Search

Background: Endoscopic ultrasound-guided biliary drainage (EUS-BD), classified as choledochoduodenostomy (CDS) and hepaticogastrostomy (HGS), is a feasible and effective alternative for distal malignant biliary obstruction (MBO) in failed endoscopic retrograde cholangiopancreatography. However, the preferred technique for better outcomes has not yet been evaluated. Objectives: We compared the long-term outcomes between the techniques. Design: Retrospective comparative study. Methods: We reviewed consecutive patients who underwent EUS-CDS or EUS-HGS with transmural stent placement for distal MBO between 2009 and 2022. The primary outcome was the stent patency. The secondary outcomes were technical and clinical success, adverse events (AEs) of each technique, and independent risk factors for stent dysfunction. Results: In all, 115 patients were divided into EUS-CDS ( n = 56) and EUS-HGS ( n = 59) groups. Among them, technical success was achieved in 98.2% of EUS-CDS and 96.6% of EUS-HGS groups. Furthermore, clinical success was 96.4% in EUS-CDS and 88.1% in EUS-HGS groups, without significant difference ( p = 0.200). The mean duration of stent patency for EUS-CDS was 770.3 days while that for EUS-HGS was 164.9 days ( p = 0.010). In addition, the only independent risk factor for stent dysfunction was systematic treatment after EUS-BD [hazard ratio and 95% confidence interval 0.238 (0.066–0.863), p = 0.029]. The incidence of stent dysfunction of EUS-HGS was higher than EUS-CDS (35.1% versus 18.2%, 0.071), despite no significant differences even in late AEs. Conclusion: In distal MBO, EUS-CDS may be better than EUS-HGS with longer stent patency and fewer AEs. Furthermore, systematic treatment after EUS-BD is recommended for the improvement of stent patency.

Abstract It is debatable which needle has clear superiority of diagnostic performance in endoscopic ultrasound (EUS)-guided fine needle biopsy (FNB) of solid pancreatic masses. This study aimed to compare the performance of three needles and determine the variables that affect diagnostic accuracy. From March 2014 to May 2020, 746 patients with solid pancreatic masses who underwent EUS-FNB using three types of needles (Franseen needle, Menghini-tip needle, and Reverse-bevel needle) were retrospectively reviewed. Multivariate analysis using a logistic regression model was used to identify factors related to diagnostic accuracy. There were significant differences between the groups regarding the procurement rate of the histologic and optimal quality cores (Franseen vs. Menghini-tip vs. Reverse-bevel: 98.0% [192/196] vs. 85.8% [97/113] vs. 91.9% [331/360], P

Background/Aims: The combinatorial effects of prophylactic methods for postendoscopic retrograde cholangiopancreatography pancreatitis (PEP) in patients with risk factors remain unclear. In this network meta-analysis, we compared the efficacy of various prophylactic strategies to decrease the risk of PEP among patients with risk factors. Methods: A systematic review was performed to identify randomized controlled trials from PubMed, Embase, and the Cochrane Library through July 2021. We used frequentist network meta-analysis to compare the rates of PEP among patients who received prophylactic treatments as follows: class A, rectal nonsteroidal anti-inflammatory drugs; class B, prophylactic pancreatic stent; class C, aggressive hydration; or control, no prophylaxis or active control. We selected those studies that included patients with risk factors for PEP. Results: We identified 19 trials, comprising 4,328 participants. Class ABC (odds ratio [OR], 0.08; 95% confidence interval [CI], 0.03 to 0.24), class AC (OR, 0.10; 95% CI, 0.02 to 0.47), class AB (OR, 0.12; 95% CI, 0.05 to 0.26), class BC (OR, 0.13; 95% CI, 0.04 to 0.41), class A (OR, 0.16; 95% CI, 0.05 to 0.50), and class B (OR, 0.26; 95% CI, 0.14 to 0.46), were associated with a reduced risk of PEP as compared to that of the control. The most effective prophylaxis was ABC (0.87), followed by AC (0.68), AB (0.65), BC (0.56), A (0.49), and B (0.24) according to P-score. Conclusions: The results of this network meta-analysis suggest that the more prophylactic methods are employed, the better the outcomes. It appears that for patients with risk factors, we need to prevent PEP through the use of these well proven combination strategies.

