Case Study: Banner Health improves COVID chest CT diagnosis and outcomes research with Google for Education and Google Cloud

Google Cloud supports radiology research to improve COVID diagnoses on CT

Banner Health, through radiologist Dr. Kathryn Olsen, needed a faster, more collaborative way to study COVID-19 chest CT scans and determine whether a distinctive lung injury pattern could improve diagnosis and predict outcomes. Using Google for Education’s cloud ecosystem through Flywheel on Google Cloud, the team managed and analyzed deidentified imaging data from multiple Colorado health systems during a time-sensitive pandemic research effort.

Google for Education helped enable a secure, centralized research workflow with Google Compute and Google Cloud Storage, allowing blinded reader studies, data capture, and AI-assisted analysis of CT datasets. The result was a finding that hospitalized COVID-19 patients were over ten times more likely to show a lung injury pattern than a classic viral pneumonia pattern, with early AI work from Imbio also suggesting consolidation may predict poor clinical outcomes.

Banner Health

Kathryn Olsen

Cardiothoracic Radiologist