Diagnosis Of Early Gastric Cancer Research Articles

S1599 Diagnostic Performance of Linked Color Image (LCI) and Blue Laser Imaging (BLI) for the Diagnosis of Early Gastric Cancer: A Systematic Review and Meta-Analysis

S1599 Diagnostic Performance of Linked Color Image (LCI) and Blue Laser Imaging (BLI) for the Diagnosis of Early Gastric Cancer: A Systematic Review and Meta-Analysis

데이터 선별을 통한 EfficientNetV2-L 기반 조기 위암 컴퓨터 보조 진단 시스템 연구

Gastric cancer is a common cancer worldwide, especially in Korea. Early diagnosis is very important to increase the full recovery rate. However, early gastric cancer has no special symptoms and is a disease that even experts find difficult to diagnose in gastroscopy. Therefore, in this paper proposed a computer-aided diagnosis(CADx) for early gastric cancer diagnosis using EfficientNetV2-L. Due to the nature of medical data, it is difficult to collect a large amount of data. The data used for training was augmented using Cifar10 policy of the Google

Deep learning for classification and localization of early gastric cancer in endoscopic images

Gastric cancer, as a malignant tumor, is one of the most common cancer-related deaths worldwide with high mortality and incidence rates. Therefore, the endoscopic detection of gastric cancer at an early stage is essential. In this paper, we propose an automatic diagnosis method of early gastric cancer (EGC) based on deep learning (DL) techniques. Specifically, with the new annotated endoscopic image dataset collected from a single-center, this paper designs several DL architectures to realize the automatic analysis of EGC images. Particularly, a guided-attention deep network was introduced, derived from ResNet-50, for the accurate score prediction of EGC and the extraction of feature information. Furthermore, we combined a lightweight attention module and multi-scale feature extractor with U-Net for estimating the pixel-level segmentation of EGC pathological regions. Experiment results on the presented dataset showed outstanding performance in the involved classification and segmentation assignments with an accuracy of 98.84% and an intersection over union (IOU) of 0.64, revealing the potential applications of DL in aiding and improving EGC diagnosis using endoscopic images.

Primary Investigation of Deep Learning Models for Japanese "Group Classification" of Whole-Slide Images of Gastric Endoscopic Biopsy.

Background Accurate pathological diagnosis of gastric endoscopic biopsy could greatly improve the opportunity of early diagnosis and treatment of gastric cancer. The Japanese “Group classification” of gastric biopsy corresponds well with the endoscopic diagnostic system and can guide clinical treatment. However, severe shortage of pathologists and their heavy workload limit the diagnostic accuracy. This study presents the first attempt to investigate the applicability and effectiveness of AI-aided system for automated Japanese “Group classification” of gastric endoscopic biopsy. Methods In total, 260 whole-slide images of gastric endoscopic biopsy were collected from Dalian Municipal Central Hospital from January 2015 to January 2021. These images were annotated by experienced pathologists according to the Japanese “Group classification.” Five popular convolutional neural networks, i.e., VGG16, VGG19, ResNet50, Xception, and InceptionV3 were trained and tested. The performance of the models was compared in terms of widely used metrics, namely, AUC (area under the receiver operating characteristic curve, i.e., ROC curve), accuracy, recall, precision, and F1 score. Results Results showed that ResNet50 achieved the best performance with accuracy 93.16% and AUC 0.994. Conclusion Our results demonstrated the applicability and effectiveness of DL-based system for automated Japanese “Group classification” of gastric endoscopic biopsy.

Feasibility of olfactomedin 4 as a molecular biomarker for early diagnosis of gastric neoplasia after intestinal metaplasia

Objectives This study discusses whether olfactomedin 4 (OLFM4) could be used as a sensitive and specific biomarker in the early diagnosis of gastric cancer (GC) after gastric intestinal metaplasia (GIM). Methods An integrative analysis combining data derived from the Gene Expression Omnibus (GEO) and cBioPortal databases was performed to investigate the potential molecular biomarker. Immunohistochemistry and quantitative real-time polymerase chain reactions were used to measure the expression of messenger ribonucleic acid (mRNA) and protein by OLFM4. In combination with the gastroscopic findings and the OLFM4 expression in GIM-GC, a predictive model was established. The receiver operator characteristic curve (ROC) was applied to assess the diagnostic value of the model for GIM-GC. Results According to the GEO and cBioPortal databases, OLFM4 was identified as a key gene in the diagnosis of GIM-GC. Higher protein expression of OLFM4 was found in GIM and GIM-GC compared with chronic superficial gastritis (GS) (p < 0.05). The positive expression rate of OLFM4 in paracancerous tissue (GCP) was higher than in GIM (p > 0.05). There was no significant difference between GIM-GC and GCP (p > 0.05). The mRNA expression of OLFM4 was similar to the protein expression, and the positive expression rate was higher in early GIM-GC than in GIM (p < 0.05). Conclusion Olfactomedin 4 could be used as a biomarker for the early diagnosis of GIM-GC, and the logistic predictive model could be an effective tool for increasing the early diagnostic rate.

