early identification of high-risk patients is critical for timely intervention, yet routine testing for anti-MDA5 antibodies is not consistently available. This study aims to address the gap in risk assessment by utilizing data collected within the first 48 hours of patient admission.

Data Highlights

Key Findings

• The early-fusion random forest model showed the best performance with an AUC of 0.967.
• SHAP analysis identified arthritis, pulmonary function indices, laboratory markers, and latent CT features as influential predictors.
• Data were collected within the first 48 hours of admission, excluding anti-MDA5 results.
• The model was internally tested on a cohort of 124 adult patients with DM-ILD.
• Respiratory failure was chosen as a clinically relevant outcome for prediction.

Clinical Implications

This model may assist clinicians in stratifying patients at risk for respiratory failure in DM-ILD when serologic testing is delayed.

Conclusion

The admission-based multimodal model demonstrates performance for predicting 90-day respiratory failure in DM-ILD patients.

Related Resources & Content

1. npj Digital Medicine, 2026 -- A device-invariant multi-modal learning framework for respiratory disease classification
2. Frontiers in Medicine, 2026 -- A machine learning-based classification model for interstitial lung disease in rheumatoid arthritis
3. Frontiers in Medicine, 2026 -- Multimodal machine learning predicts type 2 respiratory failure in COPD exacerbations: a multicenter XGBoost model with clinical nomogram
4. ERS/EULAR clinical practice guidelines for connective tissue disease-associated interstitial lung disease, 2026
5. Interstitial Lung Disease Clinical Practice Guidelines | American College of Rheumatology, 2023
6. Frontiers in Medicine — Multimodal data integration and machine learning methods for early detection and risk prediction of pulmonary diseases in athletes
7. ERS/EULAR clinical practice guidelines for connective tissue disease-associated interstitial lung disease
8. Interstitial Lung Disease Clinical Practice Guidelines | American College of Rheumatology
9. Rituximab versus intravenous cyclophosphamide in patients with connective tissue disease-associated interstitial lung disease in the UK (RECITAL): a double-blind, double-dummy, randomised, controlled, phase 2b trial - ScienceDirect
10. Risk factors for mortality in anti-MDA5 antibody-positive dermatomyositis with interstitial lung disease: a systematic review and meta-analysis - PubMed
11. Prognostic analysis of MDA5‐associated clinically amyopathic dermatomyositis with interstitial lung disease - PMC
12. Mortality Risk Prediction in Patients With Antimelanoma Differentiation-Associated, Gene 5 Antibody-Positive, Dermatomyositis-Associated Interstitial Lung Disease: Algorithm Development and Validation - PubMed