Conventional Endoscopy Research Articles

Cap-assisted endoscopic treatment of esophageal food bolus impaction and/or foreign body ingestion: a systematic review and meta-analysis.

Esophageal food bolus and/or foreign body (FB) impaction is a common gastrointestinal emergency. This meta-analysis reports on the pooled outcomes of cap-assisted endoscopic removal of esophageal FB. We conducted a comprehensive search of several databases (inception to February 2022) to identify studies reporting on the use of a cap in the endoscopic treatment of esophageal FB ingestion. A random effects model was used to calculate the pooled odds ratio (OR) and mean difference (MD), and I2 values were used to assess the heterogeneity. Six studies were analyzed that included 677 patients treated with cap-assisted and 694 with conventional endoscopy. The cap-assisted method demonstrated statistically significant superiority regarding technical success (pooled OR 7.1, 95% confidence interval [CI] 1.9-26.9; P=0.004), en bloc removal (pooled OR 26.6, 95%CI 17.6-40.2; P<0.001), as well as a significantly shorter procedure time (4.6 min, 95%CI -6.5 to -2.8; P<0.001), compared to conventional methods. Better technical success was achieved with the cap-assisted method performed under anesthesia (OR 8.7, 95%CI 1.6-47.7; P=0.01); however, a shorter procedure time was noted for the cap-assisted method without anesthesia (MD -1.5, 95%CI -2.7 to -0.4; P=0.01). Pooled adverse events were comparable. Pooled OR for mucosal tear was significantly lower with cap in food bolus impaction (OR 0.07, 95%CI 0.01-0.38; P=0.02). Cap-assisted endoscopic removal of esophageal FB is associated with better technical success and en bloc removal, and a shorter procedure time compared to conventional methods, with comparable adverse events.

Endoscopic suspect of subacute occlusive mesenteric ischaemia

A 74-year-old patient presented to our emergency department with melena for 24 hours, associated with postprandial abdominal pain, predominantly in the epigastrium, for one month. Urgent gastroscopy showed a pale mucosa with loss of vascular pattern in the gastric antrum, as well as several superficial ulcers, Forrest III, at that level and in the duodenal bulb highly suggestive of ischaemia. An abdominal computed tomography angiography (CTA) revealed a filiform celiac trunk, with calcified atherosclerotic plaques in the ostium, superior mesenteric artery and both renal arteries, with absence of enhancement in a large part of the intestinal wall, suggestive of ischaemia. An exploratory laparoscopy confirmed patchy ischaemia of the small intestine, as well as of the cecum, ascending colon, and portion of the transversus. Atherosclerosis is the most common cause of occlusive chronic mesenteric ischaemia. The 'classic triad' consisting of postprandial pain, weight loss and abdominal bruit, is found in only a minority of patients, with a significant percentage of paucisymptomatic patients due to abundant colateral circulation, making diagnosis difficult. The endoscopic finding of edema, erythema or signs of mucosal atrophy, as well as gastric or duodenal ulcers, not justified by other causes can guide us in the diagnosis. However, more distal sections of the digestive tract which are not accessible with the conventional endoscope can often be affected, and the absence of these findings, does not exclude the diagnosis. Abdominal CTA is the gold-standard imaging test. Early diagnosis of mesenteric ischaemia is a challenge in clinical practice. Recognizing its endoscopic signs can facilitate its early diagnosis and treatment.

Confocal laser endomicroscopy in gastro-intestinal endoscopy: technical aspects and clinical applications.

