Development of an Artificial Intelligence Algorithm

The abdominopelvic tomodensitometric examination is the gold standard for making a diagnosis and identifying the cause, the level of the lesion and looking for signs of severity related to the risk of ischemia of the intestinal tissue. Well characterized, these signs remain difficult to identify for a non-expert radiologist. The evaluation of the detection performance of these signs shows a sensitivity between 63 and 100% and a specificity between 61 and 96% with a great variation according to the signs studied. Our research project aims to develop a tool for the detection of the three main signs of severity of mechanical intestinal obstruction to raise and standardize the level of detection of the risk of ischemia on abdominal CT images.