Proton pump inhibitors induced fungal dysbiosis in patients with gastroesophageal reflux disease.

Gut mycobiota inhabits human gastrointestinal lumen and plays a role in human health and disease. We investigated the influence of proton pump inhibitors (PPIs) on gastric mucosal and fecal mycobiota in patients with gastroesophageal reflux diseases (GERD) by using Internal Transcribed Spacer 1 sequencing. A total of 65 participants were included, consisting of the healthy control (HC) group, GERD patients who did not use PPIs (nt-GERD), and GERD patients who used PPIs, which were further divided into short-term (s-PPI) and long-term PPI user (l-PPI) groups based on the duration of PPI use. The alpha diversity and beta diversity of gastric mucosal mycobiota in GERD patients with PPI use were significantly different from HCs, but there were no differences between s-PPI and l-PPI groups. LEfSe analysis identified Candida at the genus level as a biomarker for the s-PPI group when compared to the nt-GERD group. Meanwhile, Candida, Nothojafnea, Rhizodermea, Ambispora, and Saccharicola were more abundant in the l-PPI group than in the nt-GERD group. Furthermore, colonization of Candida in gastric mucosa was significantly increased after PPI treatment. However, there was no significant difference in Candida colonization between patients with endoscopic esophageal mucosal breaks and those without. There were significant differences in the fecal mycobiota composition between HCs and GERD patients regardless whether or not they used PPI. As compared to nt-GERD patient samples, there was a high abundance of Alternaria, Aspergillus, Mycenella, Exserohilum, and Clitopilus in the s-PPI group. In addition, there was a significantly higher abundance of Alternaria, Aspergillus, Podospora, Phallus, and Monographella in the l-PPI group than nt-GERD patients. In conclusion, our study indicates that dysbiosis of mycobiota was presented in GERD patients in both gastric mucosal and fecal mycobiota. PPI treatment may increase the colonization of Candida in the gastric mucosa in GERD patients.

Novel deep learning-based computer-aided diagnosis system for predicting inflammatory activity in ulcerative colitis.

BACKGROUND AND AIMS: Endoscopy is increasingly performed for evaluating patients with ulcerative colitis (UC). However, its diagnostic accuracy is largely affected by the subjectivity of endoscopists' experience and scoring methods, and scoring of selected endoscopic images cannot reflect the inflammation of the entire intestine. We aimed to develop an automatic scoring system using deep-learning technology for consistent and objective scoring of endoscopic images and full-length endoscopic videos of patients with UC. METHODS: We collected 5875 endoscopic images and 20 full-length videos from 332 patients with UC who underwent colonoscopy between January 2017 and March 2021. We trained the artificial intelligence (AI) scoring system using these images, which was then used for full-length video scoring. To more accurately assess and visualize the full-length intestinal inflammation, we divided the large intestine into a fixed number of "areas" (cecum, 20; transverse colon, 20; descending colon, 20; sigmoid colon, 15; rectum, 10). The scoring system automatically scored inflammatory severity of 85 areas from every video and generated a visualized result of full-length intestinal inflammatory activity. RESULTS: Compared with endoscopist scoring, the trained convolutional neural network achieved 86.54% accuracy in the Mayo-scored task, whereas the kappa coefficient was .813 (95% confidence interval [CI], .782-.844). The metrics of the Ulcerative Colitis Endoscopic Index of Severity-scored task were encouraging, with accuracies of 90.7%, 84.6%, and 77.7% and kappa coefficients of .822 (95% CI, .788-.855), .784 (95% CI, .744-.823), and .702 (95% CI, .612-.793) for vascular pattern, erosions and ulcers, and bleeding, respectively. The AI scoring system predicted each bowel segment's score and displayed distribution of inflammatory activity in the entire large intestine using a 2-dimensional colorized image. CONCLUSIONS: We established a novel deep learning-based scoring system to evaluate endoscopic images from patients with UC, which can also accurately describe the severity and distribution of inflammatory activity through full-length intestinal endoscopic videos.

Artificial Intelligence-Assisted Colonoscopy for Colorectal Cancer Screening: A Multicenter Randomized Controlled Trial.

