Magnifying narrowband imaging is more accurate than conventional white-light imaging in diagnosis of gastric mucosal cancer

Clinical Features of False-Negative Early Gastric Cancers: A Retrospective Study of Endoscopic Submucosal Dissection Cases

Background

We frequently encounter early gastric cancer (EGC) that could not be detected in the previous esophagogastroduodenoscopy even if the procedure was annually performed. However, little evidence exists regarding the characteristics of false-negative EGCs. Our aim was to reveal the clinical features of false-negative EGCs.

Methods

We retrospectively reviewed cases of endoscopic submucosal dissection (ESD) for EGCs in Fukuchiyama City Hospital between January 2013 and May 2019. False-negative EGCs were defined as EGCs within 3 years of negative endoscopy. We evaluated the clinical characteristics of false-negative and initially detected EGCs and the difference in the detected and last missed endoscopy in false-negative EGCs. The miss rates of false-negative EGCs were compared between trainees (nonboard-certified endoscopists) and experienced endoscopists (board-certified endoscopists); thereafter, the characteristics of false-negative EGCs missed by trainees were investigated.

Results

Of 219 cases, 119 were classified as false-negative EGCs. False-negative EGCs were characterized as smaller lesions, which presented with normal color or gastritis-like appearance, and were diagnosed after ESD and H. pylori eradication (P < 0.01). The rate of trainees in the last missed endoscopy was significantly higher than that in the detected endoscopy. The miss rate of false-negative EGC by trainees was higher than that of experienced endoscopists but not significantly different (0.70% vs. 0.57%, P = 0.08). The false-negative EGCs missed by trainees were characterized as reddish or well-differentiated lesions, which were located in the lower or lesser curvature of the stomach (P < 0.05).

Conclusion

The characteristics of false-negative EGCs were similar to those of H. pylori-eradicated EGC. Procedures with shortened examination time and those performed by trainees were risk factors of missing false-negative EGCs. Trainees should pay attention to reddish or well-differentiated EGCs located in the lower or lesser curvature of the stomach.

Application of artificial intelligence using a convolutional neural network for diagnosis of early gastric cancer based on magnifying endoscopy with narrow-band imaging

BACKGROUND AND AIM

Magnifying endoscopy with narrow-band imaging (ME-NBI) has made a huge contribution to clinical practice. However, acquiring skill at ME-NBI diagnosis of early gastric cancer (EGC) requires considerable expertise and experience. Recently, artificial intelligence (AI), using deep learning and a convolutional neural network (CNN), has made remarkable progress in various medical fields. Here, we constructed an AI-assisted CNN computer-aided diagnosis (CAD) system, based on ME-NBI images, to diagnose EGC and evaluated the diagnostic accuracy of the AI-assisted CNN-CAD system.

METHODS

The AI-assisted CNN-CAD system (ResNet50) was trained and validated on a dataset of 5574 ME-NBI images (3797 EGCs, 1777 non-cancerous mucosa and lesions). To evaluate the diagnostic accuracy, a separate test dataset of 2300 ME-NBI images (1430 EGCs, 870 non-cancerous mucosa and lesions) was assessed using the AI-assisted CNN-CAD system.

RESULTS

The AI-assisted CNN-CAD system required 60 s to analyze 2300 test images. The overall accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of the CNN were 98.7%, 98%, 100%, 100%, and 96.8%, respectively. All misdiagnosed images of EGCs were of low-quality or of superficially depressed and intestinal-type intramucosal cancers that were difficult to distinguish from gastritis, even by experienced endoscopists.

CONCLUSIONS

The AI-assisted CNN-CAD system for ME-NBI diagnosis of EGC could process many stored ME-NBI images in a short period of time and had a high diagnostic ability. This system may have great potential for future application to real clinical settings, which could facilitate ME-NBI diagnosis of EGC in practice.

Magnifying endoscopy in detecting early gastric cancer: A network meta-analysis of prospective studies

BACKGROUND

Conventional white-light imaging endoscopy (C-WLI) had a significant number of misdiagnosis in early gastric cancer (EGC), and magnifying endoscopy (ME) combined with different optical imaging was more accurate in the diagnosis of EGC. This study aimed to evaluate the accuracy of ME and compare the accuracy of ME with different optical imaging in detecting EGC.

METHODS

A comprehensive literature search was conducted to identify all relevant studies. Pair-wise meta-analysis was conducted to evaluate the accuracy of ME, and Bayesian network meta-analysis was performed to combine direct and indirect evidence and estimate the relative effects.

