Designed to help prevent errors, the AI software analyzes lung CT scans and associated radiology reports to find any discrepancies between its findings and the radiologist reports. It then flags these cases for further review by clinicians.
In a pilot study at Sutter, a chest radiology expert reviewed these potential discrepancies and followed up with the initial radiologist if needed. During the first 90 days of the pilot, the software reviewed over 10,000 CT scans that contained lung tissue; 83 findings flagged by the algorithm were considered to warrant additional review and intervention. A subset of these cases warranted additional care, according to Ferrum.
Copyright © 2020 AuntMinnie.com