Along with the dramatic advance in imaging technologies, the quality and quantity of endoscopic images have increased significantly, which, however, raises the hurdle to high-performance diagnosis. Thus, only a limited number of expert endoscopists can achieve such highly accurate diagnoses. To overcome this issue and allow any endoscopist to provide high-quality diagnosis regardless of their skill and environment, computer-aided diagnosis (CAD) using artificial intelligence (AI) is gaining increasing attention. In fact, the recent rapid development of AI in terms of hardware and software has resulted in implementation of CAD into real clinical endoscopy and the first approved AI for colonoscopy is now available since March 2019 in Japan. This review overviews the current situation of research on AI for endoscopy in the esophagus, stomach, and colon. At the same time, we mention barriers that are hindering clinical implementation of AI for endoscopy and the importance of regulatory approval, which will facilitate the understanding of the current status of AI-assisted endoscopy.

Video capsule endoscopy was first developed in 2000 for the purpose of diagnosing small intestinal diseases and has been used for diagnosis of various gastrointestinal diseases. In 2006, colon capsule endoscopy was reported. Further improved 2^nd-generation colon capsule endoscopy was developed. In Japan, colon capsule endoscopy was approved by the Ministry of Health, Labor and Welfare, and covered by health insurance in 2014. The efficacy of colon capsule endoscopy in monitoring inflammatory bowel disease, mainly ulcerative colitis, which causes inflammation in the colon, has been reported. Colon capsule endoscopy enables minimally invasive observation of the large intestinal mucosa, which is appropriate as a monitoring tool for ulcerative colitis. This article outlines the current status of colon capsule endoscopy for the assessment of ulcerative colitis.

A 76-year-old woman with tarry stools and vomiting of blood was admitted to our hospital. She has been followed as a patient with Osler-Weber-Rendu disease in our clinic for five years. She had previously been treated with endoscopic hemostasis (high-frequency soft coagulation and argon plasma coagulation) and blood transfusion therapy multiple times for oozing bleeding of gastric angioectasia. At the current admission, endoscopic band ligation (EBL) was performed with the expectations that gastric telangiectasia would disappear due to ulcer scar formation and the treatment time would be shortened. Recurrent gastric telangiectasia disappeared in the area where ulcer scar formation by EBL was obtained. An endoscopic band ligating device was useful for hemostasis of Osler-Weber-Rendu disease in the stomach.

A 79-year-old male patient was transported to our emergency department due to loss of consciousness and hematochezia. Under the diagnosis of hemorrhagic shock, he underwent emergent colonoscopy. A tumorous lesion was found in the ascending colon and the site of active bleeding was confined to the lesion. Endoscopic mucosal resection (EMR) was performed with the aim of hemostasis, and pathological examination of the resected specimen revealed mixed adenoneuroendocrine carcinoma (MANEC). Since the tumor had invaded the muscularis propria, additional laparoscopic ileocecal resection with D3 lymph node dissection was performed. The pathological diagnosis of the tumor was pT2 pN1 M0, stage Ⅲa carcinoma, and postoperatively the patient received 8 courses of adjuvant XELOX chemotherapy. Colon MANEC is associated with early lymph node and distant organ metastases, and is resistant to multimodal and multidisciplinary treatments leading to an unfavorable prognosis. This report describes a case of colon MANEC presenting with hemorrhage as the initial symptom.

An 83-year-old woman who had undergone cholecystectomy for cholecystolithiasis at the age of 40 years presented with pain in the right hypochondriac region. Laboratory analysis showed elevated levels of hepatic and biliary enzymes. Computed tomography revealed an enhanced nodule at the junction of the cystic duct. Peroral cholangioscopy revealed a smooth protrusion that was covered with normal mucosa at the junction of the cystic duct. An amputation neuroma was suspected; thus, we performed cholangioscopy-guided boring biopsy to sample the submucosal tissue. Histopathological examination of the biopsy specimen showed connective tissue containing S-100 protein-positive nerve fibers; this finding is consistent with an amputation neuroma. The patient presented with symptoms secondary to biliary obstruction; therefore, extrahepatic bile duct resection and hepaticoduodenostomy were performed. The postoperative histopathological diagnosis was amputation neuroma of the cystic duct remnant. It is difficult to diagnose amputation neuroma by imaging, and boring biopsy under peroral cholangioscopy is useful for histological diagnosis before surgery is performed.

A 76-year-old woman was admitted to our hospital for removal of a common bile duct stone. Endoscopic retrograde cholangiopancreatography (ERCP) revealed a 25mm×16mm stone in the common bile duct. When the stone was grasped with a mechanical lithotripsy basket made of Nytinol, fracture of the metal wire occurred, resulting in basket impaction. Then, a 15mm snare was inserted into the bile duct over the metal wire of the basket, and the tip of the basket was grasped with the snare. Next, the snare was pulled, and the basket became inverted and deformed in the bile duct. As a result, the stone that had been grasped by the basket was successfully released. We herein report a case in which a basket inversion technique using a snare was useful for complicated mechanical lithotripsy basket impaction.

