The automated system for diagnosing chest X-ray images using artificial intelligence (AI) has recently shown great potential in helping doctors make quick and accurate decisions. However, the performance of these systems has not been evaluated closely to actual performance in clinical practice. In this study, the authors demonstrate a systematic evaluation mechanism for abnormal detection on chest X-ray images VinDr-CXR deployed at Phu Tho General Hospital – a provincial hospital in the North of Vietnam. Male. This AI system is directly integrated into the hospital’s medical image storage and transmission system (PACS). Performance evaluation ensures objectivity based on matching between AI-generated results and diagnostic reports of 6,285 chest X-rays extracted from Hospital Information System (HIS) over two months By the end of 2020. Evaluation shows that VinDr-CXR scores F1 0.653 (95% CI 0.635, 0.671) in detecting abnormalities on chest X-ray images.
A clinical validation of VinDr-CXR, an AI system for detecting abnormal chest radiographs
Tác giả: 
VBD
VBD 
