Use of Proton Pump Inhibitors and the Risk for the Development of Gastric Cancers: A Nationwide Population-Based Cohort Study Using Balanced Operational Definitions

Simple Summary

Previous cohort studies using national claim data in Korea have shown conflicting results about the association between the use of proton pump inhibitors (PPIs) and the risk of gastric cancer. In this population-based cohort analysis using balanced operational definitions, proton pump inhibitor use was not associated with an increased risk of gastric cancer (Hazard ratio: 1.30, 95% confidence interval: 0.75–2.27). Previous cohort studies with an inappropriate operational definition for the inclusion criteria of the study subjects or index dates could be the reason of conflicting results.

Abstract

Objectives: Previous cohort studies using national claim data in Korea have shown conflicting results about the association between the use of proton pump inhibitors (PPIs) and the risk of gastric cancer. This may be due to differences in the inclusion criteria or index dates of each study. This study aims to evaluate the association between PPI use and the risk of gastric cancer using balanced operational definitions. Design: A population-based cohort analysis was conducted using the Korean National Health Insurance Service database. Subjects who used PPIs or histamine-2 receptor antagonist (H₂RA) for more than 60 days after Helicobacter pylori eradication were included. The study subjects were those who had never used H₂RAs (PPI users) and controls were those who had never used PPIs (H₂RA users). For comparison, the index dates of previous studies were adopted and analyzed. The subjects were followed until the development of gastric cancer, death, or study end. Results: A total of 10,012 subjects were included after propensity score matching. During a median follow-up of 6.56 years, PPI was not associated with an increased risk of gastric cancer (Hazard ratio: 1.30, 95% confidence interval: 0.75–2.27). This was consistent if the cumulative daily dose was adjusted (90/120/180 days), or if the index date was changed to the first day of PPI prescription or the last day of Helicobacter pylori eradication. There was no significant difference in mortality between both groups. Conclusion: PPI use was not associated with an increased risk of gastric cancer.

Impact of the Volume and Distribution of Training Datasets in the Development of Deep-Learning Models for the Diagnosis of Colorectal Polyps in Endoscopy Images

Background: Establishment of an artificial intelligence model in gastrointestinal endoscopy has no standardized dataset. The optimal volume or class distribution of training datasets has not been evaluated. An artificial intelligence model was previously created by the authors to classify endoscopic images of colorectal polyps into four categories, including advanced colorectal cancer, early cancers/high-grade dysplasia, tubular adenoma, and nonneoplasm. The aim of this study was to evaluate the impact of the volume and distribution of training dataset classes in the development of deep-learning models for colorectal polyp histopathology prediction from endoscopic images. Methods: The same 3828 endoscopic images that were used to create earlier models were used. An additional 6838 images were used to find the optimal volume and class distribution for a deep-learning model. Various amounts of data volume and class distributions were tried to establish deep-learning models. The training of deep-learning models uniformly used no-code platform Neuro-T. Accuracy was the primary outcome on four-class prediction. Results: The highest internal-test classification accuracy in the original dataset, doubled dataset, and tripled dataset was commonly shown by doubling the proportion of data for fewer categories (2:2:1:1 for advanced colorectal cancer: early cancers/high-grade dysplasia: tubular adenoma: non-neoplasm). Doubling the proportion of data for fewer categories in the original dataset showed the highest accuracy (86.4%, 95% confidence interval: 85.0–97.8%) compared to that of the doubled or tripled dataset. The total required number of images in this performance was only 2418 images. Gradient-weighted class activation mapping confirmed that the part that the deep-learning model pays attention to coincides with the part that the endoscopist pays attention to. Conclusion: As a result of a data-volume-dependent performance plateau in the classification model of colonoscopy, a dataset that has been doubled or tripled is not always beneficial to training. Deep-learning models would be more accurate if the proportion of fewer category lesions was increased.

Deep-Learning for the Diagnosis of Esophageal Cancers and Precursor Lesions in Endoscopic Images: A Model Establishment and Nationwide Multicenter Performance Verification Study