Can Chromoendoscopy Improve the Early Diagnosis of Gastric Carcinoma in Dogs?

Simple SummaryCurrently, canine gastric carcinoma is mainly diagnosed in its late, incurable phase, and strategies for early diagnosis are lacking. In human medicine, chromoendoscopic (CE) methods such as staining the gastric mucosal surface with indigo carmine (IC), and narrow band imaging (NBI), have improved the diagnosis of precancerous gastric mucosal changes and early gastric carcinoma. This study aimed at investigating whether IC-CE and NBI-CE can improve the diagnostic yield of endoscopy in dogs. Belgian Shepherd dogs are predisposed to gastric carcinoma; thus, 30 dogs of the breed served as the study population. As a result, the study revealed that especially the combination of standard white light endoscopy (WLE) with NBI-CE allows better recognition of gastric mucosal structural changes than WLE alone. However, CE assessment templates used to predict the type of mucosal change in humans were not applicable in dogs. The value of the study lies in providing evidence that CE can improve the diagnosis of precancerous changes and early gastric carcinoma in dogs. However, current image assessment templates from human medicine need major adjustments to comprehend canine gastric mucosal conditions.Chromoendoscopy has improved the early diagnosis of gastric cancer in humans but its usefulness in dogs is unknown. This study aimed at assessing whether adding narrow band imaging (NBI) or indigo carmine (IC) chromoendoscopy (CE) can improve the diagnostic yield of standard white light endoscopy (WLE). We compared the real-time findings of canine WLE, NBI-CE, and IC-CE and corresponding histology reports with endoscopic mucosal pattern assessment templates used in human medicine. Belgian Shepherd dogs are predisposed to gastric carcinoma. Therefore, 30 dogs of this breed served as the study population. According to histology, 17/30 dogs had mucosal changes (mucous metaplasia, glandular dysplasia, and gastric carcinoma). Diagnostic yield was best when targeted biopsies were taken with WLE and NBI-CE combined (15/17 cases). WLE alone positively identified only 8/17 cases and missed a gastric carcinoma in 3/6 cases. CE assessment templates based on macroscopic mucosal patterns, broadly used in human medicine, were not readily applicable in dogs. In conclusion, the study provides evidence that using CE in dogs has the potential to improve the diagnosis of precancerous gastric mucosal pathology and early gastric carcinoma. However, current image assessment templates from human medicine need major adjustments to the patterns of canine gastric mucosa.

Diagnostic Value of Endoscopic Narrow-Band Imaging Technique in Early Gastric Cancer and Precancerous Lesions.

Objective To investigate the diagnostic value of endoscopic narrow-band imaging technique in early gastric cancer and precancerous lesions. Methods A total of 100 patients with recurrent upper gastrointestinal symptoms in our hospital from January 2017 to January 2022 were selected and divided into group A and group B according to the random number table method, with 50 cases in each group. Group A received white light endoscopy, and group B received narrow-band imaging technology combined with endoscopy. Narrow-band imaging combined with magnifying endoscopy was used to stain the area with suspicious mucosal lesions with indigo carmine and magnified observation. Results The endoscopic image clarity of group B was significantly better than that of group A in terms of lesion outline, gastric pit, and microvascular morphology (P < 0.05). There were 10 cases of early gastric cancer, 18 cases of benign lesions, and 9 cases of gastric cancer (nonearly stage); 17 cases of precancerous lesions, 12 cases of early gastric cancer, 13 cases of benign lesions, and 6 cases of gastric cancer (nonearly stage) were diagnosed by ordinary white light endoscopy. Pathological results confirmed that among the 50 patients in group B, there were 15 cases of precancerous lesions, 11 cases of early gastric cancer, 17 cases of benign lesions, and 7 cases of gastric cancer (nonearly stage). Among the 50 patients in group A, 16 were precancerous lesions, 11 were early gastric cancer, 15 were benign lesions, and 8 were gastric cancer (non early stage). In the diagnosis of precancerous lesions and early gastric cancer, the diagnostic consistency, sensitivity, and specificity of group B were better than those of group A (P < 0.05); NBI combined with endoscopy in the diagnosis of precancerous lesions and early gastric cancer (kappa = 0.860, kappa = 0.883) was more consistent with pathological diagnosis than common white light endoscopy (kappa = 0.433, kappa = 0.535). Conclusion The value of narrow-band imaging technology combined with endoscopy in the diagnosis of precancerous lesions and early gastric cancer is better than that of ordinary white light endoscopy, and it can be widely used in clinical practice.

