The role of artificial intelligence in predicting COPD exacerbations using multimodal data: a systematic review

Area of Science:

• Pulmonary Medicine
• Medical Informatics
• Artificial Intelligence

Background:

• Chronic Obstructive Pulmonary Disease (COPD) is a major global health burden.
• Acute exacerbations significantly worsen COPD progression and increase healthcare use.
• Artificial intelligence (AI) offers novel approaches for predicting COPD exacerbation risk.

Purpose of the Study:

• To systematically review AI-based models for COPD exacerbation prediction.
• To evaluate models integrating multimodal data sources.
• To assess the performance and generalizability of these AI models.

Main Methods:

• Systematic review following PRISMA 2020 guidelines.
• Searched PubMed, Embase, and Google Scholar for relevant studies.
• Included studies published between 2021-2025 focusing on AI models for COPD exacerbation prediction using combined data.

Main Results:

• Five studies met inclusion criteria, utilizing diverse AI models (e.g., random forests, CNNs).
• Models demonstrated moderate to high discriminative performance (AUC 0.73-0.92, accuracy up to 0.92).
• High sensitivity (0.94) noted in wearable-driven models, but external validation was limited.

Conclusions:

• Multimodal AI shows potential for early COPD exacerbation detection.
• Methodological heterogeneity and limited external validation hinder clinical translation.
• Future research requires transparent reporting, external validation, and real-world integration.