Clinical Motivation
CT is very sensitive to detection of respiratory disease
-
- Requires improved specificity to discriminate COVID-19 from other pneumonia
Chest radiography allows rapid deployment and fast workflow in challenging emergent / field scenarios
-
- Significant role in beside monitoring of disease progression / treatment response
Essential Groundwork
Large image datasets (COVID +/- patients)) for analysis and AI development
-
- IRB protocol submitted
- Coordination with Radiological Society of North America (RSNA)
- Radiology (Yi and Lin) + MCEH (Siewerdsen and Sulam)
Imaging Protocols (CT and Chest X-Ray)
-
- Ensure consistent and optimal imaging protocols across the Hopkins Radiology enterprise (ED and Radiology)
- More consistent imaging protocols
- Improved large-scale retrospective studies
- Radiology (Mahesh) + MCEH (Siewerdsen) + Siemens (Fung) + Carestream (Yorkston)
Image Analytics
High-Resolution CT – Improved Detection of COVID-19
-
- COVID-19 may present distinct patterns of morphology from other pneumonia
- Hopkins has the first “Ultra-High-Resolution” (UHR) CT system (Canon Precision CT) ~2x improvement in spatial resolution
- Siewerdsen (MCEH) + Lima (Cardiology) + Hansel, Hassoun (Pulmonology) + Lin, Yi (Radiology) + Canon USA
Chest Radiography (CXR) – Improved Detection of COVID-19
-
- Image analytics could improve the sensitivity / specificity of CXR
- Exploit advantages in CT in cost, workflow, deployability, and reduced infection risk
- Siewerdsen (BME) + Lin, Yi (Radiology) + Mahesh (Medical Physics) + Siemens
Investigators
Jeff Siewerdsen, John C. Malone Professor of Biomedical Engineering, Johns Hopkins University
Paul Yi, Co-Director, Radiology Artificial Intelligence Lab (RAIL), Johns Hopkins Medicine
Mahadevappa Mahesh, Professor of Radiology and Radiological Sciences, Johns Hopkins School of Medicine
Cheng Ting Lin, Assistant Professor of Radiology and Radiological Science, Johns Hopkins School of Medicine