Background: Suspicion of lesions and prediction of the histology of esophageal cancers or premalignant lesions in endoscopic images are not yet accurate. The local feature selection and optimization functions of the model enabled an accurate analysis of images in deep learning. Objectives: To establish a deep-learning model to diagnose esophageal cancers, precursor lesions, and non-neoplasms using endoscopic images. Additionally, a nationwide prospective multicenter performance verification was conducted to confirm the possibility of real-clinic application. Methods: A total of 5162 white-light endoscopic images were used for the training and internal test of the model classifying esophageal cancers, dysplasias, and non-neoplasms. A no-code deep-learning tool was used for the establishment of the deep-learning model. Prospective multicenter external tests using 836 novel images from five hospitals were conducted. The primary performance metric was the external-test accuracy. An attention map was generated and analyzed to gain the explainability. Results: The established model reached 95.6% (95% confidence interval: 94.2–97.0%) internal-test accuracy (precision: 78.0%, recall: 93.9%, F1 score: 85.2%). Regarding the external tests, the accuracy ranged from 90.0% to 95.8% (overall accuracy: 93.9%). There was no statistical difference in the number of correctly identified the region of interest for the external tests between the expert endoscopist and the established model using attention map analysis (P = 0.11). In terms of the dysplasia subgroup, the number of correctly identified regions of interest was higher in the deep-learning model than in the endoscopist group, although statistically insignificant (P = 0.48). Conclusions: We established a deep-learning model that accurately classifies esophageal cancers, precursor lesions, and non-neoplasms. This model confirmed the potential for generalizability through multicenter external tests and explainability through the attention map analysis.

No-Code Platform-Based Deep-Learning Models for Prediction of Colorectal Polyp Histology from White-Light Endoscopy Images: Development and Performance Verification

Background: The authors previously developed deep-learning models for the prediction of colorectal polyp histology (advanced colorectal cancer, early cancer/high-grade dysplasia, tubular adenoma with or without low-grade dysplasia, or non-neoplasm) from endoscopic images. While the model achieved 67.3% internal-test accuracy and 79.2% external-test accuracy, model development was labour-intensive and required specialised programming expertise. Moreover, the 240-image external-test dataset included only three advanced and eight early cancers, so it was difficult to generalise model performance. These limitations may be mitigated by deep-learning models developed using no-code platforms. Objective: To establish no-code platform-based deep-learning models for the prediction of colorectal polyp histology from white-light endoscopy images and compare their diagnostic performance with traditional models. Methods: The same 3828 endoscopic images used to establish previous models were used to establish new models based on no-code platforms Neuro-T, VLAD, and Create ML-Image Classifier. A prospective multicentre validation study was then conducted using 3818 novel images. The primary outcome was the accuracy of four-category prediction. Results: The model established using Neuro-T achieved the highest internal-test accuracy (75.3%, 95% confidence interval: 71.0–79.6%) and external-test accuracy (80.2%, 76.9–83.5%) but required the longest training time. In contrast, the model established using Create ML-Image Classifier required only 3 min for training and still achieved 72.7% (70.8–74.6%) external-test accuracy. Attention map analysis revealed that the imaging features used by the no-code deep-learning models were similar to those used by endoscopists during visual inspection. Conclusion: No-code deep-learning tools allow for the rapid development of models with high accuracy for predicting colorectal polyp histology.

Computer-Aided Diagnosis of Gastrointestinal Protruded Lesions Using Wireless Capsule Endoscopy: A Systematic Review and Diagnostic Test Accuracy Meta-Analysis

Background: Wireless capsule endoscopy allows the identification of small intestinal protruded lesions, such as polyps, tumors, or venous structures. However, reading wireless capsule endoscopy images or movies is time-consuming, and minute lesions are easy to miss. Computer-aided diagnosis (CAD) has been applied to improve the efficacy of the reading process of wireless capsule endoscopy images or movies. However, there are no studies that systematically determine the performance of CAD models in diagnosing gastrointestinal protruded lesions. Objective: The aim of this study was to evaluate the diagnostic performance of CAD models for gastrointestinal protruded lesions using wireless capsule endoscopic images. Methods: Core databases were searched for studies based on CAD models for the diagnosis of gastrointestinal protruded lesions using wireless capsule endoscopy, and data on diagnostic performance were presented. A systematic review and diagnostic test accuracy meta-analysis were performed. Results: Twelve studies were included. The pooled area under the curve, sensitivity, specificity, and diagnostic odds ratio of CAD models for the diagnosis of protruded lesions were 0.95 (95% confidence interval, 0.93–0.97), 0.89 (0.84–0.92), 0.91 (0.86–0.94), and 74 (43–126), respectively. Subgroup analyses showed robust results. Meta-regression found no source of heterogeneity. Publication bias was not detected. Conclusion: CAD models showed high performance for the optical diagnosis of gastrointestinal protruded lesions based on wireless capsule endoscopy.