Endoscopic retrograde cholangiopancreatography (ERCP) has been established as a standard modality for the palliation of malignant hilar biliary obstruction (MHBO). However, endoscopic stent placement is not possible in some patients because of failed biliary cannulation or tumor invasion that limits transpapillary approaches. In this situation, percutaneous transhepatic biliary drainage (PTBD) can be a reasonable alternative for failed ERCP. However, PTBD has a relatively high rate of adverse events and is frequently associated with patient inconvenience related to the external drainage catheter. Endoscopic ultrasound (EUS)-guided biliary drainage (BD) has therefore been introduced as a reliable alternative modality to PTBD in cases of failed ERCP due to an inaccessible papilla, gastric outlet obstruction, or surgically altered anatomy for MHBO. Although the field of dedicated stents for EUS-BD is rapidly advancing with increasing innovations, the debate on the most appropriate stent for EUS-guided drainage has resurfaced. Furthermore, some important questions remain unaddressed, such as which stent best improves clinical outcomes and safety in EUS-BD for MHBO. This review summarizes the fundamental principles of BD in MHBO, basic technique, technical challenges, clinical outcomes, safety profiles, comparison with other modalities, and the future perspectives of EUS-BD, especially EUS-guided hepaticogastrostomy or hepaticoduodenostomy, for MHBO.

Background/Aims: The appropriate number of band ligations during the first endoscopic session for acute variceal bleeding is debatable. We aimed to compare the technical aspects of endoscopic variceal ligation (EVL) in patients with variceal bleeding according to the number of bands placed per session. Methods: We retrospectively reviewed multicenter data from patients who underwent EVL for acute variceal bleeding. Patients were classified into minimal EVL (targeting only the foci with active bleeding or stigmata of recent bleeding) and maximal EVL (targeting potential bleeding sources in addition to the aforementioned targets) groups. The primary endpoint was 5-day treatment failure. The secondary endpoints were 30-day rebleeding, 30-day mortality, and intraprocedural adverse events. Results: Minimal EVL was associated with lower rates of hypoxia and shock during EVL than maximal EVL (hypoxia, 0.9% vs 2.9%; shock, 1.3% vs 3.4%). However, treatment failure was higher in the minimal EVL group than in the maximal EVL group (odds ratio, 1.60; 95% confidence interval, 1.06 to 2.41). Age ≥60 years, Model for End-Stage Liver Disease score ≥15, Child-Turcotte-Pugh classification C, presence of hepatocellular carcinoma, and systolic blood pressure

Background/Aims: Sphincterotomes are essential for endoscopic sphincterotomy (EST) and can also be used for cannulation in ERCP. A domestic new pull-type sphincterotome (Optimos™, Taewoong, Goyang, Korea) provides acceptable technical feasibility and safety, but there are no comparison results. Thus, this study compared the clinical performance and safety of Optimos™ sphincterotome to a conventional sphincterotome (CleverCut3™, Olympus, Tokyo, Japan) in patients who underwent ERCP. Methods: From April 2021 to July 2021, a randomized prospective comparative study was conducted on 104 consecutive patients who underwent ERCP in three medical centers. The primary endpoint was the clinical performance and safety of sphincterotomes during ERCP. Results: One hundred and four patients were assigned randomly to the Optimos™ group (n=51) or CleverCut3™ group (n=53). All demographic characteristics did not differ between the groups except the BMI. The technical success rate for cannulation, performance of EST, and total procedure time were similar in the two groups. The adverse events did not differ, even though two cases of post-ERCP pancreatitis occurred in CleverCut3™. On the other hand, in questionnaire analysis, CleverCut™ showed a better user’s convenience (median [interquartile range] 4.0 [3.0-4.0] vs. 3.0 [3.0-4.0], p=0.013) and manipulability (median [interquartile range], 4.0 [3.0-4.0] vs. 3.0 [3.0-4.0], p=0.039) than Optimos™, even though the other profiles did not reveal any differences. Conclusions: New domestic pull-type sphincterotome can offer comparable clinical performance and safety profiles to conventional sphincterotome, but it needs refinements to increase the user’s convenience and manipulability. Further improvement and innovation will be required to advance domestic medical devices.