Background and aim: Cold snare polypectomy (CSP) is a safe and easy procedure that has recently become widely performed for resecting small colorectal polyps. However, the polyps are sometimes unexpectedly polyps with advanced neoplasia. It is thus important to evaluate both the horizontal and vertical margins of resection. Here, we retrospectively investigated the resection depth of CSP.

Methods: We examined 503 small colorectal polyps that were resected by CSP between August 2017 and July 2018. We histologically evaluated the resected layers, and divided the samples into two groups according to the histological resection depth: the M group included samples whose resection depth reached the mucous membrane, and the S group included samples whose resection depth reached the muscularis mucosa or submucosa. We performed a comparative review of the size, location, horizontal margin, appearance of the cut surface of endoscopic resection, and inter-operator variations between the two groups.

Results: The resected layers varied among the samples: the resection depth reached the mucous membrane in 274 samples (55%), reached the muscularis mucosa in 193 samples (38%), and reached the submucosa in 36 samples (7%). The size, appearance of the cut surface of endoscopic resection, and inter-operator variations did not differ significantly between the M and S groups. In contrast, the location of the polyps differed significantly between the two groups, with the cecum more prevalent in the M group and the rectum more prevalent in the S group. Also, excluding the 106 samples that had an unclear horizontal margin, the negative horizontal margin rate was significantly higher in the S group (109/173; 63%) than in the M group (85/224; 28%).

Conclusions: The resection depth differed according to the location of the polyp, with the cecum more prevalent in the M group and the rectum more prevalent in the S group. The S group was more likely to have a negative horizontal margin. However, the resection depth achieved with CSP reached the muscularis mucosa in less than half of the cases. Given the inconsistency of resection depth, indication of CSP should be carefully decided as for larger tumor diameter.

Small-bowel bleeding is most frequently caused by a small-bowel vascular lesion. Small-bowel angioectasia makes up the majority of small-bowel vascular lesions. Small-bowel hemangioma is a rare disease but can sometimes cause small-bowel bleeding. However, there is no consensus on the classification and optimal treatment of both lesions.

We report cases of small-bowel angioectasia and small-bowel hemangioma that were treated endoscopically with polidocanol injection therapy using double-balloon endoscopy and obtained good results.

Lower rectal lesions close to the dentate line can be difficult to resect endoscopically because of the risk of bleeding from the rectal venous plexus, the sensory nerves in the squamous epithelium below the dentate line which cause pain during the procedure, the narrow lumen in proximity of the anal sphincter which makes it difficult to have a good visual field, and the presence of hemorrhoids. Special measures employed for endoscopic submucosal dissection (ESD) of lesions close to the dentate line are as follows: A transparent hood is attached to the tip of the endoscope. Lidocaine is locally injected into the submucosal squamous epithelial layer on the anal side of the lesion. The first mucosal incision is performed, and submucosal dissection is initiated on the anal side of the lesion. A shallow peripheral mucosal incision is made. Blood vessels are appropriately handled with hemostatic forceps. Using these measures, ESD for anorectal tumors close to the dentate line can be a safe and effective therapeutic procedure regardless of the presence of hemorrhoids.

Background and Aim: With an aging population, an increasing number of individuals on antithrombotic agents are diagnosed with large bile duct stones. Studies have shown the effectiveness of endoscopic papillary large balloon dilation (EPLBD) for removal of large bile duct stones. EPLBD without endoscopic sphincterotomy (EST) may reduce the risk of procedure-related bleeding, but the safety of this procedure for users of antithrombotic agents remains unclear.

Methods: In this multicenter retrospective study, we included patients who underwent EPLBD without EST for bile duct stones between March 2008 and December 2017. We compared adverse events and other clinical outcomes between users and non-users of antithrombotic agents (antiplatelet agents and anticoagulants).

Results: We analyzed a total of 144 patients (47 users and 97 non-users of antithrombotic agents). Among the users, the agents were continued in 13% and were replaced with heparin in 62% during the periprocedural period. We did not observe clinically significant bleeding and thrombotic events irrespective of the use of antithrombotic agents. Overall rate of early adverse events did not differ between users and non-users (6.4% and 7.2%, P=0.99). Procedural outcomes did not differ between the groups (necessity for lithotripsy, 28% vs. 29%; and complete stone removal in a single session, 72% vs. 71%, for users and non-users, respectively).

Conclusions: Endoscopic papillary large balloon dilation without EST may be done without a substantial increase in procedure-related bleeding for users of antithrombotic agents. A larger study is required to refine the management strategy for those agents during the periprocedural period.