Establishing Machine Learning Models to Predict Curative Resection in Early Gastric Cancer with Undifferentiated Histology: Development and Usability Study

Background

Undifferentiated type of early gastric cancer (U-EGC) is included among the expanded indications of endoscopic submucosal dissection (ESD); however, the rate of curative resection remains unsatisfactory. Endoscopists predict the probability of curative resection by considering the size and shape of the lesion and whether ulcers are present or not. The location of the lesion, indicating the likely technical difficulty, is also considered.

Objective

The aim of this study was to establish machine learning (ML) models to better predict the possibility of curative resection in U-EGC prior to ESD.

Methods

A nationwide cohort of 2703 U-EGCs treated by ESD or surgery were adopted for the training and internal validation cohorts. Separately, an independent data set of the Korean ESD registry (n=275) and an Asan medical center data set (n=127) treated by ESD were chosen for external validation. Eighteen ML classifiers were selected to establish prediction models of curative resection with the following variables: age; sex; location, size, and shape of the lesion; and whether ulcers were present or not.

Results

Among the 18 models, the extreme gradient boosting classifier showed the best performance (internal validation accuracy 93.4%, 95% CI 90.4%-96.4%; precision 92.6%, 95% CI 89.5%-95.7%; recall 99.0%, 95% CI 97.8%-99.9%; and F1 score 95.7%, 95% CI 93.3%-98.1%). Attempts at external validation showed substantial accuracy (first external validation 81.5%, 95% CI 76.9%-86.1% and second external validation 89.8%, 95% CI 84.5%-95.1%). Lesion size was the most important feature in each explainable artificial intelligence analysis.

Conclusions

We established an ML model capable of accurately predicting the curative resection of U-EGC before ESD by considering the morphological and ecological characteristics of the lesions.

Use of Endoscopic Images in the Prediction of Submucosal Invasion of Gastric Neoplasms: Automated Deep Learning Model Development and Usability Study

BACKGROUND

In a previous study, we examined the use of deep learning models to classify the invasion depth (mucosa-confined versus submucosa-invaded) of gastric neoplasms using endoscopic images. The external test accuracy reached 77.3%. However, model establishment is labor intense, requiring high performance. Automated deep learning (AutoDL) models, which enable fast searching of optimal neural architectures and hyperparameters without complex coding, have been developed.

OBJECTIVE

The objective of this study was to establish AutoDL models to classify the invasion depth of gastric neoplasms. Additionally, endoscopist-artificial intelligence interactions were explored.

METHODS

The same 2899 endoscopic images that were employed to establish the previous model were used. A prospective multicenter validation using 206 and 1597 novel images was conducted. The primary outcome was external test accuracy. Neuro-T, Create ML Image Classifier, and AutoML Vision were used in establishing the models. Three doctors with different levels of endoscopy expertise were asked to classify the invasion depth of gastric neoplasms for each image without AutoDL support, with faulty AutoDL support, and with best performance AutoDL support in sequence.

RESULTS

The Neuro-T-based model reached 89.3% (95% CI 85.1%-93.5%) external test accuracy. For the model establishment time, Create ML Image Classifier showed the fastest time of 13 minutes while reaching 82.0% (95% CI 76.8%-87.2%) external test accuracy. While the expert endoscopist's decisions were not influenced by AutoDL, the faulty AutoDL misled the endoscopy trainee and the general physician. However, this was corrected by the support of the best performance AutoDL model. The trainee gained the most benefit from the AutoDL support.

CONCLUSIONS

AutoDL is deemed useful for the on-site establishment of customized deep learning models. An inexperienced endoscopist with at least a certain level of expertise can benefit from AutoDL support.

Efficacy of eradication therapy in Helicobacter pylori-negative gastric mucosa-associated lymphoid tissue lymphoma: A meta-analysis

BACKGROUND AND AIMS

The role of eradication therapy in Helicobacter pylori-negative gastric mucosa-associated lymphoid tissue (MALT) lymphoma remains controversial. The aim of this study was to investigate the efficacy of H. pylori eradication therapy as a first-line treatment for H. pylori-negative gastric MALT lymphoma.

METHODS

A literature search of studies published until October 2019 was performed using electronic databases. Studies that reported treatment response to eradication therapy as an initial treatment for patients with H. pylori-negative gastric MALT lymphoma were eligible for inclusion. The primary outcome was the complete remission rate after eradication therapy.

RESULTS

Twenty-five studies were included in the analyses. The overall pooled complete remission rate was 29.3% (95% confidence interval [CI], 22.2%-37.4%, I²  = 41.5%). There was no publication bias, and the sensitivity analyses showed consistent results. The pooled complete remission rates were lower in the subgroups of studies that had a higher incidence of translocation t(11;18)(q21;q21) (19.9%, 95% CI, 11.6%-32.0%), studies that used serological tests to exclude H. pylori infection (27.5%, 95% CI, 20.1%-36.4%), and studies where non-response to eradication therapy was determined at <12 months after treatment (27.0%, 95% CI, 15.5%-42.7%). Meta-regression analysis revealed that the pooled estimate was not significantly different in terms of the characteristics of individual studies.