The Roles of Immune Cells in Gastric Cancer: Anti-Cancer or Pro-Cancer?

Simple SummaryGastric cancer is still one of the leading causes of death caused by cancer in developing countries. The emerging role of immunotherapy in cancer treatment has led to more research to elucidate the roles of essential immune cells in gastric cancer prognosis. We reviewed the roles of immune cells including T cells, B cells, dendritic cells, macrophages and natural killer cells in gastric cancer. Although the studies conducted on the roles of immune cells in gastric cancer pathogenesis produced conflicting results, understanding the roles of immune cells in gastric cancer will help us to harness them for application in immunotherapy for better prognosis and management of gastric cancer patients.Despite the fact that the incidence of gastric cancer has declined over the last decade, it is still the world’s leading cause of cancer-related death. The diagnosis of early gastric cancer is difficult, as symptoms of this cancer only manifest at a late stage of cancer progression. Thus, the prognosis of gastric cancer is poor, and the current treatment for improving patients’ outcomes involves the application of surgery and chemotherapy. Immunotherapy is one of the most recent therapies for gastric cancer, whereby the immune system of the host is programmed to combat cancer cells, and the therapy differs based upon the patient’s immune system. However, an understanding of the role of immune cells, namely the cell-mediated immune response and the humoral immune response, is pertinent for applications of immunotherapy. The roles of immune cells in the prognosis of gastric cancer have yielded conflicting results. This review discusses the roles of immune cells in gastric cancer pathogenesis, specifically, T cells, B cells, macrophages, natural killer cells, and dendritic cells, as well as the evidence presented thus far. Understanding how cancer cells interact with immune cells is of paramount importance in designing treatment options for gastric cancer immunotherapy.

A new biomarker for the early diagnosis of gastric cancer: gastric juice- and serum-derived SNCG.

Aim: To explore the possibility of gastric juice (GJ)- and serum-derived SNCG as a potential biomarker for the early diagnosis of gastric cancer (GC). Materials & methods: GJ and serum samples were collected from 87 patients with GC, 38 patients with gastric precancerous lesions and 44 healthy volunteers. The levels of SNCG in GJ and serum samples were detected by ELISA. Results: The levels of SNCG in GJ and serum were significantly higher in the GC group when compared with the GPL group or the control group. The expression of SNCG in GJ and serum was associated with tumor node metastasis stage, lymph node metastasis, tumor size and drinking, and it is important for the diagnosis and prognosis of GC (p<0.05). Conclusion: The findings highlight the significance of SNCG in GC diagnosis and prognosis and implicate SNCG as a promising candidate for GC treatment.

Application of deep learning in the real-time diagnosis of gastric lesion based on magnifying optical enhancement videos

Background and aimMagnifying image-enhanced endoscopy was demonstrated to have higher diagnostic accuracy than white-light endoscopy. However, differentiating early gastric cancers (EGCs) from benign lesions is difficult for beginners. We aimed to determine whether the computer-aided model for the diagnosis of gastric lesions can be applied to videos rather than still images.MethodsA total of 719 magnifying optical enhancement images of EGCs, 1,490 optical enhancement images of the benign gastric lesions, and 1,514 images of background mucosa were retrospectively collected to train and develop a computer-aided diagnostic model. Subsequently, 101 video segments and 671 independent images were used for validation, and error frames were labeled to retrain the model. Finally, a total of 117 unaltered full-length videos were utilized to test the model and compared with those diagnostic results made by independent endoscopists.ResultsExcept for atrophy combined with intestinal metaplasia (IM) and low-grade neoplasia, the diagnostic accuracy was 0.90 (85/94). The sensitivity, specificity, PLR, NLR, and overall accuracy of the model to distinguish EGC from non-cancerous lesions were 0.91 (48/53), 0.78 (50/64), 4.14, 0.12, and 0.84 (98/117), respectively. No significant difference was observed in the overall diagnostic accuracy between the computer-aided model and experts. A good level of kappa values was found between the model and experts, which meant that the kappa value was 0.63.ConclusionsThe performance of the computer-aided model for the diagnosis of EGC is comparable to that of experts. Magnifying the optical enhancement model alone may not be able to deal with all lesions in the stomach, especially when near the focus on severe atrophy with IM. These results warrant further validation in prospective studies with more patients. A ClinicalTrials.gov registration was obtained (identifier number: NCT04563416).Clinical Trial RegistrationClinicalTrials.gov, identifier NCT04563416.