Confocal laser endomicroscopy (CLE) is an advanced endoscopic imaging technology that provides a magnified, cellular level view of gastrointestinal epithelia. In conjunction with topical or intravenous fluorescent dyes, CLE allows for an "optical biopsy" for real-time diagnosis. Two different CLE system have been used in clinical endoscopy, probe-based CLE (pCLE) and endoscope-based CLE (eCLE). Using pCLE, the device can be delivered: (I) into the luminal gastrointestinal tract through the working channel of standard endoscopes; (II) into extraluminal cystic and solid parenchymal lesions through an endoscopic ultrasound (EUS) needle; or (III) into the biliary system through an endoscopic retrograde cholangiopancreatography (ERCP) accessory channel. With eCLE, the probe is directly integrated into the tip of a conventional endoscope, however, these endoscopes are no longer commercially available. CLE has moderate to high diagnostic accuracy for neoplastic and inflammatory conditions through the gastrointestinal tract including: oesophageal, gastric and colonic neoplasia, pancreatic cysts and solid lesions, malignant pancreatobiliary strictures and inflammatory bowel disease. Some studies have demonstrated the diagnostic benefit of CLE imaging when combined with either conventional white light endoscopy or advanced imaging technologies. Therefore, optical biopsies using CLE can resolve diagnostic dilemmas in some cases where conventional imaging fails to achieve conclusive results. CLE could also reduce the requirement for extensive tissue sampling during surveillance procedures. In the future, CLE in combination with molecular probes, could allow for the molecular characterization of diseases and assess response to targeted therapy. However, the narrow field of view, high capital costs and specialized operator training requirements remain the main limitations. Future multi-center, randomized trials with a focus on conventional diagnostic applications, cost-effectiveness and standardized training will be required for definitive evidence. The objective of this review is to evaluate the technical aspects and current applications of CLE in patients with gastrointestinal and pancreatobiliary diseases and discuss future directions for this technique.

Esophageal tuberculosis.

Esophageal tuberculosis (ET) is an extremely rare disease and herein, we present an ET case. Endoscopic findings of ET are variable and diverse and can be easily mistaken for malignancy. A definitive diagnosis of ET is difficult to make with white light endoscopy alone, and the diagnostic yield of a biopsy is low in secondary ET cases with normal overlying mucosa. Although the findings of conventional endoscopy and endoscopic ultrasonography in ET have been reported so far, few reports have described the findings of magnifying endoscopy with narrow-band imaging (ME-NBI). Dilated microvessels without irregularities on ME-NBI may be useful to differentiate secondary ET from esophageal carcinoma, since the findings suggest compression from the depth. Although rare, ET has to be considered in the differential diagnosis for any unexplained esophageal lesions.

Deep Learning for Automatic Identification and Characterization of the Bleeding Potential of Enteric Protruding Lesions in Capsule Endoscopy.

Capsule endoscopy (CE) revolutionized the study of the small intestine, overcoming the limitations of conventional endoscopy. Nevertheless, reviewing CE images is time-consuming. Convolutional Neural Networks (CNNs) are an artificial intelligence architecture with high performance levels for image analysis. Protruding lesions of the small intestine exhibit enormous morphologic diversity in CE images. We aimed to develop a CNN-based algorithm for automatic detection of varied small-bowel protruding lesions. A CNN was developed using a pool of CE images containing protruding lesions or normal mucosa/other findings. A total of 2565 patients were included. These images were inserted into a CNN model with transfer learning. We evaluated the performance of the network by calculating its sensitivity, specificity, accuracy, positive predictive value, and negative predictive value. A CNN was developed based on a total of 21,320 CE images. Training and validation data sets comprising 80% and 20% of the total pool of images, respectively, were constructed for development and testing of the network. The algorithm automatically detected small-bowel protruding lesions with an accuracy of 97.1%. Our CNN had a sensitivity, specificity, positive, and negative predictive values of 95.9%, 97.1%, 83.0%, and 95.7%, respectively. The CNN operated at a rate of approximately 355 frames per second. We developed an accurate CNN for automatic detection of enteric protruding lesions with a wide range of morphologies. The development of these tools may enhance the diagnostic efficiency of CE.

Histopathological Characteristics and Therapeutic Outcomes of Endoscopic Submucosal Dissection for Gastric High-Grade Intraepithelial Neoplasia.