CONCLUSIONS

Although the complete remission rate after eradication therapy is not high, it can be used as an initial treatment option in a subset of patients with H. pylori-negative gastric MALT lymphoma. Further studies to identify subgroups of patients who may benefit from eradication therapy are needed.

Artificial Intelligence for the Prediction of Helicobacter Pylori Infection in Endoscopic Images: Systematic Review and Meta-Analysis Of Diagnostic Test Accuracy

BACKGROUND

Helicobacter pylori plays a central role in the development of gastric cancer, and prediction of H pylori infection by visual inspection of the gastric mucosa is an important function of endoscopy. However, there are currently no established methods of optical diagnosis of H pylori infection using endoscopic images. Definitive diagnosis requires endoscopic biopsy. Artificial intelligence (AI) has been increasingly adopted in clinical practice, especially for image recognition and classification.

OBJECTIVE

This study aimed to evaluate the diagnostic test accuracy of AI for the prediction of H pylori infection using endoscopic images.

METHODS

Two independent evaluators searched core databases. The inclusion criteria included studies with endoscopic images of H pylori infection and with application of AI for the prediction of H pylori infection presenting diagnostic performance. Systematic review and diagnostic test accuracy meta-analysis were performed.

RESULTS

Ultimately, 8 studies were identified. Pooled sensitivity, specificity, diagnostic odds ratio, and area under the curve of AI for the prediction of H pylori infection were 0.87 (95% CI 0.72-0.94), 0.86 (95% CI 0.77-0.92), 40 (95% CI 15-112), and 0.92 (95% CI 0.90-0.94), respectively, in the 1719 patients (385 patients with H pylori infection vs 1334 controls). Meta-regression showed methodological quality and included the number of patients in each study for the purpose of heterogeneity. There was no evidence of publication bias. The accuracy of the AI algorithm reached 82% for discrimination between noninfected images and posteradication images.

CONCLUSIONS

An AI algorithm is a reliable tool for endoscopic diagnosis of H pylori infection. The limitations of lacking external validation performance and being conducted only in Asia should be overcome.

TRIAL REGISTRATION

PROSPERO CRD42020175957; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=175957.

Amoxicillin or tetracycline in bismuth-containing quadruple therapy as first-line treatment for Helicobacter pylori infection

AIM

To compare the efficacy and safety between modified quadruple- and bismuth-containing quadruple therapy as first-line eradication regimen for Helicobacter pylori infection.

METHODS

This study was a multicenter, randomized-controlled, non-inferiority trial. Subjects endoscopically diagnosed with H. pylori infection were randomly allocated to receive modified quadruple- (rabeprazole 20 mg bid, amoxicillin 1 g bid, metronidazole 500 mg tid, bismuth subcitrate 300 mg qid [elemental bismuth 480 mg]; PAMB) or bismuth-containing quadruple therapy (rabeprazole 20 mg bid, bismuth subcitrate 300 mg qid, metronidazole 500 mg tid, tetracycline 500 mg qid; PBMT) for 14 days. Rates of eradication success and adverse events were investigated. Antibiotic resistance was determined using the agar dilution and DNA sequencing of the clarithromycin resistance point mutations in the 23 S rRNA gene of H. pylori.

RESULTS

In total, 233 participants were randomized, 27 were lost to follow-up, and four violated the protocol. Both regimens showed an acceptable eradication rate in the intention-to-treat (PAMB: 87.2% vs. PBMT: 82.8%, P = .37), modified intention-to-treat (96.2% vs. 96%, P > .99), and per-protocol (96.2% vs. 96.9%, P > .99) analyses. Non-inferiority in the eradication success between PAMB and PBMT was confirmed. The amoxicillin-, metronidazole-, tetracycline-, clarithromycin-, and levofloxacin-resistance rates were 8.3, 40, 9.4, 23.5, and 42.2%, respectively. Antimicrobial resistance did not significantly affect the efficacy of either therapy. Overall compliance was 98.1%. Adverse events were not significantly different between the two therapies.

CONCLUSION

Modified quadruple therapy comprising rabeprazole, amoxicillin, metronidazole, and bismuth is an effective first-line treatment for the H. pylori infection in regions with high clarithromycin and metronidazole resistance.