Automatic captioning of early gastric cancer using magnification endoscopy with narrow-band imaging

The detection rate for early gastric cancer (EGC) is unsatisfactory, and mastering the diagnostic skills of magnifying endoscopy with narrow-band imaging (ME-NBI) requires rich expertise and experience. We aimed to develop an EGC captioning model (EGCCap) to automatically describe the visual characteristics of ME-NBI images for endoscopists. ME-NBI images (n= 1886) from 294 cases were enrolled from multiple centers, and corresponding 5658 text data were designed following the simple EGC diagnostic algorithm. An EGCCap was developed using the multiscale meshed-memory transformer. We conducted comprehensive evaluations for EGCCap including the quantitative and quality of performance, generalization, robustness, interpretability, and assistant value analyses. The commonly used metrics were BLEUs, CIDEr, METEOR, ROUGE, SPICE, accuracy, sensitivity, and specificity. Two-sided statistical tests were conducted, and statistical significance was determined when P< .05. EGCCap acquired satisfying captioning performance by outputting correctly and coherently clinically meaningful sentences in the internal test cohort (BLEU1= 52.434, CIDEr= 36.734, METEOR= 27.823, ROUGE= 49.949, SPICE= 35.548) and maintained over 80% performance when applied to other centers or corrupted data. The diagnostic ability of endoscopists improved with the assistance of EGCCap, which was especially significant (P< .05) for junior endoscopists. Endoscopists gave EGCCap an average remarkable score of 7.182, showing acceptance of EGCCap. EGCCap exhibited promising captioning performance and was proven with satisfying generalization, robustness, and interpretability. Our study showed potential value in aiding and improving the diagnosis of EGC and facilitating the development of automated reporting in the future.

Academy of Medicine, Singapore clinical guideline on endoscopic surveillance and management of gastric premalignant lesions

Gastric cancer (GC) has a good prognosis, if detected at an early stage. The intestinal subtype of GC follows a stepwise progression to carcinoma, which is treatable with early detection and intervention using high-quality endoscopy. Premalignant lesions and gastric epithelial polyps are commonly encountered in clinical practice. Surveillance of patients with premalignant gastric lesions may aid in early diagnosis of GC, and thus improve chances of survival. An expert professional workgroup was formed to summarise the current evidence and provide recommendations on the management of patients with gastric premalignant lesions in Singapore. Twenty-five recommendations were made to address screening and surveillance, strategies for detection and management of gastric premalignant lesions, management of gastric epithelial polyps, and pathological reporting of gastric premalignant lesions.

Comparison between Endoscopic Submucosal Dissection and Surgery in Patients with Early Gastric Cancer.

Simple SummaryConventional gastrectomy combined with regional lymph node dissection has been the standard treatment for early gastric cancer (EGC). This retrospective case–control study aimed to compare the clinical outcomes of endoscopic submucosal dissection (ESD) and surgical resection for EGC in China. After propensity score-matching, there were no significant differences between the two groups for OS, RFS, and DSS. Additionally, with similar R0 resection and recurrence rates, the ESD group showed less blood loss, fewer adverse events, lower hospital cost, and a shorter operative time and hospital duration than the surgery group. Therefore, ESD can be a first-line treatment of EGC in addition to surgery.Background: Endoscopic submucosal dissection (ESD) has become a preferred treatment option for early gastric cancer (EGC). This study aimed to compare the clinical outcomes of ESD and surgical resection for EGC. Methods: This was a retrospective case–control study. Patients with a diagnosis of EGC who underwent ESD or surgery in our hospital from 2011 to 2020 were enrolled. We compared the clinical characteristics and treatment outcomes of these two groups according to propensity score-matching. The primary outcome comparison was overall survival (OS). Secondary outcomes were disease-specific survival (DSS), recurrence-free survival (RFS), and treatment-related events. Results: In the matched cohort, the ESD group showed comparable OS, RFS, and DSS with the surgery group. Statistical differences were shown in blood loss and adverse events. Furthermore, the ESD group showed lower hospital cost, as well as a shorter operative time and hospital duration than the surgery group. The R0 resection and recurrence rates were similar between the two groups. In Cox regression analysis, age, tumor size, poor differentiation, and lymphovascular invasion were regarded as independent factors of OS. Conclusions: With sufficient safety and advantages, ESD can be a first-line treatment of EGC. Preoperative evaluation is vital to the appropriate treatment and prognosis.