Background: The endoscopic resection of suspected gastric high-grade intraepithelial neoplasia (HGIN) may incidentally cause the patient to suffer from early gastric cancer (EGC), complicating the subsequent clinical management. Identifying the risk factors for such misstaging may help guide the clinical management. Methods: The information obtained from 123,460 patients, who underwent conventional upper gastrointestinal endoscopy at the First Affiliated Hospital of Nanjing Medical University from January 2010 to December 2015, were retrospectively reviewed. Patients with an initial diagnosis of HGIN underwent endoscopic submucosal dissection (ESD), and received a final diagnosis of EGC. The risk factors for the upgraded pathology and noncurative resection were analyzed. Results: Among the 134 patients initially diagnosed with HGIN, 35 (26.12%) patients were finally diagnosed with EGC after ESD. A lesion size of ≥2 cm (odds ratio [OR] = 5.16, 95% confidence interval [CI] = 2.04-13.05; P < .01), ≤4 biopsies taken (OR = 2.73, 95% CI = 1.15-6.48; P < .05), and the presence of upper gastrointestinal bleeding (UGIB; OR = 15.64, 95% CI = 1.29-189.75; P < .05) were the independent risk factors for upgraded pathology. In addition, patients >65 years old (OR = 0.022, 95% CI = 0.901-6.549; P < .05) or with a lesion size of ≥2 cm (OR = 4.237, 95% CI = 1.650-10.878; P < .01) were more likely to endure the noncurative resection. Conclusion: For suspected gastric HGIN patients, age, lesion size, the number of biopsies, and UGIB should be taken into account before deciding on the ESD.

New concept for colonoscopy including side optics and artificial intelligence

Adenoma detection rate is the crucial parameter for colorectal cancer screening. Increasing the field of view with additional side optics has been reported to detect flat adenomas hidden behind folds. Furthermore, artificial intelligence (AI) has also recently been introduced to detect more adenomas. We therefore aimed to combine both technologies in a new prototypic colonoscopy concept. A 3-dimensional-printed cap including 2 microcameras was attached to a conventional endoscope. The prototype was applied in 8 gene-targeted pigs with mutations in the adenomatous polyposis coli gene. The first 4 animals were used to train an AI system based on the images generated by microcameras. Thereafter, the conceptual prototype for detecting adenomas was tested in a further series of 4 pigs. Using our prototype, we detected, with side optics, adenomas that might have been missed conventionally. Furthermore, the newly developed AI could detect, mark, and present adenomas visualized with side optics outside of the conventional field of view. Combining AI with side optics might help detect adenomas that otherwise might have been missed.

Utilization of Ultrasonic Image Characteristics Combined with Endoscopic Detection on the Basis of Artificial Intelligence Algorithm in Diagnosis of Early Upper Gastrointestinal Cancer.

The aim of this study was to evaluate the diagnostic value of artificial intelligence algorithm combined with ultrasound endoscopy in early esophageal cancer and precancerous lesions by comparing the examination of conventional endoscopy and artificial intelligence algorithm combined with ultrasound endoscopy, and by comparing the real-time diagnosis of endoscopy and the ultrasonic image characteristics of artificial intelligence algorithm combined with endoscopic detection and pathological results. 120 cases were selected. According to the inclusion and exclusion criteria, 80 patients who met the criteria were selected and randomly divided into two groups: endoscopic examination combined with ultrasound imaging based on intelligent algorithm processing (cascade region-convolutional neural network (Cascade RCNN) model algorithm group) and simple use of endoscopy group (control group). This study shows that the ultrasonic image of artificial intelligence algorithm is effective, and the detection performance is better than that of endoscopic detection. The results are close to the gold standard of doctor recognition, and the detection time is greatly shortened, and the recognition time is shortened by 71 frames per second. Compared with the traditional convolutional neural network (CNN) algorithm, the accuracy and recall of image analysis and segmentation using feature pyramid network are increased. The detection rates of CNN model, Cascade RCNN model, and endoscopic detection alone in early esophageal cancer and precancerous lesions are 56.3% (45/80), 88.8% (71/80), and 44.1% (35/80), respectively. The detection rate of Cascade RCNN model and CNN model was higher than that of endoscopy alone, and the difference was statistically significant (P < 0.05). The sensitivity, specificity, positive predictive value, and negative predictive value of Cascade RCNN model were higher than those of CNN model, which was close to the gold standard for physician identification. This provided a reference basis for endoscopic ultrasound identification of early upper gastrointestinal cancer or other gastrointestinal cancers.

