The development of endoscopic AI advanced rapidly during the third wave of artificial intelligence due to the progress in deep learning technology. In 2018, the computer-aided diagnosis of endocytoscopy was approved as the first AI-equipped medical device in Japan. Since 2024, additional reimbursements have been introduced. Endoscopic AI is broadly classified into lesion detection and differential diagnosis.

In colonoscopy, meta-analyses have shown that computer-aided detection improves adenoma detection rates by 8-10% and significantly reduces miss rates. Prospective studies have also been conducted on gastric and esophageal squamous cell carcinomas.

However, in clinical practice, issues such as false positives, physician dependence on AI (deskilling), and optimization of human-AI interactions have become apparent. Although the clinical implementation of endoscopic AI is steadily progressing, further prospective studies are required to establish its true clinical utility.

The diagnosis of superficial non-ampullary duodenal epithelial tumors (SNADET), defined as adenomas and mucosal or submucosal adenocarcinomas, is increasingly common due to the widespread use of screening EGD. Most SNADETs are diagnosed in the first and second parts of the duodenum around the papilla of Vater. Diagnosis based on white-light endoscopy and image-enhanced endoscopy with magnification helps differentiate tumors from non-tumors, assessing tumor histological grade (e.g., low-grade adenoma [LGA] vs. high-grade adenoma or adenocarcinoma [HGA/Ca]), and identifying epithelial subtypes (e.g., gastric vs. intestinal). Gastric epithelial subtypes account for 10-20% of SNADETs and exhibit a higher histological grade than intestinal types. Tumor-like lesions located in the duodenal bulb can sometimes be difficult to distinguish between tumors and non-tumors, requiring comprehensive evaluation, including biopsy and endoscopic ultrasonography. For optimal treatment selection of SNADETs, differential diagnosis between LGA and HGA/Ca, considering the epithelial phenotypes, is crucial.

Objective: This study aimed to prospectively evaluate the efficacy of bowel preparation with PEG-Asc monotherapy for CS and identify factors associated with prolonged preparation time.

Methods: In total, 237 patients (mean age, 64.2 years; male-to-female ratio, 151:86) were enrolled. Laxatives were not administered, and all patients consumed a diet meal the day before the CS. The preparation was performed using the 2:1 method, in which the patients drink two cups of PEG-Asc and one cup of water or tea on the day of CS. The efficacy, safety, and factors that prolonged the preparation time were examined.

Results: The mean volume of PEG-Asc consumed was 1,474 mL, and the mean preparation time was 189.6 min. Adequate bowel cleansing was achieved in 99.2% of patients, and no severe adverse events were observed. According to multivariate analysis, the factors that prolonged the preparation time to more than 180 min were age over 65 years and no defecation on the day before and morning of the CS. The preparation time was significantly shorter if the patients defecated on the day before and on the morning of the CS, even if the amount was small.

Conclusions: This method is safe and effective for CS. Our results suggest that laxatives for adjuvant use are not necessary for patients who defecate the day before or on the morning of CS.

A 76-year-old man who underwent subtotal esophagectomy with gastric tube reconstruction for esophageal cancer was diagnosed with gastric tube cancer two years after the initial surgery. An open local resection of the gastric tube was performed. However, anastomotic leakage and a pulmonary fistula were observed postoperatively. The fistula was refractory to conservative management. Although surgery was considered, minimally invasive endoscopic treatment was selected. Endoscopic fistula closure using a polyglycolic acid sheet was successfully performed, and the patient recovered successfully. Endoscopic closure for refractory fistulas after gastric tube cancer surgery appears to be a minimally invasive and effective treatment method.

