COVID-19 therapy optimization by AI-driven biomechanical simulations.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Pulmonology

Background:

• COVID-19 can lead to severe pneumonia and Acute Distress Respiratory Syndrome (ARDS).
• Early identification of ARDS risk is crucial for effective patient management and resource allocation in intensive care units.

Purpose of the Study:

• To develop and assess an AI-based prognostic system for predicting respiratory efficiency in COVID-19 patients.
• To identify patients at high risk of developing ARDS for timely clinical intervention.

Main Methods:

• Utilized lung Computed Tomography (CT) scans, biomechanical simulations of lung air flux, and Arterial Blood Gas (ABG) analysis as input.
• Developed a preliminary AI algorithm to predict oxygen exchange and correlate ABG parameters with CT-derived morphological information and disease outcomes.
• Investigated the system's feasibility on a small clinical database of COVID-19 patients.

Main Results:

• Demonstrated promising results from a preliminary version of the AI prognostic algorithm.
• Found correlations between the time evolution of ABG parameters, CT scan morphological data, and patient outcomes.
• Showcased the potential for predicting the progression of respiratory efficiency in COVID-19 patients.

Conclusions:

• The proposed AI system shows potential for predicting ARDS development in COVID-19 patients.
• Integrating CT imaging, biomechanical simulations, and ABG analysis can enhance prognostic accuracy.
• Predicting respiratory efficiency evolution is vital for optimizing COVID-19 patient care and intensive care unit resource management.