BACKGROUND AND AIMS: Artificial intelligence (AI)-assisted colonoscopy improves polyp detection and characterization in colonoscopy. However, data from large-scale multicenter randomized controlled trials (RCT) in an asymptomatic population are lacking. METHODS: This multicenter RCT aimed to compare AI-assisted colonoscopy with conventional colonoscopy for adenoma detection in an asymptomatic population. Asymptomatic subjects 45-75 years of age undergoing colorectal cancer screening by direct colonoscopy or fecal immunochemical test were recruited in 6 referral centers in Hong Kong, Jilin, Inner Mongolia, Xiamen, and Beijing. In the AI-assisted colonoscopy, an AI polyp detection system (Eagle-Eye) with real-time notification on the same monitor of the endoscopy system was used. The primary outcome was overall adenoma detection rate (ADR). Secondary outcomes were mean number of adenomas per colonoscopy, ADR according to endoscopist's experience, and colonoscopy withdrawal time. This study received Institutional Review Board approval (CRE-2019.393). RESULTS: From November 2019 to August 2021, 3059 subjects were randomized to AI-assisted colonoscopy (n = 1519) and conventional colonoscopy (n = 1540). Baseline characteristics and bowel preparation quality between the 2 groups were similar. The overall ADR (39.9% vs 32.4%; P < .001), advanced ADR (6.6% vs 4.9%; P = .041), ADR of expert (42.3% vs 32.8%; P < .001) and nonexpert endoscopists (37.5% vs 32.1%; P = .023), and adenomas per colonoscopy (0.59 ± 0.97 vs 0.45 ± 0.81; P < .001) were all significantly higher in the AI-assisted colonoscopy. The median withdrawal time (8.3 minutes vs 7.8 minutes; P = .004) was slightly longer in the AI-assisted colonoscopy group. CONCLUSIONS: In this multicenter RCT in asymptomatic patients, AI-assisted colonoscopy improved overall ADR, advanced ADR, and ADR of both expert and nonexpert attending endoscopists. (ClinicalTrials.gov, Number: NCT04422548).

A Novel Convolutional Neural Network Model as an Alternative Approach to Bowel Preparation Evaluation Before Colonoscopy in the COVID-19 Era: A Multicenter, Single-Blinded, Randomized Study.

INTRODUCTION: Adequate bowel preparation is key to a successful colonoscopy, which is necessary for detecting adenomas and preventing colorectal cancer. We developed an artificial intelligence (AI) platform using a convolutional neural network (CNN) model (AI-CNN model) to evaluate the quality of bowel preparation before colonoscopy. METHODS: This was a colonoscopist-blinded, randomized study. Enrolled patients were randomized into an experimental group, in which our AI-CNN model was used to evaluate the quality of bowel preparation (AI-CNN group), or a control group, which performed self-evaluation per routine practice (control group). The primary outcome was the consistency (homogeneity) between the results of the 2 methods. The secondary outcomes included the quality of bowel preparation according to the Boston Bowel Preparation Scale (BBPS), polyp detection rate, and adenoma detection rate. RESULTS: A total of 1,434 patients were enrolled (AI-CNN, n = 730; control, n = 704). No significant difference was observed between the evaluation results ("pass" or "not pass") of the groups in the adequacy of bowel preparation as represented by BBPS scores. The mean BBPS scores, polyp detection rate, and adenoma detection rate were similar between the groups. These results indicated that the AI-CNN model and routine practice were generally consistent in the evaluation of bowel preparation quality. However, the mean BBPS score of patients with "pass" results were significantly higher in the AI-CNN group than in the control group, indicating that the AI-CNN model may further improve the quality of bowel preparation in patients exhibiting adequate bowel preparation. DISCUSSION: The novel AI-CNN model, which demonstrated comparable outcomes to the routine practice, may serve as an alternative approach for evaluating bowel preparation quality before colonoscopy.

[Value of Blood Urea Nitrogen-to-Creatinine Ratio for Differentiating the Site of Gastrointestinal Bleeding and Assessing the Prognosis of Patients with Acute Nonvariceal Upper Gastrointestinal Bleeding].

