CT/Chest Radiography COVID-19 Open-source AI Prognostication Predictive Model

COVID-19 Research Area(s): Diagnostics, Genomics & Transmission Dynamics, Equipment & Technology Innovations

To develop an open-source prognostic and predictive machine learning model for COVID-19 on Chest X-rays and CT images, in partnership with AWS. A dataset of 15,000 DICOM imaging studies with the corresponding Radiologists' final report, clinical and pathology data will be collected from the Radiology (PACs, RIS) and clinical databases and anonymized by SapienSecure software, which is provided freely by SapienML. SapienSecure will completely remove and encrypt the direct patient identifiers within DICOM headers and the Radiologists' final report. Indirect patient identifiers will be limited with a k-anonymization