Endoscopic and histopathological findings in adult dyspeptic patients, and their association with Helicobacter pylori infection in Dhaka, Bangladesh

Conventional upper-intestinal endoscopy is usually performed to diagnose Helicobacter pylori (H. pylori) associated diseases, using gastric mucosa from the biopsy. The objective of our study was to identify the prevalence of H. pylori and its relation with endoscopic findings and histopathological features in dyspeptic adult patients. Gastroduodenal biopsy specimens were collected from 143 adult dyspeptic outpatients who attended the Department of Gastroenterology, Bangabandhu Sheikh Mujib Medical University (BSMMU) and Dhaka Medical College Hospital (DMCH), for endoscopy. H. pylori was identified by rapid urease test (RUT), ureC gene PCR, and histological staining (Giemsa). The study population was divided into H. pylori-positive cases (47; 32.9%) and H. pylori-negative cases (96; 67.1%), based on the case definitions used in the study. The highest rate of H. pylori infection was found in the 41-50 years age group (25.5%). Endoscopically, 101 (97.1%) dyspeptic patients had gastritis, with the majority of H. pylori infections found among histopathologically diagnosed duodenal ulcer patients. Endoscopic findings were significantly correlated with histological findings (p < 0.001). Significant correlations between endoscopic and histopathological findings were observed. Early detection and prompt treatment of H. pylori infection are essential for the prevention of serious complications.

Magnifying endoscopy for the diagnosis of early gastric cancer: Establishment of technique, diagnostic system, and scientific evidence from Japan

Magnifying endoscopy for the diagnosis of early gastric cancer: Establishment of technique, diagnostic system, and scientific evidence from Japan

Systematic Endoscopic Approach to Early Gastric Cancer in Clinical Practice.

Early gastric cancers (EGCs) are defined as gastric cancers confined to the mucosa or submucosa, regardless of regional lymph node metastasis. The proportion of EGCs has been increasing due to the increase in screening endoscopy for gastric cancers; therefore, the paradigm shift from surgical resection to endoscopic resection as a treatment modality for selected EGCs is accelerating. For successful endoscopic resection of EGCs, it is important to detect EGCs at an early stage and to accurately predict the histological type, depth of invasion, and horizontal margins of the tumor. The diagnostic process of EGCs can be divided into three steps: presence diagnosis, qualitative diagnosis, and quantitative diagnosis. The presence diagnosis of EGCs is mainly based on two endoscopic findings: a well-demarcated lesion and irregularity in the color/surface pattern. Qualitative diagnosis refers to the prediction of histological type, which is mainly possible based on the macroscopic shape and color of the lesion. Quantitative diagnosis of EGCs consists of predicting the depth of invasion by detailed examination of the macroscopic morphology and determining horizontal margins using chromoendoscopy. Although advanced diagnostic modalities, such as endosonography or magnifying endoscopy, are helpful for the qualitative and quantitative diagnosis of EGCs, these modalities are not available in most hospitals. Therefore, it is still very important to evaluate EGCs systematically during conventional endoscopy for successful endoscopic treatment.

The feasibility and safety of disposable endoscope vs. conventional endoscope for upper gastrointestinal tract examination: a multicenter, randomized, parallel, non-inferiority trial.

Background A disposable upper gastrointestinal endoscope can effectively decrease infectious outbreaks associated with endoscope reuse. In the present study, we aimed to evaluate the feasibility and safety of a disposable endoscope for upper gastrointestinal examination. Methods In a prospective, randomized trial, 144 upper endoscopic procedures were allocated to either the disposable endoscope group or the conventional endoscope group. The primary outcomes were rates of excellent and good image qualities and maneuverability satisfaction. The second outcome included procedure duration, endoscopic diagnosis, and adverse events. Results A total of 144 subjects were enrolled in the present analysis and prospectively randomized to 2 study groups. Finally, 70 and 69 subjects were enrolled in the novel disposable endoscope group and the conventional endoscope group, respectively, due to the schedule cancellation of 5 subjects. The baseline characteristics of the patients were similar in both groups. The excellent and good image quality rates and maneuverability satisfaction of the novel disposable endoscope were not inferior to the conventional endoscope (p = 0.99 and p = 0.99, respectively). Moreover, no significant between-group difference was observed in the endoscopic results and adverse events (p = 0.30 and p = 1, respectively). However, the procedure duration in the novel disposable endoscope was longer compared with the conventional endoscope (8.40 ± 4.28 min vs. 5.12 ± 2.65 min, p < 0.001). Conclusions The novel disposable endoscope was as safe, effective, and maneuverable as a conventional endoscope. However, the novel disposable endoscope was associated with a longer procedure duration.

