Efficient clinical decision-making process via AI-based multimodal data fusion: A COVID-19 case study.

Area of Science:

• Artificial Intelligence in Medicine
• Medical Imaging Analysis
• Clinical Decision Support Systems

Background:

• COVID-19 pandemic necessitated efficient resource management and risk stratification.
• Computer-aided diagnosis aids clinicians in identifying high-risk patients.

Purpose of the Study:

• To develop a fully-automatic method for COVID-19 risk estimation using multimodal data fusion.
• To assess the risk of hospitalization and mortality in COVID-19 patients.

Main Methods:

• Utilized multimodal data fusion combining clinical features and deep features from chest X-ray images.
• Developed a novel, efficient, and fully-automatic machine learning model.
• Evaluated model performance using Area Under the Receiver Operating Characteristic Curve (AUC-ROC).

Main Results:

• Achieved high performance in predicting hospitalization (AUC-ROC 0.8452 ± 0.0133) and mortality (AUC-ROC 0.8285 ± 0.0210).
• Identified key features for each risk scenario, showing distinct roles for clinical and imaging data.
• Demonstrated that multimodal data fusion outperforms single-source data approaches.

Conclusions:

• The developed method effectively aids clinical decision-making for COVID-19 risk stratification.
• Multimodal data fusion offers advantages, even with reduced feature sets, beneficial for resource-limited settings.
• The approach shows potential for broader applications in managing other clinical scenarios.