Artificial intelligence to quantify important segments from CT in COVID-19

We are developing a diagnostic tool based on deep learning to quantify the most important pathological tissue subclasses in covid-19 from CT images: Ground-glass opacification, consolidation, and pleural effusion. These volumes may be important quantitative correlates for estimating prognosis in covid-19, as well as important signs to map out the disease itself. We hypothesize that the volumes of the pathological findings are prognostic for survival. The artificial intelligence - after calibration and testing - will be made freely available to anyone who wants to use it, and we will implement it for use into the OUS radiology workflow.

Project leader:

Atle Bjørnerud

Page visits: 738