Improving tissue characterization, differentiation and diagnosis in gynecology with the narrow-band imaging technique: A systematic review.

Narrow-band imaging (NBI), an on-demand, real-time endoscopic imaging technique, was developed to enhance visualization of the mucosal vascular network and surface texture. The present article provides a systematic review of studies that assessed the use of NBI in gynecological endoscopy. The following electronic databases were searched: PubMed (1950-2020), Google Scholar (2004-2020) and Cochrane Library (2010-2020). In the initial search, 3,836 entries were identified, of which 31 were finally included in the systematic review. Of the selected studies, 10 (32%) were case reports, 19 (61.2%) were prospective studies and 2 (6.4%) were randomized controlled trials with control groups. The selected studies reported on the use of NBI in hysteroscopy, laparoscopy and colposcopy. It was revealed that NBI utilization in hysteroscopy increased the accuracy, sensitivity and specificity in detecting malignant and premalignant lesions. NBI improved the specificity and sensitivity in the detection of endometriotic lesions and cervical lesions. Conventional white light endoscopy in gynecology may be significantly improved by the use of NBI. Further studies with larger cohorts and improved design are required to achieve more reliable results. It is of special interest that utilization of this method requires apparatus which is expensive; concerns are the long training and experience of staff required and the long learning curve.

Transoral Endoscopic Examination of the Oropharynx With Tongue Protrusion, Phonation, and Open Mouth.

We examined the diagnostic performance of the tongue protrusion with phonation and open mouth (TOPPOM) method for visualizing structures of the oropharynx. Transoral endoscopy was performed on 20 healthy participants to evaluate 12 oropharynx subsites under three conditions: open mouth (OM), phonation with open mouth (POM), and TOPPOM. Each subsite was scored from 0 to 2 depending on subsite visualization, and the scores were summed. Images of subsite-adjacent mucosa were similarly scored. The total scores were significantly higher for TOPPOM than for POM and for POM than for OM. Such scores were observed for both the palatine arches, both palatine tonsils, the left lingual tonsillar sulcus, and the vallecula. TOPPOM enables visualization of the oropharynx through transoral endoscopic examination, and TOPPOM with conventional transnasal endoscopy may enable early detection of oropharyngeal carcinomas and lesions and improve the performance of pre- and post-treatment evaluations.

Small bowel to closest human body surface distance calculation through a custom-made software using CT-based datasets.

Screening of the gastrointestinal tract is imperative for the detection and treatment of physiological and pathological disorders in humans. Ingestible devices (e.g., magnetic capsule endoscopes) represent an alternative to conventional flexible endoscopy for reducing the invasiveness of the procedure and the related patient's discomforts. However, to properly design localization and navigation strategies for capsule endoscopes, the knowledge of anatomical features is paramount. Therefore, authors developed a semi-automatic software for measuring the distance between the small bowel and the closest human external body surface, using CT colonography images. In this study, volumetric datasets of 30 patients were processed by gastrointestinal endoscopists with the dedicated custom-made software and results showed an average distance of 79.29 ± 23.85 mm.

Feature extraction for abnormality detection in capsule endoscopy images

Capsule endoscopy is a non-invasive method for diagnosing gastrointestinal diseases. This new technology has many advantages over conventional endoscopy. However, investigating endoscopic video frames in search of diseases is a tedious task for physicians. Hence, a system is required to automatically detect suspicious frames for further medical examination. Different abnormalities may exist in capsule endoscopy images. In this paper, a novel method is proposed to investigate capsule endoscopy images for abnormalities such as bleeding and angiodysplasia lesions. The proposed method identifies potential regions of interest by using an expectation maximization based image segmentation algorithm, and extracts features from them using a combination of color histogram analysis and statistical features to classify frames into normal and abnormal classes. The results show that the proposed method can distinguish the two objective classes with an approximate precision and recall of 96.5% and 95.9%, respectively.