Background/Aims: The endoscopic step-up approach is accepted as the preferred treatment for complicated or symptomatic walled-off necrosis (WON). Direct endoscopic necrosectomy (DEN) is an effective therapeutic option, but few reports describe long-term follow-up in this patient population. Thus, we aim to assess the long-term outcomes of DEN following severe necrotizing pancreatitis. Methods: The data of all acute pancreatitis patients who underwent DEN following endoscopic transmural drainage from six referral centers between 2007 and 2017 were retrospectively collected. Results: Sixty patients (76.7% male, mean age 48.3 years) underwent a median of 4 sessions of DEN starting at a median of 45.5 days after the onset of acute pancreatitis. Clinical success was achieved in 51 patients (85%), with a 35% complication rate and a 5% mortality rate. Using multivariate analysis, the risk factor associated with DEN failure or major DEN complications requiring intervention or surgery was an identified bacterial/fungal WON infection (odds ratio, 19.3; 95% confidence interval, 1.5 to 261.7). During the median follow-up period of 27 months, complicated WON recurrence was observed in 5.3% of patients, and long-term complications occurred in 24.6% of patients (four exocrine insufficiency, nine newly developed diabetes mellitus, one recurrent small bowel obstruction, one chylous ascites). Conclusions: Considering that long-term complications are similar to those observed after pancreatectomy, DEN should be performed meticulously while minimizing damage to the viable pancreatic parenchyma with adequate antibiotic escalation.

Endoscopic ultrasound (EUS)-guided tissue acquisition of pancreatic solid tumor requires a strict recommendation for its proper use in clinical practice because of its technical difficulty and invasiveness. The Korean Society of Gastrointestinal Endoscopy appointed a Task Force to draft clinical practice guidelines for EUS-guided tissue acquisition of pancreatic solid tumor. The strength of recommendation and the level of evidence for each statement were graded according to the Minds Handbook for Clinical Practice Guideline Development 2014. The committee, comprising a development panel of 16 endosonographers and an expert on guideline development methodology, developed 12 evidence-based recommendations in eight categories intended to help physicians make evidence-based clinical judgments with regard to the diagnosis of pancreatic solid tumor. This clinical practice guideline discusses EUS-guided sampling in pancreatic solid tumor and makes recommendations on circumstances that warrant its use, technical issues related to maximizing the diagnostic yield (e.g., needle type, needle diameter, adequate number of needle passes, sample obtaining techniques, and methods of specimen processing), adverse events of EUS-guided tissue acquisition, and learning-related issues. This guideline was reviewed by external experts and suggests best practices recommended based on the evidence available at the time of preparation. This guideline may not be applicable for all clinical situations and should be interpreted in light of specific situations and the availability of resources. It will be revised as necessary to cover progress and changes in technology and evidence from clinical practice.

Endoscopic ultrasound (EUS)-guided tissue acquisition of pancreatic solid tumor requires a strict recommendation for its proper use in clinical practice because of its technical difficulty and invasiveness. The Korean Society of Gastrointestinal Endoscopy (KSGE) appointed a task force to draft clinical practice guidelines for EUS-guided tissue acquisition of pancreatic solid tumor. The strength of recommendation and the level of evidence for each statement were graded according to the Minds Handbook for Clinical Practice Guideline Development 2014. The committee, comprising a development panel of 16 endosonographers and an expert on guideline development methodology, developed 12 evidence-based recommendations in eight categories intended to help physicians make evidence-based clinical judgments with regard to the diagnosis of pancreatic solid tumor. This clinical practice guideline discusses EUS-guided sampling in pancreatic solid tumor and makes recommendations on circumstances that warrant its use, technical issues related to maximizing the diagnostic yield (e.g., needle type, needle diameter, adequate number of needle passes, sample obtaining techniques, and methods of specimen processing), adverse events of EUS-guided tissue acquisition, and learning-related issues. This guideline was reviewed by external experts and suggests best practices recommended based on the evidence available at the time of preparation. This guideline may not be applicable for all clinical situations and should be interpreted in light of specific situations and the availability of resources. It will be revised as necessary to cover progress and changes in technology and evidence from clinical practice.

