At the present time, real-time web-based imaging AI analytics for chest radiography is proving to be useful in assisting radiologists as a companion, in detecting an abnormality or mass on checkup chest films. It may reduce some cost for unnecessary studies, for instance chest tomosynthesis or high resolution computed tomography (CT) chest examinations. Sampling-case studies found that suspicious lesions especially lung nodules, parenchymatous disease and pneumothorax are sensitive for detection. However, AI analytics are not yet achieving satisfactory results in the detection of bony structures, heart and mediastinal lesions. Furthermore, it also may cause false positive results when a chest film examination is not a true PA (posterior-anterior) or AP (anterior-posterior) position.
artificial intelligence, AI, chest image, chest radiograph, chest tomosynthesis, fractured ribs, chest screening, pulmonary tuberculosis
Thanis Saksirinukul, MD Radiology Department, Bangkok Hospital, 2 Soi Soonvijai 7, New Petchburi Rd, Bangkok 10310, Thailand. email: [email protected]
Received: June 5, 2018
Revision received: June 7, 2018
Accepted after revision: July 5, 2018
BKK Med J 2018;14(2): 59-63
DOI: 10.31524/bkkmedj.2018.09.011