Endoscopic Add-On System to Classify Ulcerative Colitis (ENACT)
Endoscopic evaluation of the severity of disease in ulcerative colitis patients is subjective. As a result, there’s often significant variance between doctors – and even between evaluations by the same doctor. This leads to overtreating, undertreating or delays in treatment, all of which increase hospitalisations and costs.
The Inspiration Behind the Innovation
Master teaching/supervision is the current clinical best practice for evaluating the endoscopic severity of ulcerative colitis. Many of these evaluations are performed by young doctors who have minimal supervision and training in grading the severity of this disease. There's also a significant variance among assessments performed by all gastrointestinal doctors, including substantial disagreement between those with advanced experience, such as consultants, professors and chief physicians. This inconsistency among even experienced doctors indicates that the current “gold standard” is inadequate.
ENACT is an AI-enabled system that analyses videos from endoscopic procedures performed on ulcerative colitis patients. The system, which can differentiate between usable and unusable video frames, has achieved an accuracy of 89% in classifying each endoscopic frame. The next iteration of the system is an add-on to the existing endoscopic column. It analyses live video frames, enabling on-screen classification scores of the disease severity in real time and with minimal disturbance to daily clinical practice.
Other current solutions are only able to perform such analysis after the fact, requiring each procedure to be recorded and uploaded, which provides less-timely information while introducing GDPR challenges. ENACT’s real-time analysis improves the speed and accuracy with which doctors can evaluate the severity of each case, optimising early intervention, disease monitoring and treatment-related decisions.
Bobby Zhao Sheng Lo: PhD, MD; Gastrounit, Hvidovre Hospital
Flemming Bendtsen: Professor, MD, Senior Doctor; Gastrounit, Hvidovre Hospital
Ida Vind: Associate Professor, MD, Senior Doctor; Gastrounit, Hvidovre Hospital
Johan Burisch: PhD, MD, Specialist Registrar; Gastrounit, Hvidovre Hospital
Christian Igel: Professor; Department of Computer Science, Copenhagen University
Bulat Ibragimov: Associate Professor; Promotion Programme
Bjørn Leth Møller: PhD Student, MSc in Computer Science; Copenhagen University