Backgrounds/Aims: PPARγ, farnesoid X receptor (FXR) and CYP7A1 are associated with solubility of bile. This study was performed to understand a mechanism and interactions of statin-induced PPARγ, PGC-1α and HNF-4α related to the statin-induced activation of FXR and CYP7A1, and verify whether the mevalonate pathway is involved in the mechanism. Methods: MTT assays were performed using cultured human Hep3B cells to determine the effect of atorvastatin on the cell proliferation. Expression levels of indicated proteins were measured using Western blotting assays by inhibiting the protein expression or not. Results: Atorvastatin increased expression of PPARγ, PGC-1α, HNF-4α, FXR, and CYP7A1 in Hep3B cells. PPARγ ligand of troglitazone upregulated the expression of PGC-1α, HNF-4α, FXR, and CYP7A1 in Hep3B cells. Silencing of PPARγ, PGC1α, and HNF4α using respective siRNA demonstrated that atorvastatin-induced FXR and CYP7A1 activation required sequential action of PPARγ /PGC-1α/HNF-4α. The silencing of PPARγ completely inhibited atorvastatin-induced PGC-1α expression, and the PGC1α silencing partially inhibited atorvastatin-induced PPARγ expression. The inhibition of HNF4α did not affect atorvastatin-induced PPARγ expression, but partially inhibited atorvastatin-induced PGC-1α expression. Besides, mevalonate completely reversed the effect of atorvastatin on PPARγ, PGC-1α, HNF-4α, FXR, and CYP7A1. Conclusions: Atorvastatin induces FXR and CYP7A1 activation as a result of sequential action of PPARγ/PGC-1α/HNF-4α in human hepatocytes. We propose that atorvastatin enhances solubility of cholesterol in bile by simultaneously activating of FXR and CYP7A1.

Background/Aims: Bisphosphonates are increasingly recognized for their anti-neoplastic properties, which are the result of their action on the mevalonate pathway. Our primary aim was to investigate the association between bisphosphonate use and survival in patients with pancreatic cancer. Since statins also act on the mevalonate pathway, we also investigated the effect of the combined use of bisphosphonates and statins on survival. Methods: The Surveillance, Epidemiology, and End Results : registry (SEER)-Medicare linked database was used to identify patients with pancreatic ductal adenocarcinoma (PDAC) between 2007 and 2015. Kaplan-Meier models were used to examine the association between survival with bisphosphonate use alone and in combination with statins within 1 year prior to the diagnosis of PDAC. Propensity score matching analysis and Cox-proportional hazard models were used to determine the association between overall survival with bisphosphonate use alone and combined with statins, after adjusting for relevant confounders, such as the Charlson comorbidity index score, stage, treatment, sociodemographic characteristics, and propensity score. Results: In total, 13,639 patients with PDAC were identified, and 1,203 (8.82%) used bisphosphonates. There was no difference in the mean survival duration between bisphosphonate users (7.27 months) and nonusers (7.25 months, p=0.61). After adjustment for confounders, bisphosphonate use was still not associated with improved survival (hazard ratio, 1.00; 95% confidence interval, 0.93 to 1.08; p=0.96). Combined bisphosphonate and statin use was also not associated with improved survival (hazard ratio, 0.97; 95% confidence interval, 0.87 to 1.07; p=0.48) after adjustment for confounders. Conclusions: Our findings suggest that the use of bisphosphonates, whether alone or in combination with statins, does not confer a survival advantage in patients with PDAC.

BACKGROUND AND AIMS:Although thicker needles theoretically allow more tissue to be collected, their decreased flexibility can cause mechanical damage to the endoscope, technical failure, and sample blood contamination. The effects of needle gauge on diagnostic outcomes of endoscopic ultrasound-guided fine-needle biopsy (EUS-FNB) of pancreatic mass lesions remain unknown. This study compared procurement rates of histologic cores obtained from EUS-FNB of pancreatic masses using 25- and 22-gauge core biopsy needles. PATIENTS AND METHODS:From March 2014 to July 2014, 66 patients with solid pancreatic mass underwent EUS-FNB with both 25- and 22-gauge core biopsy needles. Among them, 10 patients were excluded and thus 56 patients were eligible for the analyses. Needle sequences were randomly assigned, and two passes were made with each needle, consisting of 10 uniform to-and-fro movements on each pass with 10 mL syringe suction. A pathologist blinded to needle sequence evaluated specimens for the presence of histologic core. RESULTS:The mean patient age was 65.8 ± 9.5 years (range, 44-89 years); 35 patients (62.5%) were men. The mean pancreatic mass size was 35.3 ± 17.1 mm (range 14-122.3 mm). Twenty-eight patients (50%) had tumors at the pancreas head or uncinate process. There were no significant differences in procurement rates of histologic cores between 25-gauge (49/56, 87.5%) and 22-gauge (46/56, 82.1%, P = 0.581) needles or diagnostic accuracy using only histologic cores (98% and 95%). There were no technical failures or procedure-related adverse events. CONCLUSIONS:The 25-gauge core biopsy needle could offer acceptable and comparable outcomes regarding diagnostic performance including histologic core procurement rates compared to the 22-gauge core biopsy needle, although the differences were not statistically significant. TRIAL REGISTRATION:ClinicalTrials.gov NCT01795066.