Objective: To explore the diagnostic performance of blood urea nitrogen-to-creatinine (BUN/Cr) ratio in differentiating the site of gastrointestinal bleeding, and to assess the predictive value of early elevated BUN/Cr ratio for clinical outcomes in patients with acute nonvariceal upper gastrointestinal bleeding (ANVUGIB). Methods: The adult patients diagnosed with gastrointestinal bleeding who were hospitalized in the Department of Gastroenterology, Zhongshan Hospital, Xiamen University between May 2020 and May 2021 were retrospectively enrolled. According to the site of gastrointestinal bleeding, the patients were divided into the upper gastrointestinal tract group, the proximal small intestinal bleeding group, and the distal small intestinal and colonic bleeding group. According to the early dynamic changes of BUN/Cr ratio within 6-48 hours after admission, patients with ANVUGIB were divided into early dynamic elevated BUN/Cr ratio group and non-early dynamic elevated BUN/Cr ratio group. Receiver operating characteristic (ROC) curve was used to analyze the diagnostic performance of BUN/Cr ratio in differentiating the site of gastrointestinal bleeding and examine the predictive efficacy of early dynamic elevated BUN/Cr ratio after admission, Rockall scoring system, and the combined indicator of the two for estimating the primary clinical outcomes in ANVUGIB patients. Results: A total of 266 patients were enrolled. Among them, 204 cases were in the upper gastrointestinal bleeding group, 15 cases were in the proximal small intestinal bleeding group, and 47 cases were in the distal small intestinal and colonic bleeding group. In the ANVUGIB patients, 16 were in the group with early dynamic elevated BUN/Cr ratio after admission, and 146 were in the group with non-early dynamic elevated BUN/Cr ratio after admission. The area under the ROC curve of the BUN/Cr ratio was 0.831 (95% CI: 0.780-0.874), the optimal cut-off value being 34.59 mg/g for differentiation between upper and lower gastrointestinal bleeding. The area under the ROC curve of the BUN/Cr ratio was 0.901 (95% CI: 0.798-0.963) and the optimal cut-off value was 19.27 mg/g for differentiation between proximal small intestinal bleeding and the distal small intestinal and colonic bleeding. The area under the ROC curve of the early dynamic elevated BUN/Cr ratio after admission was 0.806 (95% CI: 0.737-0.864) for predicting the primary clinical outcome in patients with ANVUGIB. The area under the ROC curve of the combined indicator included the early dynamic elevated BUN/Cr ratio after admission and the Rockall scoring system was 0.909 (95% CI: 0.854-0.949) for predicting the primary clinical outcome in patients with ANVUGIB. Conclusion: The BUN/Cr ratio shows rather reliable diagnostic performance for identifying the site of gastrointestinal bleeding, and the early dynamic elevated BUN/Cr ratio after admission is a reliable indicator for predicting clinical outcomes in patients with ANVUGIB.

Real-time landmark detection for precise endoscopic submucosal dissection via shape-aware relation network.

We propose a novel shape-aware relation network for accurate and real-time landmark detection in endoscopic submucosal dissection (ESD) surgery. This task is of great clinical significance but extremely challenging due to bleeding, lighting reflection, and motion blur in the complicated surgical environment. Compared with existing solutions, which either neglect geometric relationships among targeting objects or capture the relationships by using complicated aggregation schemes, the proposed network is capable of achieving satisfactory accuracy while maintaining real-time performance by taking full advantage of the spatial relations among landmarks. We first devise an algorithm to automatically generate relation keypoint heatmaps, which are able to intuitively represent the prior knowledge of spatial relations among landmarks without using any extra manual annotation efforts. We then develop two complementary regularization schemes to progressively incorporate the prior knowledge into the training process. While one scheme introduces pixel-level regularization by multi-task learning, the other integrates global-level regularization by harnessing a newly designed grouped consistency evaluator, which adds relation constraints to the proposed network in an adversarial manner. Both schemes are beneficial to the model in training, and can be readily unloaded in inference to achieve real-time detection. We establish a large in-house dataset of ESD surgery for esophageal cancer to validate the effectiveness of our proposed method. Extensive experimental results demonstrate that our approach outperforms state-of-the-art methods in terms of accuracy and efficiency, achieving better detection results faster. Promising results on two downstream applications further corroborate the great potential of our method in ESD clinical practice.

Appendectomy Is Associated With Alteration of Human Gut Bacterial and Fungal Communities.

Recent research has revealed the importance of the appendix in regulating the intestinal microbiota and mucosal immunity. However, the changes that occur in human gut microbial communities after appendectomy have never been analyzed. We assessed the alterations in gut bacterial and fungal populations associated with a history of appendectomy. In this cross-sectional study, we investigated the association between appendectomy and the gut microbiome using 16S and ITS2 sequencing on fecal samples from 30 healthy individuals with prior appendectomy (HwA) and 30 healthy individuals without appendectomy (HwoA). Analysis showed that the gut bacterial composition of samples from HwA was less diverse than that of samples from HwoA and had a lower abundance of Roseburia, Barnesiella, Butyricicoccus, Odoribacter, and Butyricimonas species, most of which were short-chain fatty acids-producing microbes. The HwA subgroup analysis indicated a trend toward restoration of the HwoA bacterial microbiome over time after appendectomy. HwA had higher gut fungi composition and diversity than HwoA, even 5 years after appendectomy. Compared with those in samples from HwoA, the abundance correlation networks in samples from HwA displayed more complex fungal-fungal and fungal-bacterial community interactions. This study revealed a marked impact of appendectomy on gut bacteria and fungi, which was particularly durable for fungi.