Prediction of Submucosal Invasion for Gastric Neoplasms in Endoscopic Images Using Deep-Learning

Endoscopic resection is recommended for gastric neoplasms confined to mucosa or superficial submucosa. The determination of invasion depth is based on gross morphology assessed in endoscopic images, or on endoscopic ultrasound. These methods have limited accuracy and pose an inter-observer variability. Several studies developed deep-learning (DL) algorithms classifying invasion depth of gastric cancers. Nevertheless, these algorithms are intended to be used after definite diagnosis of gastric cancers, which is not always feasible in various gastric neoplasms. This study aimed to establish a DL algorithm for accurately predicting submucosal invasion in endoscopic images of gastric neoplasms. Pre-trained convolutional neural network models were fine-tuned with 2899 white-light endoscopic images. The prediction models were subsequently validated with an external dataset of 206 images. In the internal test, the mean area under the curve discriminating submucosal invasion was 0.887 (95% confidence interval: 0.849–0.924) by DenseNet−161 network. In the external test, the mean area under the curve reached 0.887 (0.863–0.910). Clinical simulation showed that 6.7% of patients who underwent gastrectomy in the external test were accurately qualified by the established algorithm for potential endoscopic resection, avoiding unnecessary operation. The established DL algorithm proves useful for the prediction of submucosal invasion in endoscopic images of gastric neoplasms.

Endoscopic Submucosal Dissection of Papillary Gastric Adenocarcinoma; Systematic Review

This study evaluated the possibility of endoscopic submucosal dissection (ESD) for early gastric cancer with papillary adenocarcinoma (EGC-PAC). PAC, an uncommon pathologic type of stomach cancer, is classified into differentiated-type histology. However, aggressive features, including a high rate of submucosal invasion, lymphovascular invasion (LVI), and lymph node metastasis (LNM), have been reported in studies with surgical specimens. Treatment outcomes of ESD for EGC-PAC have not been precisely demonstrated. Core databases were sought for the following inclusion criteria: studies of endoscopic resection or surgery of EGC-PAC presenting the following therapeutic indicators; en bloc resection, complete resection, curative resection, recurrence, complications associated with procedures, LVI, or LNM that enabled an analysis of ESD possibility. Overall, 15 studies were included for systematic review. Frequent submucosal invasion and high LVI were noted in EGC-PAC. However, PAC was not significantly associated with LNM. Pooled en bloc resection, complete resection, and curative resection rates were 89.7% (95% confidence interval: 55.3%-98.4%), 85.3% (67.7%-94.2%), and 67% (43%-84.5%), respectively. No LNM was observed if EGC-PAC satisfied the curative resection criteria. ESD seems technically feasible, although a high LVI rate results in a lower rate of curative resection.

Helicobacter pylori Infection in Patients with Chronic Kidney Disease: A Systematic Review and Meta-Analysis

Background/Aims

Insufficient systematic reviews were conducted in the previous meta-analyses about the prevalence of Helicobacter pylori infection in patients with chronic kidney disease (CKD). The aim of this study was to evaluate the prevalence of H. pylori infection in patients with CKD.

Methods

A systematic review of studies that evaluated the prevalence of H. pylori infection in patients with CKD compared to a control group was performed. Only studies with adult patients were included, and studies with renal transplant recipients or diabetic nephropathy patients were excluded. Random-effects model meta-analyses with sensitivity analyses and subgroup analyses were conducted to confirm the robustness of the main result. A meta-regression analysis was conducted to explore the influence of potential heterogeneity on the outcomes. The methodological quality of the included publications was evaluated using the Risk of Bias Assessment tool for Nonrandomized Studies. Publication bias was also assessed.

Results

In total, 47 studies were identified and analyzed. The total prevalence of H. pylori infection was 48.2% (1,968/4,084) in patients with CKD and 59.3% (4,097/6,908) in the control group. Pooled analysis showed a significantly lower prevalence of H. pylori infection in patients with CKD (vs control group: odds ratio, 0.64; 95% confidence interval, 0.52 to 0.79). Sensitivity analyses revealed consistent results, and meta-regression analysis showed no significant confounders. No publication bias was detected.

Conclusions

The results of this study suggest a lower prevalence of H. pylori infection in patients with CKD.