Background: Gallstones are a well-known risk factor for acute cholecystitis. However, their role as a risk factor for gallbladder perforation (GBP) remains unclear. Therefore, this study aimed to determine the effect of gallstones on the development of GBP. Materials and methods: This large-scale retrospective cohort study enroled consecutive patients who underwent cholecystectomy for acute cholecystitis. The primary endpoint was the role of gallstones as a risk factor for developing GBP. Secondary endpoints included the clinical characteristics of GBP, other risk factors for GBP, differences in clinical outcomes between patients with acalculous cholecystitis (AC) and calculous cholecystitis (CC), and the influence of cholecystectomy timing. Results: A total of 4497 patients were included in this study. The incidence of GBP was significantly higher in the AC group compared to the CC group (5.6% vs. 1.0%, P< 0.001). However, there were no differences in ICU admission and hospital stay durations. The incidence of overall complications was significantly higher in the AC group than in the CC group (2.2% vs. 1.0%, P <0.001). Patients with AC had a higher risk of developing GBP than those with CC (odds ratio, 5.00; 95% CI, 2.94-8.33). In addition, older age (≥ 60 years), male sex, comorbidities, poor performance status, and concomitant acute cholangitis were associated with the development of GBP. Furthermore, the incidence of GBP was significantly higher in the delayed cholecystectomy group than in the early cholecystectomy group (2.0% vs. 0.9%, P <0.001). Conclusions: AC is a significant risk factor for GBP. Furthermore, early cholecystectomy can significantly reduce GBP-related morbidity and mortality. [ABSTRACT FROM AUTHOR]

Insulin-like growth factor 1 (IGF-1) and IGF-1 receptor (IGF-1R) signaling plays an important role in tumor progression in patients with certain gastrointestinal tract cancers. In addition to lowering cholesterol in serum, statins have pleiotropic effects, including anti-oxidative, anti-inflammatory or anti-neoplastic effects. Therefore, the present study investigated whether statins could induce the apoptosis of colon cancer cells and regulate the expression of IGF-1R and its associated signaling pathways in the present study. It was demonstrated that simvastatin and pravastatin suppressed cell proliferation and induced cell death in human HT-29 cells, but simvastatin was more potent than pravastatin. Simvastatin induced apoptosis in a concentration-dependent manner. In addition, simvastatin suppressed the expression of IGF-1R and inhibited the activity of phosphorylated-extracellular signal-regulated kinase (ERK)1/2 and phosphorylated-Akt activated by IGF-1. Simvastatin and IGF-1 each stimulated the activity of phosphorylated ERK1/2. However, simvastatin inhibited cell proliferation and IGF-1 stimulated cell proliferation. Mevalonic acid and PD98059 reversed the ERK activation and apoptosis induced by treatment with simvastatin. It was concluded that simvastatin induces the apoptosis of human colon cancer cells and inhibits IGF-1-induced ERK and Akt expression via the downregulation of IGF-1R expression and proapoptotic ERK activation. Simvastatin may be beneficial for the treatment of colon cancer. The present study suggested that statin may possess therapeutic potential for the treatment of colon cancer. [ABSTRACT FROM AUTHOR]

Background/Aims: Statins act as antineoplastic agents through the inhibition of cell proliferation. This study sought to demonstrate the effects of statins on extrahepatic bile duct cancer cell apoptosis and to document the changes in protein expression involved in tumor growth and suppression. Methods: Human extrahepatic bile duct cancer cells were cultured. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays were performed to determine the effect of statins on cell proliferation. Apoptosis was measured by a cell death detection enzyme-linked immunosorbent assay and caspase-3 activity assay, and flow cytometry was used to determine the percentage of cells in each phase of the cell cycle. The protein expression of Bax, Bcl-2, insulin-like growth factor 1 (IGF-1) receptor, extracellular signal-regulated kinase 1/2 (ERK1/2), and Akt was measured by Western blot analysis. Results: Simvastatin suppressed cell proliferation by inducing G1 phase cell cycle arrest in bile duct cancer cells. Furthermore, it induced apoptosis via caspase-3 activation, downregulated the expression of the Bcl-2 protein, and enhanced the expression of the Bax protein. Moreover, simvastatin suppressed the expression of the IGF-1 receptor and IGF-1-induced ERK/Akt activation. Conclusions: Simvastatin induces apoptosis in bile duct cancer cells, which suggests that it could be an antineoplastic agent for bile duct cancer. [ABSTRACT FROM AUTHOR]