S2345 A Rare Case of Bleeding Ulceration From the Excluded Stomach After Roux-en-Y Gastric Bypass Surgery

Introduction: Roux-en-Y gastric bypass (RYGB) is among the most common bariatric surgeries for treatment of morbid obesity, with >450,000 RYGB performed in the United States since 2011. Formation of an excluded stomach and duodenum that are difficult to access via conventional endoscopy can pose a unique challenge for diagnosis and treatment of gastrointestinal (GI) bleeding in post-RYGB patients. Case Description/Methods: A 60-year-old man with a history of morbid obesity and status post RYGB in 2017 presented in January 2021 with black tarry stools and a hemoglobin of 7.4 from a baseline of 13 g/dL without recent NSAID or tobacco use. He underwent upper endoscopy, colonoscopy, push enteroscopy and video capsule endoscopy without identification of a bleeding source. He was ultimately discharged home after stabilization and no further bleeding. Due to concern for an unidentified source of bleeding, further outpatient workup was pursued. CT enterography was unrevealing. Attempted endoscopic ultrasound (EUS)-guided gastrotomy to access the excluded stomach was unsuccessful due to inability to pass an Axios stent into the excluded stomach. Antegrade balloon enteroscopy failed to reach the jejuno-jejunal anastomosis. The patient was re-admitted in April 2021 for recurrent melena and a hemoglobin of 8.3 g/dL. Laparoscopic gastrotomy was performed with intraoperative transgastric endoscopy which revealed blood in the excluded stomach and duodenum. An ulcer in the excluded stomach was discovered and excised via subtotal gastrectomy. Pathology revealed focal transmural ulceration without dysplasia; H. pylori testing was negative. The patient was discharged home without recurrence of bleeding. Discussion: Alteration in GI anatomy after RYGB impedes access to excluded segments of the upper GI tract. Access to the excluded segments may be achieved by balloon enteroscopy depending on the length of the alimentary limb, EUS-assisted gastrotomy or laparoscopic formation of a temporary gastrotomy with transgastric endoscopy. A review of post-RYGB cases of excluded segment ulceration found that 93% of patients required surgery for definitive treatment. The most common interventions were remnant gastrectomy and oversewing of perforated ulcers with a 94% cure rate at short-term follow-up. Our case demonstrates that a high degree of suspicion for bleeding from the excluded gastric or small bowel segments should be maintained in patients with RYGB presenting with upper GI bleeding and initial negative GI tract evaluation.Figure 1.: Transmural ulceration in the remnant stomach status post subtotal gastrectomy

Endoscopic Ultrasound versus Conventional Ultrasonography as A Diagnostic Tool for Patients with Obstructive Jaundice

Abstract Background Patient presented with jaundice, initial evaluation should be ultrasonography for evaluation of liver parenchymal changes, cirrhosis and assessment of hepatic vascularity and evaluation of intrahepatic, extrahepatic biliary system. Aim of the Work to compare between the Conventional US and Endoscopic Ultrasound in diagnosis of obstructive jaundice and other diagnostic modality including CT/MRI, ERCP for diagnosis and evaluation of the cause of OJ. Patients and Methods This study was conducted on 60 patients with obstructed jaundice as evident by ultrasonography in the form of dilated CBD with IHBRDs, with visualization of Gall bladder content and pancreas if can be well visualized, all patients had clinically apparent jaundice with disturbance of liver biochemical profile, all patient underwent EUS and compare finding with US including CBD diameter and IHBRDs degree. Results EUS provide better visualization of CBD and IHBRDs with diagnosis of pathological lesion including stones, malignancy. EUS also has better visualization of pancreatic pathology including inflammation and malignancy and extent of the tumor including local extent, lymphatic spread with vascular invasion. Conclusion EUS is better than US and other modality CT&amp;MRI and ERCP in diagnosis of distal CBD stricture including malignant obstruction especially in cases of early malignancy and small tumors, and calculary distal obstruction. CT is insensitive in diagnosis of periampullary cancer, EUS provide early diagnosis of Pancreaticobiliary malignancy and give better advandage for early resectability .