Delta Neutrophil Index for the Prediction of Prognosis in Acute Gastrointestinal Diseases; Diagnostic Test Accuracy Meta-Analysis

Delta neutrophil index (DNI) is a novel diagnostic and prognostic biomarker of various infectious or inflammatory conditions. However, data on optimal measurement time are scarce, and no studies have evaluated the potential role of the DNI as a prognostic biomarker of gastrointestinal diseases with diagnostic test accuracy meta-analysis. Core databases were searched. The inclusion criteria were as follows: patients who have gastrointestinal diseases and DNI measurements presenting diagnostic indices for predicting the prognosis, including severity, surgical outcomes, and mortality from gastrointestinal diseases. We identified twelve studies for the systematic review and ten studies for the quantitative analysis. Pooled area under the curve, sensitivity, specificity, and diagnostic odds ratio of DNI at the initial admission date were 0.82 (95% confidence interval: 0.78–0.85), 0.75 (0.52–0.89), 0.76 (0.63–0.86), and 10 (3–35), respectively. Meta-regression showed no reasons for heterogeneity and publication bias was not detected. Fagan’s nomogram indicated that the posterior probability of ‘poor prognosis’ was 76% if the test was positive, and ‘no poor prognosis’ was 25% if the test was negative. The DNI can be considered as a reliable initial measurement biomarker for predicting prognosis in patients with gastrointestinal diseases,

Endoscopic Submucosal Dissection of Early Gastric Cancer with Mixed-Type Histology: A Systematic Review

BACKGROUND

Endoscopic submucosal dissection (ESD) criteria are histologically categorized by early gastric cancer (EGC) with differentiated- and undifferentiated-type histology. However, EGC is histologically heterogenous and there have been no separate criteria for EGC with mixed-type histology [EGC-MH; differentiated-type predominant EGC mixed with an undifferentiated component (EGC-MD) or undifferentiated-type predominant EGC mixed with a differentiated component (EGC-MU)]. Moreover, therapeutic outcomes of ESD for EGC-MH have not been clearly described.

AIM

This study aimed to evaluate the feasibility of ESD for EGC-MH.

METHODS

We searched core databases for specific inclusion factors: patients with EGC-MH, intervention of ESD, and at least one of the following outcomes: rate of en bloc, complete, curative resection, recurrence, procedure-related adverse event, lymphovascular invasion (LVI), or lymph node metastasis (LNM) that enabled evaluation of feasibility of ESD.

RESULTS

A total of eight (systematic review) and four studies (meta-analysis) were included. There was no robustness in age, location, or morphology of EGC-MH. Moderately differentiated adenocarcinoma was frequent in pre-ESD biopsy. EGC-MH showed larger size, deeper invasion, and higher rates of LVI/LNM than pure-type EGC. Total en bloc, complete resection, and curative resection rates were 94.6% (95% confidence interval 86.6-97.9%), 77.8% (57.9-89.9%), and 55.1% (50.4-59.6%), respectively. There was no LNM or extra-gastric recurrence after ESD if the EGC-MD met the curative resection criteria. However, the EGC-MD itself was a risk factor for non-curative resection. (Margin positivity was the most common reason.) CONCLUSIONS: Although ESD seems to be technically feasible, inaccurate prediction of lateral or vertical margin leads to lower curative resection rate. Application of more strict indication is needed for EGC-MH.

Automated classification of gastric neoplasms in endoscopic images using a convolutional neural network

BACKGROUND

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 non-neoplasm. 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 5017 images were collected from 1269 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).

CONCLUSION

The evaluated deep-learning models have the potential for clinical application in classifying gastric cancer or neoplasm on endoscopic white-light images.

The most influential articles in Helicobacter pylori research: A bibliometric analysis

OBJECTIVE

The number of articles that researchers must be familiar with is increasing, along with the importance of selective searching and summarization. This study aimed to assess and characterize the most influential articles in Helicobacter pylori research.

METHODS

We performed a search of the top-100 cited articles using the Web of Science Core Collection (WoSCC) and Google Scholar from their inception to 2018. The top-100 Altmetric Attention Score (AAS) articles based on online media mentions were also searched using the term H pylori. Each article was evaluated for the following characteristics: citation number, title, journal, publication year, and authorship.

RESULTS

The citation number for the top-100 WoSCC articles ranged from 44 to 367. Gut published the largest number of articles (11%). In the top-100 Google Scholar articles, Lancet had the largest number of articles (13%); however, among the top-1000 cited articles published after 2012, Helicobacter published the largest number (46%). The largest number of top-100 AAS articles was published by PLOS Pathogens (6%). PubMed Central articles' citations in WoSCC or Google Scholar showed significant correlation with those from each metric; however, AAS showed no correlation. The proportion of basic research was 36%-37% in top-cited articles; but, 52% in the top-100 AAS articles. No time trend in the number of publications or citations of basic/clinical research in the top-100 bibliometrics was found. "Meta-analysis/systematic review," "gastric cancer," "eradication," and "association" were the most influential title words.

