Thoracic Point-of-Care Ultrasound: A SARS-CoV-2 Data Repository for Future Artificial Intelligence and Machine...

Area of Science:

• Medical Informatics
• Pulmonology
• Infectious Diseases

Background:

• Artificial intelligence (AI) and machine learning (ML) show potential in managing hospitalized patients, including those with COVID-19.
• Diagnostic and prognostic indicators for SARS-CoV-2 infection present challenges.
• Thoracic ultrasound is a valuable tool in critical care settings.

Purpose of the Study:

• To establish an international clinical protocol for collecting standardized thoracic point-of-care ultrasound (POCUS) data in hospitalized COVID-19 patients.
• To create a curated dataset of ultrasound images and clinical metadata for AI/ML modeling.
• To explore the future role of AI/ML in improving SARS-CoV-2 patient management and outcomes.

Main Methods:

• Development of an international clinical protocol for data collection.
• Standardized collection of thoracic point-of-care ultrasound (POCUS) data.
• Linking ultrasound images with patient clinical metadata.

Main Results:

• A protocol for standardized thoracic POCUS data collection in COVID-19 patients has been established.
• A corpus of curated ultrasound images and linked clinical metadata is being created.
• The foundation for future AI/ML modeling in SARS-CoV-2 patient management is being laid.

Conclusions:

• AI/ML holds potential for assisting in the management of SARS-CoV-2 patients.
• Standardized POCUS data is crucial for developing effective AI/ML models.
• This research resource may lead to improved patient outcomes in future pandemics.