The outbreak and spread of Coronavirus (COVID-19) has created an urgent need in hospitals worldwide for clinicians to swiftly diagnose COVID-19 in patients and identify those most likely to become acute cases. This project is developing AI which can quickly and accurately diagnose whether a patient is COVID -19 positive /negative and support clinicians in deciding the most appropriate treatment approaches.
Aims & Objectives
This work package aims to improve understanding and develop rapid clinical impact in the following areas:
● Diagnosis and stratification of patients presenting with suspected COVID-19
● Improved understanding of the presentation of COVID-19 in imaging, which in combination with non-imaging data sources will improve our knowledge of the disease aetiology, supporting the development of enhanced clinical decision support for the treatment of COVID-19
This will be achieved through the following approaches:
● Identifying abnormal chest x-ray (CXR) images to enable a point-of-care risk stratification tool
● Identifying imaging biomarkers and quantifying the features in non-contrast chest CT scans that are associated with different stages of disease severity. This will further the community’s understanding of the disease aetiology and will enable objective measurement of disease severity/progression
● Combining features extracted from imaging with information from the patient’s EHR (including comorbidities, medications and other relevant structured and unstructured patient data) to develop a risk stratification tool enabling patients of greatest concern to be identified for timely targeted intensive treatment.
Work Package Lead: NHS Greater Glasgow & Clyde
Contributing Partners: Bering Ltd, Canon Medical Research Europe Ltd & NHS Greater Glasgow & Clyde