Missed diagnosis on chest X-ray: Auditing large volumes of data with the help of comprehensive artificial intelligence
Author
Talwar, Arpit | St. Vincent’s Hospital Melbourne, Australia
Scientific poster presentation (W5B-SPCH-2) at RSNA 2023, 26 – 30. November 2023 in Chicago, US
Purpose
Can comprehensive AI decision-support help audit missed diagnosis on chest X-rays?
Method
Retrospective analysis of 1,595 CXR studies performed in 2016 at St. Vincent’s Hospital Melbourne using Annalise Enterprise CXR for detecting a selected list of 60 findings, deemed significant/critical. Radiologist review in case of disagreement.
Results
The AI detected true missed critical findings in 6% of cases (n=97), including pulmonary nodules (16%), pleural effusions (16%), spinal compression fractures (12%), airspace opacities (11%), acute rib fractures (9%), and others.
Key Takeaway
Comprehensive decision-support AI can be a time-efficient and effective means of auditing large volumes of CXR studies for missed findings, potentially improving quality and patient care.
Disclaimer
Harrison.ai Radiology Solutions were previously marketed as Annalise.ai solutions.