The occurrence of valporic acid (VPA)-induced pancreatitis is a rare condition, predominantly observed in adolescent. Also, the occurrence of VPA-associated with hemorrhagic pseudocyst is extremely rare. We report the case of a 54-year-old man who had been taking VPA for uncontrolled seizures. He was admitted to our hospital with complaints of abdominal pain and diagnosed with acute on chronic pancreatitis. There were no other causes explaining pancreatitis, and it was thought to be due to VPA therapy. Despite of cessation of VPA, there was ongoing severe abdominal pain with fever. The patient underwent follow-up CT, which revealed a large loculated fluid collection that was observed with intra-cystic hemorrhage. After treatment with percutaneous catheter drainage, he was discharged with regression of the pancreatic pseudocyst. VPA-associated pancreatitis with hemorrhagic pseudocyst is rare but possible. Therefore, this possibility should be considered in the cause of hemorrhagic pseudocyst in a patient taking VPA. [ABSTRACT FROM AUTHOR]

Background/Aims Etomidate-based sedation during endoscopy was superior to propofol in terms of the overall hemodynamic and respiratory events in previous reported trials. However, there is a lack of evidence on the safety of etomidate in patients with liver cirrhosis (LC). The aim of this study was to compare efficacy and safety profiles including hepatic encephalopathy of etomidate and propofol for endoscopic sedation in patients with LC. Methods This single-center, randomized, double-blind trial included patients with chronic liver disease who had evidence of LC who were undergoing diagnostic or therapeutic endoscopy. Either etomidate or propofol was administered according to group allocation. The primary endpoint was the proportion of patients with any cardiopulmonary adverse events. Also, the trail making test (TMT) A consists of documenting the time required to sequentially connect randomly placed circles which are labeled from 1 to 25 was also assessed. Results Fifty-three and 56 patients were enrolled in the etomidate and propofol groups, respectively. There were no significant differences between the groups regarding the total procedure time, induction time, awake time, and recovery time. The overall incidence of respiratory and cardiovascular events was not different between the groups. The time for TMT tended to be higher in the propofol group than etomidate group without significant difference (mean±standard deviation, 92.4±116.8 vs 67.1±56.6 seconds; p=0.151). Also, the proportion of patients without any cognitive dysfunction was not different between the groups (etomidate vs propofol, 12 [22.6%] vs 11 [19.6%]; p=0.570). Conclusions Etomidate-based sedation in patients with LC was noninferior to propofol- based sedation in terms of the overall incidence of respiratory, cardiovascular and neurologic adverse events. [ABSTRACT FROM AUTHOR]

Background/Aims Visual inspection, lesion detection, and differentiation between malignant and benign features are key aspects of an endoscopist's role. The use of machine learning for the recognition and differentiation of images has been increasingly adopted in clinical practice. This study aimed to establish convolutional neural network (CNN) models to automatically classify gastric neoplasms based on endoscopic images. Methods Endoscopic white-light images of pathologically confirmed gastric lesions were collected and classified into five categories: advanced gastric cancer, early gastric cancer, high grade dysplasia, low grade dysplasia, and nonneoplasm. Three pretrained CNN models were fine-tuned using a training dataset. The classifying performance of the models was evaluated using a test dataset and a prospective validation dataset. Results A total of 5,017 images were collected from 1,269 patients, among which 812 images from 212 patients were used as the test dataset. An additional 200 images from 200 patients were collected and used for prospective validation. For the five-category classification, the weighted average accuracy of the Inception-Resnet-v2 model reached 84.6%. The mean area under the curve (AUC) of the model for differentiating gastric cancer and neoplasm was 0.877 and 0.927, respectively. In prospective validation, the Inception-Resnet-v2 model showed lower performance compared with the endoscopist with the best performance (five-category accuracy 76.4% vs 87.6%; cancer 76.0% vs 97.5%; neoplasm 73.5% vs 96.5%; p<0.001). However, there was no statistical difference between the Inception-Resnet-v2 model and the endoscopist with the worst performance in the differentiation of gastric cancer (accuracy: 76.0% vs 82.0%) and neoplasm (AUC: 0.776 vs 0.865). Conclusions The evaluated deep-learning models have the potential for clinical application in classifying gastric cancer or neoplasm on endoscopic white-light images. [ABSTRACT FROM AUTHOR]