Artificial intelligence in respiratory pandemics—ready for disease X? A scoping review
-
By
-
Jennifer Straub
-
Enrique Estrada Lobato
-
Diana Paez
-
Georg Langs
-
Helmut Prosch
-
November 21, 2024
-
Clinical Scorecard: The Role of Artificial Intelligence in Managing Respiratory Disease Outbreaks: A Scoping Review on Preparedness for Future Threats
At a Glance
| Category | Detail |
|---|
| Condition | Respiratory disease outbreaks including SARS, H1N1, MERS, and COVID-19 |
| Key Mechanisms | Use of medical imaging and artificial intelligence (AI) for diagnosis, prognosis, therapy guidance, outbreak detection, and forecasting |
| Target Population | Patients affected by respiratory pandemics globally |
| Care Setting | Hospital and public health settings with radiology and data-sharing infrastructure |
Key Highlights
- Medical imaging is critical for early diagnosis, prognosis, and therapy guidance in respiratory pandemics.
- AI has potential to enhance outbreak detection, case forecasting, drug development, and optimize imaging interpretation.
- International data collection and sharing remain limited by legal and procedural barriers, hindering AI development.
Guideline-Based Recommendations
Diagnosis
- Use chest radiography as a cornerstone for diagnosis and triage in early pandemic phases.
- Incorporate radiographic changes in case definitions to support early identification.
Management
- Monitor therapy response through serial chest radiographs.
- Base treatment choices on radiographic findings and clinical guidelines.
Monitoring & Follow-up
- Recommend radiographic follow-up to assess disease progression and recovery.
- Use imaging data to guide discharge decisions following WHO guidelines.
Risks
- AI applicability is limited if training data lacks diversity or has poorly defined diagnoses.
- Legal and publication rights issues impede timely international data sharing.
Patient & Prescribing Data
Patients with confirmed or suspected respiratory viral infections during pandemics
Radiographic monitoring informs treatment response; no standardized therapy existed early in outbreaks like SARS.
Clinical Best Practices
- Implement routine surveillance and early imaging to detect novel respiratory pathogens.
- Develop and maintain international standardized data collection protocols to facilitate AI algorithm development.
- Ensure diversity and quality in AI training datasets to improve generalizability and accuracy.
- Address legal and ethical barriers to enable rapid data sharing during outbreaks.
References
