Can AI Match Radiologists in ED Chest Radiograph Interpretation, Achieving Comparable Clinical Accuracy and Textual Quality?
7 Oct 2023 • In the emergency department (ED), timely interpretation of diagnostic imaging is a crucial component in clinical decision-making for otherwise undifferentiated patients. In a recent diagnostic study which was published in JAMA & conducted in an emergency department setting, researchers evaluated the performance of a generative artificial intelligence (AI) model in producing chest radiograph reports compared to reports generated by in-house radiologists and a teleradiology service.
- They analyzed 500 chest radiographs from different patients and had six emergency department physicians rate the quality and accuracy of each report using a 5-point Likert scale.
- The findings revealed that the AI-generated reports were comparable in clinical accuracy to those generated by radiologists, and both AI and radiologist reports received significantly higher ratings in terms of textual quality compared to teleradiology reports.
- Secondary analyses showed no significant differences in the probability of reports from any of the three sources containing clinically significant discrepancies for various findings, including cardiomegaly, pulmonary edema, pleural effusion, infiltrate, pneumothorax, and support devices.
Overall, the study suggested that implementing this generative AI tool in clinical workflows could aid emergency department physicians in interpreting chest radiographs more efficiently.
Source: JAMA | Read full story