CONCLUSION

This study presents a detailed list of top-100 articles, journals, authors, and topic title words.

Delta neutrophil index as a predictor of disease severity, surgical outcomes, and mortality rates in gastrointestinal diseases: Rationale for a meta-analysis of diagnostic test accuracy

BACKGROUND

Delta neutrophil index (DNI) is the ratio of the number of immature granulocytes and the total neutrophil count in peripheral circulation. DNI precedes changes in white blood cell or neutrophil counts due to the course of granular leukocyte differentiation in infectious and inflammatory conditions, beginning with immature granulocyte formation. The role of DNI as a biomarker of various infectious or inflammatory conditions has been reported. However, no studies explored the potential role of DNI as an initial biomarker for predicting disease severity, surgical outcomes, and mortality rates of gastrointestinal diseases with pooled diagnostic test accuracy. This study aims to provide evidence that DNI is a predictor of disease severity, surgical outcomes, and mortality rates in patients with gastrointestinal diseases in emergency medical departments.

METHODS

MEDLINE, EMBASE, and the Cochrane Library will be searched using common keywords (inception to July 2019) by 2 independent evaluators. Inclusion criteria will be patients with gastrointestinal diseases, DNI measurements performed in the emergency department, indices of diagnostic performance (sensitivity, specificity, predictive values, and likelihood ratios) of DNI for predicting severity, surgical outcomes, and mortality rate of gastrointestinal diseases. True and false positives and negatives will be calculated based on the diagnostic indices of each study. All types of study designs with full-text literature written in English will be included. Risk of bias will be assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool. Descriptive data synthesis will be conducted and quantitative synthesis (bivariate and hierarchical summary receiver operating characteristic model) will be performed if the included studies are sufficiently homogenous. Meta-regression, sensitivity analysis, publication bias, and Fagan nomogram will be analyzed and described.

RESULTS

The pooled synthesis of the diagnostic performance of various gastrointestinal diseases with different cut-off values for DNI may limit the interpretation of uniform diagnostic validity. The authors will contact the corresponding authors for the missing values, requesting the original data in each study. However, if there are no responses from these authors, these studies will be excluded.

CONCLUSION

This study will provide diagnostic validity of DNI as an initial marker for the prediction of severity, surgery, and mortality of gastrointestinal diseases.

Alcohol consumption is associated with the risk of developing colorectal neoplasia: Propensity score matching analysis

Although alcohol intake is known to be associated with the development of colorectal cancer, the effect of alcohol consumption on the development of colorectal neoplasm (CRN) is unclear. We performed a retrospective cohort analysis with 1 to 1 propensity score matching in a single center of Korea. Among 1,448 patients who underwent index and surveillance colonoscopy, 210 matched pairs were analyzed. The 5-year cumulative occurrence of overall CRN after index colonoscopy was higher in the significant alcohol consumption group (defined as alcohol consumption more than 30 g/day in men and 20 g/day in women) (vs. without significant alcohol consumption group) (40% vs. 27.6%, p = 0.004). Significant alcohol consumption increased the development of overall CRN (adjusted hazard ratio [aHR]: 1.86, 95% confidence interval [CI]: 1.28–2.70, p = 0.001) at surveillance colonoscopy. However, this effect was not valid on the development of advanced CRN. In subgroup analysis considering the risk classification of index colonoscopy, significant alcohol consumption increased the overall CRN development at surveillance colonoscopy in the normal group (patients with no detected adenoma in the index colonoscopy) (aHR: 1.90, 95% CI: 1.16–3.13, p = 0.01). Alcohol consumption habits should be considered in optimizing time intervals of surveillance colonoscopy.

Prediction of Chronic Atrophic Gastritis and Gastric Neoplasms by Serum Pepsinogen Assay: A Systematic Review and Meta-Analysis of Diagnostic Test Accuracy

Serum pepsinogen assay (sPGA), which reveals serum pepsinogen (PG) I concentration and the PG I/PG II ratio, is a non-invasive test for predicting chronic atrophic gastritis (CAG) and gastric neoplasms. Although various cut-off values have been suggested, PG I ≤70 ng/mL and a PG I/PG II ratio of ≤3 have been proposed. However, previous meta-analyses reported insufficient systematic reviews and only pooled outcomes, which cannot determine the diagnostic validity of sPGA with a cut-off value of PG I ≤70 ng/mL and/or PG I/PG II ratio ≤3. We searched the core databases (MEDLINE, Cochrane Library, and Embase) from their inception to April 2018. Fourteen and 43 studies were identified and analyzed for the diagnostic performance in CAG and gastric neoplasms, respectively. Values for sensitivity, specificity, diagnostic odds ratio, and area under the curve with a cut-off value of PG I ≤70 ng/mL and PG I/PG II ratio ≤3 to diagnose CAG were 0.59, 0.89, 12, and 0.81, respectively and for diagnosis of gastric cancer (GC) these values were 0.59, 0.73, 4, and 0.7, respectively. Methodological quality and ethnicity of enrolled studies were found to be the reason for the heterogeneity in CAG diagnosis. Considering the high specificity, non-invasiveness, and easily interpretable characteristics, sPGA has potential for screening of CAG or GC.