ADAN: An Adversarial Domain Adaptation Neural Network for Early Gastric Cancer Prediction.

Gastric cancer is a highly prevalent cancer world-wide. Accurate diagnosis of Early Gastric Cancer (EGC) is of great significance to improve the treatment and survival rate of patients. However, EGC and gastric ulcers have similar en-doscopic image characteristics, resulting in a high misdiagnosis rate. Most existing deep learning and machine learning models for EGC recognition have the disadvantages of cumbersome pre-processing steps and high leakage ratios. To address the above challenges, we propose an end-to-end Adversarial Do-main Adaptation Neural network (ADAN) for EGC prediction on endoscopic images. ADAN network consists of a source domain feature extractor, a source domain classifier, two target domain feature extractors, a target domain classifier, and a domain discriminator. A source domain feature extractor is designed to train the model on public gastrointestinal datasets, which effectively solves the problem of insufficient training data. In addition, an adaptive source-target domain mapping classifier is added to each target domain feature extractor for automatically adjusting the number of classification categories in the target domain. Experimental results show that the proposed ADAN network is superior to the most advanced methods and can accurately predict EGC in clinical practice. Clinical relevance-In this study, the EGC diagnosis model based on the adversarial domain adaptive construction will be more applicable to the real clinical scenario, with higher accuracy and sensitivity and assist the endoscopist to make more accurate diagnosis for EGC and reduce the workload.

(Invited) New Challenges in Early Diagnosis of Gastric Cancer

Graphene proved to be an excellent material for sensors’ design. Therefore, we decorated it with different heteroatoms (e.g., N, B, S) and use it as electrode material for the design of new stochastic sensors. The graphene paste was modified with oleamides, inulins, dextrins in order to get the stochastic response of the sensor. Biomarkers like MSH6, MLH1, KRAS, p53 were analysed by screening the whole blood, urine, saliva, and tissue samples with the proposed sensors. High sensitivities and low limits of determination were obtained. Excellent correlations between the results obtained using the proposed screening method and the standard method were obtained.

Quantitative and amplification-free detection of SOCS-1 CpG methylation percentage analyses in gastric cancer by fiber optic nanoplasmonic biosensor

A new innovative approach is essential for early and effective diagnosis of gastric cancer, using promoter hypermethylation of the tumor suppressor, SOCS-1, that is frequently inactivated in human cancers. We have developed an amplification-free fiber optic nanoplasmonic biosensor for detecting DNA methylation of the SOCS-1 human genome. The method is based on the fiber optic nanogold-linked sorbent assay of PCR-free DNA from human gastric tumor tissue and cell lines. We designed a specific DNA probe fabricated on the fiber core surface while the other probe is bioconjugated with gold nanoparticles in free form to allow percentage determination and differentiating the methylated and unmethylated cell lines, further demonstrating the SOCS-1 methylation occurs in cancer patients but not in normal cell lines. The observed detection limit is 0.81 fM for methylated DNA, and the detection time is within 15 min. In addition, our data were significantly correlated to the data obtained from PCR-based pyrosequencing, and yet with superior accuracy. Hence our results provide new insight to the quantitative evaluation of methylation status of the human genome and can act as an alternative to PCR with a great potential.

Leveraging Multimodal Semantic Fusion for Gastric Cancer Screening via Hierarchical Attention Mechanism

Gastroscopy is a widely adopted method for locating gastric lesions and performing the early screening and diagnosis of gastric cancer (GC). However, the effectiveness of traditional GC screening methods depends on the medical skills of the gastroscopy specialist. A lack of knowledge and experience may lead to misdiagnosis and mistreatment, especially in small-scale hospitals. Recently, there has been a significant increase in studies on data-driven computer-aided diagnosis techniques. In this article, we propose a novel intelligent decision-making method for GC screening (ID-GCS), a multimodal semantic fusion-based data-driven decision-making system. ID-GCS exploits a hybrid attention mechanism to extract textual semantics from multimodal gastroscopy reports and performs semantic fusion to integrate the semantics of textual gastroscopy reports and images, resulting in improved interpretability of gastroscopy findings. We evaluated ID-GCS using a real gastroscopy report dataset, and experimental results show that compared with state-of-the-art methods, ID-GCS achieves better sensitivity and accuracy in GC screening.

