Article Sidebar
Views pdf: 68
ROLE OF ARTIFICIAL INTTELIGENCE IN CHEST X RAY INTERPRETATION: PRELIMINARY RESULTS AT CHO RAY HOSPITAL
Main Article Content
Abstract
Background: Chest X-ray is currently a widely used diagnostic modality, applied in screening, diagnosis and post treatment follow-up. In order to make a quick and accurate diagnosis, artificial intelligence (AI) - a computer science branch has developed rapidly with many applications in medicine - has been applied in many health facilities in Viet Nam. At Cho Ray hospital, the AI system supporting chest X-ray interpretation has been applied for more than 1 year with 24120 cases processed by AI. This study aims to initially evaluate the role of artificial intelligence in chest X-ray interpretation at Cho Ray hospital.
Methods: Randomly collecting data from chest X-ray images taken at the exit examination department of Cho Ray hospital from April 1, 2023 to May 1, 2023. Radiographic images were analyzed by a 5-year experienced radiologist at Cho Ray hospital and used as reference data. Chest X-ray imaging characteristic including consolidation, interstitial lesions, pulmonary cavities, linear interstitial thickening, calcified nodules, pulmonary nodules, atelectasis, pleural effusion, pleural thickening were observed. The interpreting time was also recorded. The data was then analyzed by two residents with and without AI assistance, the variables were collected for comparison.
Results: The average time for an experienced doctor to analyze the results is: 55.17 ± 32.43 seconds, with AI assistance, this time is shortened to: 16.57 ± 13.78 seconds. Sensitivity for detecting general signs on PA chest radiographs of resident without AI was 73.01%, specificity was 83.68%, with AI-assisted, the sensitivity and specificity increased to 97.51% and 94.90%, respectively. For the group of important signs, suggesting TB, the resident's sensitivity and specificity were 73.01% and 83.68%, respectively; with AI-assisted, the sensitivity and specificity increased, reaching 97.49% and 94.83%. In addition, with AI application on PACS/RIS platform, the average time for technicians to perform a chest radiograph is also significantly reduced from 93.57 ± 7.06 seconds to 34.86 ± 2.67 seconds.
Conclusions: AI application assist in chest radiograph interpreting time reduction while improving the sensitivity and specificity.
Article Details
Keywords
AI, artificial intelligence, chest x-ray, PACS/RIS.
References
2. Cicero, M., et al., Training and validating a deep convolutional neural network for computer-aided detection and classification of abnormalities on frontal chest radiographs. Investigative radiology, 2017. 52(5): p. 281-287.
3. Hurt, B., et al., Augmenting interpretation of chest radiographs with deep learning probability maps. Journal of thoracic imaging, 2020. 35(5): p. 285.
4. Lee, C.S., et al., Cognitive and system factors contributing to diagnostic errors in radiology. American Journal of Roentgenology, 2013. 201(3): p. 611-617.
5. Nam, J.G., et al., Development and validation of deep learning–based automatic detection algorithm for malignant pulmonary nodules on chest radiographs. Radiology, 2019. 290(1): p. 218-228.
6. Organization, W.H., Rapid communication on systematic screening for tuberculosis. 2020.
7. Radiation, U.N.S.C.o.t.E.o.A., Effects of ionizing radiation. Scientific Annexes E, 2008: p. 203-204.
8. Rajpurkar, P., et al., Chexnet: Radiologist-level pneumonia detection on chest x-rays with deep learning. arXiv preprint arXiv:1711.05225, 2017.
9. Seah, J.C., et al., Effect of a comprehensive deep-learning model on the accuracy of chest x-ray interpretation by radiologists: a retrospective, multireader multicase study. The Lancet Digital Health, 2021. 3(8): p. e496-e506