Clinical Case Summary

The detection and management of overt gastrointestinal (GI) bleeding may be challenging, especially in case of limited resources, as during the COVID-19 pandemic. This case report shows how artificial intelligence (AI) technology applied to capsule endoscopy (CE) can contribute to this challenge.
In August 2020, a 80 year-old male patient was hospitalized for symptomatic iron-deficiency anemia three weeks after cardiovascular surgery and the start of dual antiplatelet therapy (DAPT). Since esophagogastroduodenoscopy (EGD) and ileocolonoscopy excluded any source of GI bleeding and the patient was stable, he was discharged. Two months later, the patient returned to the Emergency Department (ED) presenting with melena, and CE was recommended to rule out suspected small bowel bleeding.
Due to COVID-19, CE was not promptly scheduled. After two months, the patient was readmitted to the ED for persistent anemia (Hgb 7g/dL) and melena, without hemodynamic instability. Outpatient CE (Navicam SB system –AI tech embedded, Ankon, China) was performed two days later. The following morning the video was downloaded in 10 minutes and promptly reviewed in AI-assisted modality. After few seconds, active bleeding originating from the duodenum or proximal jejunum was detected. The patient was called back and an urgent EGD was performed after 4 hours from CE review. An actively bleeding Dieulafoy’s lesion was found in the third portion of the duodenum and treated with mechanical hemostasis. At 30-months follow-up, no episodes of re-bleeding occurred and Hgb was stable at 11 g/dL.
AI technology by Navicam SB system allows a rapid download of CE videos and offers a prompt detection of bleeding sources in patients with overt GI bleeding, leading to timely endoscopic management. This may potentially improve the long-term patients’ outcome and reduce the hospital stay, being particularly important in case of limited resources, as experienced during the COVID-19 pandemic.