Artificial intelligence-assisted system for the assessment of Forrest classification of peptic ulcer bleeding: a multicenter diagnostic study.

BACKGROUND: Inaccurate Forrest classification may significantly affect clinical outcomes, especially in high risk patients. Therefore, this study aimed to develop a real-time deep convolutional neural network (DCNN) system to assess the Forrest classification of peptic ulcer bleeding (PUB). METHODS: A training dataset (3868 endoscopic images) and an internal validation dataset (834 images) were retrospectively collected from the 900th Hospital, Fuzhou, China. In addition, 521 images collected from four other hospitals were used for external validation. Finally, 46 endoscopic videos were prospectively collected to assess the real-time diagnostic performance of the DCNN system, whose diagnostic performance was also prospectively compared with that of three senior and three junior endoscopists. RESULTS: The DCNN system had a satisfactory diagnostic performance in the assessment of Forrest classification, with an accuracy of 91.2% (95%CI 89.5%-92.6%) and a macro-average area under the receiver operating characteristic curve of 0.80 in the validation dataset. Moreover, the DCNN system could judge suspicious regions automatically using Forrest classification in real-time videos, with an accuracy of 92.0% (95%CI 80.8%-97.8%). The DCNN system showed more accurate and stable diagnostic performance than endoscopists in the prospective clinical comparison test. This system helped to slightly improve the diagnostic performance of senior endoscopists and considerably enhance that of junior endoscopists. CONCLUSION: The DCNN system for the assessment of the Forrest classification of PUB showed satisfactory diagnostic performance, which was slightly superior to that of senior endoscopists. It could therefore effectively assist junior endoscopists in making such diagnoses during gastroscopy.

Artificial intelligence-aided diagnostic imaging: A state-of-the-art technique in precancerous screening.

BACKGROUND AND AIM: Chromoendoscopy with the use of indigo carmine (IC) dye is a crucial endoscopic technique to identify gastrointestinal neoplasms. However, its performance is limited by the endoscopist's skill, and no standards are available for lesion identification. Thus, we developed an artificial intelligence (AI) model to replace chromoendoscopy. METHODS: This pilot study assessed the feasibility of our novel AI model in the conversion of white-light images (WLI) into virtual IC-dyed images based on a generative adversarial network. The predictions of our AI model were evaluated against the assessments of five endoscopic experts who were blinded to the purpose of this study with a staining quality rating from 1 (unacceptable) to 4 (excellent). RESULTS: The AI model successfully transformed the WLI of polyps with different morphologies and different types of lesions in the gastrointestinal tract into virtual IC-dyed images. The quality ratings of the real IC-dyed and AI images did not significantly differ concerning surface structure (AI vs IC: 3.08 vs 3.00), lesion border (3.04 vs 2.98), and overall contrast (3.14 vs 3.02) from 10 sets of images (10 AI images and 10 real IC-dyed images). Although the score depended significantly on the evaluator, the staining methods (AI or real IC) and evaluators had no significant interaction (P > 0.05) with each other. CONCLUSION: Our results demonstrated the feasibility of employing AI model's virtual IC staining, increasing the possibility of being employed in daily practice. This novel technology may facilitate gastrointestinal lesion identification in the future.

TACC diagnosed by transoesophageal endoscopic ultrasonography: A case report.

BACKGROUND: Primary adenoid cystic carcinoma in the trachea (TACC) is a rare tumour. Tracheal bronchoscopy is always chosen as a routine approach to obtain a pathological diagnosis, but it can be associated with an increased risk of asphyxia. CASE SUMMARY: We describe a case of TACC in a patient evaluated by chest computed tomography (CT) with three-dimensional reconstruction imaging and diagnosed by transoesophageal endoscopic ultrasonography. The pathological diagnosis confirmed tracheal adenoid cystic carcinoma. CONCLUSION: We highlight the importance of CT and provide a successful exploration of transoesophageal biopsy as a safe alternative approach.

Novel endoscopic tangential sawing technique in treatment of giant gastric bezoars: a retrospective single-center study (with video).

BACKGROUND AND AIMS: The current methods for treatment of giant gastric bezoars mainly include chemical dissolution, endoscopic fragmentation, and surgical removal, which often have limited curative effects or generate multiple adverse events. Therefore, there is an urgent need to find new methods to overcome such a dilemma. The aim of this study was to evaluate the safety, efficacy, and feasibility of a novel guidewire-based tangential sawing fragmentation (GTSF) technique to treat giant gastric bezoars. METHODS: This retrospective single-center study was performed in the Department of Gastroenterology and Hepatology, Sichuan Academy of Medical Sciences & Sichuan Provincial People's Hospital. Ten patients with giant bezoars were consecutively enrolled from December 8, 2019 to April 8, 2021. We treated the 10 patients with the GTSF technique, recorded the GTSF procedure, and followed the patients with gastroscopy 2 weeks after the procedure. RESULTS: All patients were successfully treated by the GTSF technique, and the giant bezoar was broken into small pieces (<2 cm in diameter). The average operation time was 21.73 minutes, and the average fragmentation time was 8.06 minutes. Ten patients treated with the GTSF technique attained satisfactory treatment results, with no acute adverse events or alimentary canal injury during the procedure, and no bezoar residue remained as shown by gastroscopy 2 weeks after the procedure. CONCLUSIONS: The GTSF technique is a safe, effective, and feasible method for removing giant bezoars and can be considered as an alternative treatment of this disease.