The patient was a 68-year-old woman who had undergone left iliac resection and tumor resection for left pelvic leiomyosarcoma at the age of 47 years. She visited the hospital due to vomiting, and a CT scan revealed stenosis of the descending colon adjacent to the iliac resection site and dilation at the oral side of the intestine. Although transanal ileus tube placement was considered technically difficult, successful placement was achieved with the aid of a colonoscope, which allowed release of sigmoid colon flexion. Since the introduction of self-expanding metal stents, the need for transanal ileus tube placement has decreased. However, in cases of benign intestinal strictures such as the present case, preoperative intestinal decompression may still require tube placement. When transanal ileus tube insertion is challenging, the use of a colonoscope to release colonic flexion can be considered a simple and effective method.

CT colonography (CTC) is a non-invasive examination technique with approximately 90% sensitivity and specificity for colorectal neoplasms ≥10 mm, although its use in Japan remains limited owing to a lack of awareness and shortage of experienced readers. Accuracy can be maintained through appropriate bowel preparation, colonic distention, low-dose scanning, and standardized interpretation; moreover, CTC has a low complication rate and is effective for triaging positive screening cases, reducing colonoscopy volume and waiting times. Recent legal changes enabling radiologic technologists to perform CTC independently may improve access in underserved areas. With further standardization and formal inclusion in guidelines, CTC could be integrated into colorectal cancer screening programs, improving both efficiency and outcomes.

Balloon-assisted endoscopy can facilitate most ERCP-related procedures in patients with surgically altered gastrointestinal anatomy. Nevertheless, bile duct cannulation remains technically demanding in such cases owing to anatomical modifications and the limitations of forward-viewing endoscopes, such as the lack of an elevator, and tailored procedural strategies are therefore required. This article summarizes the key considerations for deep scope insertion and bile duct cannulation according to the type of reconstruction, and describes practical techniques for bile duct cannulation using a short-type double-balloon endoscope in cases of surgically altered anatomy.

Objectives: Endoscopic ultrasound-guided fine-needle aspiration/biopsy (EUS-FNA/B) is the gold standard for diagnosing gastric subepithelial lesions (SELs), but diagnosing lesions smaller than 20 mm remains challenging. We developed traction-assisted EUS-FNB (TA-EUS-FNB) using the clip-with-thread method to enhance diagnostic accuracy by stabilizing the lesion and providing counter-traction for easier needle access. This study evaluates the effectiveness of TA-EUS-FNB in diagnosing small gastric SELs.

Methods: In this prospective, randomized, controlled cross-over trial (August 2019-November 2022), 30 patients with gastric SELs ＜20 mm were randomized to undergo TA-EUS-FNB or conventional EUS-FNB. Each patient underwent four punctures, two per method. The primary end-point was the adequate tissue sampling rate for both techniques. Secondary end-points included diagnostic yield and performance (sensitivity and specificity) in distinguishing gastrointestinal stromal tumors (GISTs) from non-GISTs.

Results: The mean tumor size was 15.0 mm, with diagnoses comprising GISTs (n = 15, 50%), leiomyomas (n = 8, 26.7%), schwannomas (n = 2, 6.7%), aberrant pancreas (n = 3, 10%), and inflammation (n = 2, 6.7%). TA-EUS-FNB demonstrated a significantly higher adequate-tissue sampling rate (90% vs. 66.7%, P = 0.035) and diagnostic yield (86.7% vs. 63.3%, P = 0.037) than conventional EUS-FNB. Sensitivity (86.7%, 95% confidence interval [CI] 62.1-96.3% vs. 66.7%, 95% CI 41.7-84.8%; P = 0.20) and specificity (100%, 95% CI 79.6-100% vs. 100%, 95% CI 79.6-100%) were comparable between the methods. No adverse events were observed in the study.

Conclusion: TA-EUS-FNB demonstrated superior tissue sampling rates and diagnostic yield for SELs ＜20 mm compared to conventional EUS-FNB, making it a viable option. Controlling lesion mobility is essential for successful EUS-FNB in small SELs.

Trial registration: This study was registered in the University Hospital Medical Information Network Clinical Trials Registry (UMIN 000037494) following CONSORT guidelines.