Preneoplastic Gastric Lesions and Helicobacter Pylori in Endoscopic Detection and Early Diagnosis of Gastric Cancer in Lebanese Population. A Retrospective Study from 2019-2020

Gastric cancer (GC) ranks as the fifth leading cause of cancer worldwide and is the fourth leading cause of cancer mortality. Premalignant gastric lesions (PGL) are considered as the most significant risk factor for GC. The purpose of this study.

A pump-free and high-throughput microfluidic chip for highly sensitive SERS assay of gastric cancer-related circulating tumor DNA via a cascade signal amplification strategy

Circulating tumour DNA (ctDNA) has emerged as an ideal biomarker for the early diagnosis and prognosis of gastric cancer (GC). In this work, a pump-free, high-throughput microfluidic chip coupled with catalytic hairpin assembly (CHA) and hybridization chain reaction (HCR) as the signal cascade amplification strategy (CHA–HCR) was developed for surface-enhanced Raman scattering (SERS) assays of PIK3CA E542K and TP53 (two GC-related ctDNAs). The chip consisted of six parallel functional units, enabling the simultaneous analysis of multiple samples. The pump-free design and hydrophilic treatment with polyethylene glycol (PEG) realized the automatic flow of reaction solutions in microchannels, eliminating the dependence on external heavy-duty pumps and significantly improving portability. In the reaction region of the chip, products generated by target-triggered CHA initiated the HCR, forming long nicked double-stranded DNA (dsDNA) on the Au nanobowl (AuNB) array surface, to which numerous SERS probes (Raman reporters and hairpin DNA-modified Cu2O octahedra) were attached. This CHA–HCR strategy generated numerous active “hot spots” around the Cu2O octahedra and AuNB surface, significantly enhancing the SERS signal intensity. Using this chip, an ultralow limit of detection (LOD) for PIK3CA E542K (1.26 aM) and TP53 (2.04 aM) was achieved, and the whole process was completed within 13 min. Finally, a tumour-bearing mouse model was established, and ctDNA levels in mouse serum at different stages were determined. To verify the experimental accuracy, the gold-standard qRT–PCR assay was utilized, and the results showed a high degree of consistency. Thus, this rapid, sensitive and cost-effective SERS microfluidic chip has potential as an ideal detection platform for ctDNA monitoring.

A deep learning-based model improves diagnosis of early gastric cancer under narrow band imaging endoscopy.

Diagnosis of early gastric cancer (EGC) under narrow band imaging endoscopy (NBI) is dependent on expertise and skills. We aimed to elucidate whether artificial intelligence (AI) could diagnose EGC under NBI and evaluate the diagnostic assistance of the AI system. In this retrospective diagnostic study, 21,785 NBI images and 20 videos from five centers were divided into a training dataset (13,151 images, 810 patients), an internal validation dataset (7057 images, 283 patients), four external validation datasets (1577 images, 147 patients), and a video validation dataset (20 videos, 20 patients). All the images were labeled manually and used to train an AI system using You look only once v3 (YOLOv3). Next, the diagnostic performance of the AI system and endoscopists were compared and the diagnostic assistance of the AI system was assessed. The accuracy, sensitivity, specificity, and AUC were primary outcomes. The AI system diagnosed EGCs on validation datasets with AUCs of 0.888-0.951 and diagnosed all the EGCs (100.0%) in video dataset. The AI system achieved better diagnostic performance (accuracy, 93.2%, 95% CI, 90.0-94.9%) than senior (85.9%, 95% CI, 84.2-87.4%) and junior (79.5%, 95% CI, 77.8-81.0%) endoscopists. The AI system significantly enhanced the performance of endoscopists in senior (89.4%, 95% CI, 87.9-90.7%) and junior (84.9%, 95% CI, 83.4-86.3%) endoscopists. The NBI AI system outperformed the endoscopists and exerted potential assistant impact in EGC identification. Prospective validations are needed to evaluate the clinical reinforce of the system in real clinical practice.