Risk factors for predicting early variceal rebleeding after endoscopic variceal ligation.

AIM: To analyze the clinical risk factors for early variceal rebleeding after endoscopic variceal ligation (EVL). METHODS: 342 cirrhotic patients with esophageal varices who received elective EVL to prevent bleeding or rebleeding at our endoscopy center between January 2005 and July 2010. were included in this study. The early rebleeding cases after EVL were confirmed by clinical signs or endoscopy. A case-control study was performed comparing the patients presenting with early rebleeding with those without this complication. RESULTS: The incidence of early rebleeding after EVL was 7.60%, and the morbidity of rebleeding was 26.9%. Stepwise multivariate logistic regression analysis showed that four variables were independent risk factors for early rebleeding: moderate to excessive ascites [odds ratio (OR) 62.83, 95% CI: 9.39-420.56, P < 0.001], the number of bands placed (OR 17.36, 95% CI: 4.00-75.34, P < 0.001), the extent of varices (OR 15.41, 95% CI: 2.84-83.52, P = 0.002) and prothrombin time (PT) > 18 s (OR 11.35, 95% CI: 1.93-66.70, P = 0.007). CONCLUSION: The early rebleeding rate after EVL is mainly affected by the volume of ascites, number of rubber bands used to ligate, severity of varices and prolonged PT. Effective measures for prevention and treatment should be adopted before and after EVL.

Endoscopic nasojejunal feeding tube placement in patients with severe hepatopancreatobiliary diseases: a retrospective study of 184 patients.

BACKGROUND: Total parenteral nutrition (TPN) has been recognized as the mainstay of nutritional support in patients with severe hepatopancreatobiliary (HPB) diseases for decades. However, recent studies advocate the utilization of endoscopic nasojejunal feeding tube placement (ENFTP), rather than the conventional approach. This study was designed to compare the clinical value of ENFTP and TPN in patients with severe HPB diseases. METHODS: Two groups of patients with severe HPB diseases were analyzed retrospectively. One group of 88 patients received ENFTP, and the other 96 received TPN. Routine blood levels, serum glucose and prealbumin, hepatic and renal function, serum lipid, and calcium were measured at baseline and after 1, 2, and 4 weeks of nutritional support. Also, complication rate, mortality, nutritional support time, mechanical ventilation time, mean length of time in intensive care unit, and duration of hospital stay were analyzed. RESULTS: After 4 weeks of nutritional support, the degree of recovery of red blood cells, prealbumin, and blood glucose was greater in the ENFTP than in the TPN group (P<0.05). Furthermore, the ENFTP group showed a lower incidence of septicemia, multiple organ dysfunction syndrome, peri-pancreatic infection, biliary infection, and nosocomial infection, in addition to shorter nutritional support time and hospital stay (P<0.05). CONCLUSIONS: ENFTP is much more effective than TPN in assisting patients with severe HPB diseases to recover from anemia, low prealbumin level, and high serum glucose, as well as in decreasing the rates of various infections (pulmonary infection excluded), multiple organ dysfunction syndrome rate, nutrition support time, and length of hospital stay. Therefore, ENFTP is safer and more economical for clinical application.

FAT10 level in human gastric cancer and its relation with mutant p53 level, lymph node metastasis and TNM staging.

AIM: To investigate the role of FAT10 and mutant p53 in the pathogenesis, severity and prognosis of gastric cancer. METHODS: FAT10, mutant p53 mRNA and protein levels were measured by reverse transcription (RT)-PCR and immunohistochemistry in gastric cancer tissue (n = 62), tumor-adjacent tissue (n = 62) and normal gastric tissue (n = 62). Relation of FAT10 and mutant p53 expression with clinicopathological features and clinical outcomes of gastric cancer patients were analyzed. RESULTS: The FAT10, mutant p53 mRNA and protein levels were significantly higher in gastric cancer than in its adjacent and normal tissue. The FAT10 and mutant p53 levels in gastric cancer tissue were significantly correlated with lymph node metastasis and tumor, nodes, metastasis (TNM) staging. Moreover, the high FAT10 level was associated with the overall survival rate of patients. Multivariate Cox-proportional hazards model analysis showed that mRNA and protein levels of FAT10 and mutant p53, lymph node metastasis, distant metastasis and TNM stage were the independent prognostic factors for gastric cancer. CONCLUSION: FAT10 may be involved in gastric carcinogenesis, and is a potential marker for the prognosis of gastric cancer patients. FAT10 and mutant p53 may play a common role in the carcinogenesis of gastric cancer.