S558 Artificial Intelligence and Capsule Endoscopy: Automatic Detection of Gastric Vascular Lesions Using a Convolutional Neural Network

Introduction: Capsule endoscopy (CE) is the main diagnostic modality for the investigation of obscure gastrointestinal bleeding. Although it is designed to identify small bowel pathology, lesions (particularly vascular lesions) are frequently documented in locations within reach of conventional endoscopy. However, the performance of CE for detection of gastric lesions has been shown to be suboptimal. Studies on the development of artificial intelligence (AI) systems, particularly convolutional Neural Networks (CNN), for automatic analysis of CE images have provided promising results. To date, performance of these automated algorithms for detection of gastric vascular lesions in CE images has not been evaluated. We aimed to develop and test a CNN-based algorithm for automatic detection of gastric vascular lesions (angiectasia, gastric antral vascular ectasia, varices, and red spots). Methods: A total of 1275 CE images were included for construction of the CNN, 470 containing vascular lesions and 805 showing normal mucosa. For automatic detection of gastric lesions, these images were inserted into a CNN model with transfer learning. A training dataset comprising 80% of the total pool of images was defined. Subsequently, we evaluated the performance of the network using an independent test dataset (20% of total image pool). The output provided by the CNN was compared to a consensus classification provided by two endoscopists experienced in CE (more than 1000 CE videos each). We calculated the sensitivity, specificity, accuracy, positive predictive and negative predictive values (PPV and NPV, respectively), and area under the curve (AUC). Results: After optimization of the neural architecture of the algorithm, our model was able to detect detected gastric vascular lesions with a sensitivity, specificity, PPV and NPV of 86.2%, 98.1%, 96.4% and 92.4%, respectively. The algorithm had an overall accuracy of 93.7%. The AUC was 0.97. The CNN read the validation dataset in 6 seconds (average processing speed of 45 images per second). Conclusion: This is the first report of a deep learning system for automatic detection of gastric vascular lesions in CE images. The implementation of these systems may potentiate the ability of CE for exploration of suspected upper gastrointestinal disease, which may be particularly helpful in patients with limited tolerance to EGD.

S3290 Thermographic Endoscopy for Dynamic Thermal Analysis of Gastrointestinal Disease

Introduction: To date the temperature has not been recognized an important factor in endoscopic fields, though it is physical finding to evaluate the disease. The medical common sense about thermic homeostasis interrupted notifying the availability of temperature measurement for digestive inflammation. Recently the real time monitoring by thermography in esophagus has been reported to prevent patients from burning during the cardiac ablation therapy to treat atrial fibrillation. The thermography has been used for evaluating the Raynaud' s symptom too. The portable thermography has been introduced into military uses or detection of COVID-19 patients in these days. Methods: 1) Thermographic images were obtained by a thermographic unit attached to a conventional colonoscope. It was inserted over the scope style. The protocol was approved by IRB in my hospital. 2) The prototype thermographic unit was composed of 3 parts (optical lens, image sensor (Seek Thermal or FLIR;USA), microprocessor and USB). 3) In the pilot study, the control group consisted of 4 subjects without inflammation and disease group consisted of 4 patients (2 with ulcerative colitis and 2 with polyp or early cancer) were examined. 4) Transient thermal response (TTR) was evaluated by measuring mucosal temperature changes just before and after spraying cool water. Results: 1) The temperature resolution of thermographic endoscope was 0.1 °C and the accuracy was 0.5°C. The measuring time was 0.4 second enough short to evaluate TTR. 2)The colonic temperatures in the control group were as stable as data measured by thermistor thermometer through the forceps channel (DDW2021). On the contrary, that in UC was 0.5∼1.0°C higher in the area with inflammation than in normal mucosa. 3) The increases of the mucosal temperature in UC patients were correlated to Mayor's endoscopic sub-score. 4) On the TTR, temperature recovery after spraying cool water was significantly delayed parts with inflammation in a UC patient compared to that of control group. 6) The hypervascular mucosa in tumor had shorter recovery time than normal mucosa on the TTR analysis. This mechanism enabled the thermographic endoscope to discriminate the lesion. Conclusion: The temperature and image resolution of thermography should be improved in future, thogh they are not adquate compared to the present conventional endoscope. The thermographic endoscope equipped the TTR function has the possibility to clarify the grade of inflammation and to figure out the tumor..