Early neoplasia identification in Barrett's esophagus via attentive hierarchical aggregation and self-distillation.

Automatic surveillance of early neoplasia in Barrett's esophagus (BE) is of great significance for improving the survival rate of esophageal cancer. It remains, however, a challenging task due to (1) the large variation of early neoplasia, (2) the existence of hard mimics, (3) the complicated anatomical and lighting environment in endoscopic images, and (4) the intrinsic real-time requirement of this application. We propose a novel end-to-end network equipped with an attentive hierarchical aggregation module and a self-distillation mechanism to comprehensively address these challenges. The hierarchical aggregation module is proposed to capture the complementariness of adjacent layers and hence strengthen the representation capability of each aggregated feature. Meanwhile, an attention mask is developed to selectively integrate the logits of each feature, which not only improves the prediction accuracy but also enhances the prediction interpretability. Furthermore, an efficient self-distillation mechanism is implemented based on a teacher-student architecture, where the student aims at capturing abstract high-level features while the teacher is applied to bring more low-level semantic details to calibrate the classification results. The proposed techniques are effective yet lightweight, improving the classification performance without sacrificing time performance, and thus achieving real-time inference. We extensively evaluate the proposed method on the MICCAI EndoVis Challenge Dataset. Experimental results demonstrate the proposed method can achieve competitive accuracy with a much faster speed than state-of-the-arts.

Linked Color Imaging Can Improve Detection Rate of Early Gastric Cancer in a High-Risk Population: A Multi-Center Randomized Controlled Clinical Trial.

BACKGROUND: Early diagnosis of gastric cancer is difficult in China due to the lack of a valid method for endoscopic screening. Early gastric cancer, especially flat gastric cancer, lacks specific endoscopic features. Many cases appear to be similar to ordinary gastritis cases under normal white light endoscopy, which can lead to misdiagnosis. AIMS: In order to find a new method to improve detection rate of early gastric cancer in China, we designed a trial to validate linked color imaging (LCI) for screening of early gastric cancer in a high-risk population, as compared to white light imaging (WLI). METHOD: Subjects were randomly allocated to either the LCI + WLI or WLI group and then subjected to gastroscopy and all endoscopies were made after special preparation. All endoscopists had knowledge of this experiment. The main indicator was the rate of detection of gastric neoplastic lesions. The difference in the detection rate between the two groups is reported. RESULTS: The detection rate was 4.31% in the WLI group and 8.01% in the LCI + WLI group. This is a difference of 3.70% with a P value < 0.001 and an OR (95% CI) of 1.934 (1.362, 2.746). The lower limit of the 95% CI was greater than 0, and the superiority margin was 1%. CONCLUSION: The detection rate of gastric neoplastic lesions was higher in the LCI + WLI group than in the WLI group, LCI might be an effective method for screening early gastric cancer.

Effects of Proton Pump Inhibitors on the Gastrointestinal Microbiota in Gastroesophageal Reflux Disease.

Proton pump inhibitors (PPIs) are commonly used to lessen symptoms in patients with gastroesophageal reflux disease (GERD). However, the effects of PPI therapy on the gastrointestinal microbiota in GERD patients remain unclear. We examined the association between the PPI usage and the microbiota present in gastric mucosal and fecal samples from GERD patients and healthy controls (HCs) using 16S rRNA gene sequencing. GERD patients taking PPIs were further divided into short-term and long-term PPI user groups. We showed that PPI administration lowered the relative bacterial diversity of the gastric microbiota in GERD patients. Compared to the non-PPI-user and HC groups, higher abundances of Planococcaceae, Oxalobacteraceae, and Sphingomonadaceae were found in the gastric microbiota from the PPI-user group. In addition, the Methylophilus genus was more highly abundant in the long-term PPI user group than in the short-term PPI-user group. Despite the absence of differences in alpha diversity, there were significant differences in the fecal bacterial composition of between GERD patients taking PPIs and those not taking PPIs. There was a higher abundance of Streptococcaceae, Veillonellaceae, Acidaminococcaceae, Micrococcaceae, and Flavobacteriaceae present in the fecal microbiota from the PPI-user group than those from the non-PPI-user and HC groups. Additionally, a significantly higher abundance of Ruminococcus was found in GERD patients on long-term PPI medication than that on short-term PPI medication. Our study indicates that PPI administration in patients with GERD has a significant effect on the abundance and structure of the gastric mucosal microbiota but only on the composition of the fecal microbiota.