Prediction of outcome in acute lower gastrointestinal hemorrhage: role of artificial neural network

Acute lower gastrointestinal hemorrhage (LGIH) has traditionally been defined as bleeding that occurs distal to the ligament of Treitz. More recently, however, it has been subdivided into mid-intestinal (small bowel) hemorrhage and bleeding that originates from the colon. Acute LGIH has diverse etiologies, is a frequent cause of hospital admission, and is associated with significant patient morbidity and mortality, as well as substantial economic cost. In contrast to hemorrhage from the upper gastrointestinal tract (UGIH), the management of acute LGIH is less well defined; furthermore, there is a paucity of published studies that evaluate predictive models in this disorder. Nonetheless, extrapolating from what is known in UGIH, the development of reliable predictive models in LGIH may lead to improved patient care and outcome, by enhancing clinical triage, and by the more cost-effective use of limited healthcare resources. In this review, we discuss the technical development and potential use of artificial neural network in patients presenting with acute LGIH.