Gastrointestinal (GI) diseases are largely influencing quality of life with about 2.8 million new GI cancers per year. The mortality is about 65%, but it is heavily influenced by early detection of neoplasia, where a 1% increase in detection can decrease the risk of cancer by 3%. The challenge is that as many as 20% of potential dangerous anomalies are missed during examinations. This is a situation that can be mitigated by an automatic disease detection system assisting the clinicians in real-time. We propose to deploy a small device running an artificial intelligence (AI) system which is connected directly to the video output of medical endoscopy equipment. The AI will, in real-time, make doctors aware of potential anomalies in the endoscopy image in order to aid the doctor and improve anomaly detection rates. The current AI prototype has GI disease detection rates at least as good as current state-of-the-art academic research, but has superior speed for scale and real-time feedback. Further, the AI can extract critical parts of the video and automatically generate a summary text report on the anomaly findings, significantly reducing the administrative burden for the physician after the procedure.
Project leader: Andreas Petlund
Institution: AUGERE MEDICAL AS