Technical Name |
A deep learning-powered novel artificial intelligence algorithmsystem to assist in the identification of pneumoperitoneum on abdominal computed tomography |
Project Operator |
CHANG GUNG MEDICAL FOUNDATION |
Project Host |
薛承君 |
Summary |
Our developed artificial intelligence system is capable of rapid identification of pneumoperitoneum on CT within 5 minutes, with accuracyconsistency on par with an experienced board-certified radiologist. This is especially so if there are multiple image slices (4 image) with evidence of pneumoperitoneum, yielding a 100 accuracy rate. |
Scientific Breakthrough |
Our novel artificial intelligence system is capable of automatically identification of pneumoperitoneum on non-contrast abdominal CT scans within 5 minutes. The high diagnostic performance may be used as a preliminary screening tool for radiologistsclinicians. Non-contrast abdominal CT scans flagged by our artificial intelligence system as having evidence of pneumoperitoneum can be prioritised for review, ultimately decreasing morbiditymortality of patients with pneumoperitoneum. |
Industrial Applicability |
Our developed artificial intelligence system is capable of rapid identification of pneumoperitoneum on CT within 5 minutes, which can be used as an automated screeningtriaging tool to facilitate early identificationprompt intervention as appropriate to improve patient clinical outcomes.We estimate the global market revenue of this technology to be NT$24.9 billion. |
Keyword |
Pneumoperitoneum Free air Deep learning Artificial Intelligence Non-contrast abdominal computed tomography |
Notes |
|