Application of artificial intelligence in gastroenterology

Artificial intelligence (AI) using deep-learning (DL) has emerged as a breakthrough computer technology. By the era of big data, the accumulation of an enormous number of digital images and medical records drove the need for the utilization of AI to efficiently deal with these data, which have become fundamental resources for a machine to learn by itself. Among several DL models, the convolutional neural network showed outstanding performance in image analysis. In the field of gastroenterology, physicians handle large amounts of clinical data and various kinds of image devices such as endoscopy and ultrasound. AI has been applied in gastroenterology in terms of diagnosis, prognosis, and image analysis. However, potential inherent selection bias cannot be excluded in the form of retrospective study. Because overfitting and spectrum bias (class imbalance) have the possibility of overestimating the accuracy, external validation using unused datasets for model development, collected in a way that minimizes the spectrum bias, is mandatory. For robust verification, prospective studies with adequate inclusion/exclusion criteria, which represent the target populations, are needed. DL has its own lack of interpretability. Because interpretability is important in that it can provide safety measures, help to detect bias, and create social acceptance, further investigations should be performed.

Pitfalls in the Interpretation of Publications about Endoscopic Submucosal Dissection of Early Gastric Cancer with Undifferentiated-Type Histology

Endoscopic submucosal dissection (ESD) is a standard treatment for patients with gastrointestinal neoplasms with a negligible risk of lymph node metastasis. ESD enables en bloc resection of gastrointestinal neoplasms and organ preservation, thereby, avoiding surgical treatment. Although small (<2 cm) intramucosal early gastric cancer with undifferentiated-type histology (EGC-UH) without ulceration is included in the expanded criteria for ESD, controversies remain due to different biology and characteristics compared to EGC with differentiated-type histology. The authors previously presented studies about the technical feasibility of ESD for these lesions using a meta-analysis and retrospective multicenter analysis. However, many pitfalls were identified in the interpretation of studies analyzing histologic discrepancy, mixed-type histology, criteria-based analysis of therapeutic outcomes, interpretation of curative resection, and long-term clinical outcomes. In this review, the authors discuss pitfalls in the interpretation of publications on ESD for EGC-UH.

Diagnostic performance of serum pepsinogen assay for the prediction of atrophic gastritis and gastric neoplasms: Protocol for a systematic review and meta-analysis

BACKGROUND

Serum pepsinogen assay (sPGA) combining concentration of pepsinogen I (PG I), and the ratio of PG I/II is the noninvasive biomarker for predicting chronic atrophic gastritis (CAG) and neoplasms reflecting mucosal secretory status. Although various cut-off values have been suggested, PG I ≤70 ng/mL and PG I/II ≤3 have been widely accepted. However, previous studies for diagnostic test accuracy presented only pooled outcomes, which cannot discriminate the diagnostic validity of sPGA with cut-off of PG I ≤70 ng/mL and PG I/II ≤3.

METHODS

We will search the core databases [MEDLINE (through PubMed), the Cochrane Library, and Embase] from their inception to December 2018 by 2 independent evaluators. The P.I.C.O. is as follows; Patients: who have histologically proven CAG or gastric neoplasms, Intervention: sPGA with cut-off of PG I ≤70 ng/mL and/or PG I/II ≤3, Comparison: none, Outcome: diagnostic performance indices of sPGA for CAG and gastric neoplasms (sensitivity, specificity, positive predictive value, negative predictive value, likelihood ratios) (if, true/false positive, true/false negative values are presented, diagnostic performance indices will be calculated). All types of study design with full text will be sought and included. The risk of bias will be assessed using the QUADAS-2 tool. Descriptive data synthesis is planned and quantitative synthesis (bivariate and HSROC model) will be used if the included studies are sufficiently homogenous. Publication bias will be assessed.

RESULTS

The results will provide clinical evidence for diagnostic validity of sPGA.

CONCLUSION

This study will provide evidence of sPGA for predicting CAG and gastric neoplasms.