RESULTS

Eight prospective studies were identified with a total of 5948 patients and 3 optical imaging in ME (ME with WLI (M-WLI), ME with narrow-band imaging (M-NBI), and ME with blue laser imaging (M-BLI)). Pair-wise meta-analysis showed a higher accuracy of ME than C-WLI (OR: 2.97, 95% CI: 1.68∼5.25). In network meta-analysis, both M-NBI and M-BLI were more accurate than M-WLI (OR: 2.56, 95% CI: 2.13∼3.13; OR: 3.13, 95% CI: 1.85∼5.71). There was no significant difference between M-NBI and M-BLI.

CONCLUSION

ME was effective in improving the detecting rate of EGC, especially with NBI or BLI.

White Opaque Substance, a New Optical Marker on Magnifying Endoscopy: Usefulness in Diagnosing Colorectal Epithelial Neoplasms

Background/Aims

A white substance that is opaque to endoscopic light is sometimes observed in the epithelium during narrowband imaging with magnifying endoscopy of gastric or colorectal epithelial neoplasms. This prospective observational study aimed to determine whether the morphology of the white opaque substance (WOS) allows differential diagnosis between colorectal adenoma and carcinoma.

Methods

A consecutive series of patients with colorectal adenomas or early carcinomas who underwent endoscopic resection or surgical excision were studied. The morphology of the WOS was determined based on endoscopic images before the histopathological diagnosis was performed. The primary outcome was the diagnostic performance of an irregular WOS as a marker of colorectal carcinoma.

Results

The study analyzed 125 lesions. A total of 33 lesions showed an irregular WOS, and 92 lesions showed a regular WOS. Among the 33 lesions found to show an irregular WOS, 30 were carcinomas. Among the 92 lesions showing a regular WOS, 79 were adenomas. With irregular WOS as a marker of carcinoma, the diagnostic accuracy was 87%, sensitivity was 91%, and specificity was 86%.

Conclusions

This study demonstrated the potential usefulness of the morphology of the WOS as a marker for the differential diagnosis between adenoma and carcinoma in cases of colorectal epithelial neoplasms.

A case of signet ring cell carcinoma and mucosa-associated lymphoid tissue lymphoma of the stomach diagnosed simultaneously via magnifying endoscopy with narrow-band imaging

We present the case of a 57-year-old woman who experienced 3 weeks of intermittent tarry stools and epigastric pain. She had a 25-year history of smoking and a 12-year history of heavy alcohol consumption but had discontinued both 12 years prior. Laboratory investigations revealed elevated anti-H. Pylori IgG antibody levels. Conventional upper gastrointestinal endoscopy revealed two lesions. Magnifying endoscopy with narrow-band imaging showed the characteristic appearance of a diffuse-type gastric cancer 0-IIc lesion with a demarcation line and abnormal vessels ("wavy micro-vessels" and a "cork-screw pattern"). There was also "ballooning" of the crypts and a "tree-like vessel appearance" in an "unstructured area" characteristic of gastric mucosa-associated lymphoid tissue lymphoma with ulceration. Accurate target biopsies were obtained. We performed a laparoscopic total gastrectomy with D1 lymphadenectomy. Pathological examination revealed poorly differentiated stage I adenocarcinoma of the stomach with features of signet ring cell carcinoma as well as stage I mucosa-associated lymphoid tissue lymphoma. In conclusion, we encountered a case of co-existing diffuse-type gastric cancer and gastric mucosa-associated lymphoid tissue lymphoma. Magnifying endoscopy with narrow-band imaging was effective for diagnosing these two co-existing tumors and resulted in the collection of adequate biopsy specimens allowing for an accurate pathological diagnosis and optimal treatment.

Current status and future perspective of artificial intelligence applications in endoscopic diagnosis and management of gastric cancer

Image recognition using artificial intelligence (AI) has progressed significantly due to innovative technologies such as machine learning and deep learning. In the field of gastric cancer (GC) management, research on AI-based diagnosis such as anatomical classification of endoscopic images, diagnosis of Helicobacter pylori infection, and detection and qualitative diagnosis of GC is being conducted, and an accuracy equivalent to that of physicians has been reported. It is expected that AI will soon be introduced in the field of endoscopic diagnosis and management of gastric cancer as a supportive tool for physicians, thus improving the quality of medical care.

Role of endoscopy in suspicion of atrophic gastritis with and without intestinal metaplasia in comparison to histopathology

Background and study aims

Atrophic gastritis (AG) and intestinal metaplasia (IM) are established premalignant gastric lesions. Many studies documented a poor correlation between esophagogastroduodenoscopy (EGD) and histopathological (HP) findings of precancerous gastric lesions. The aim was to bridge the gap between endoscopy and HP in detection of chronic gastritis, AG and IM.

Patients and methods

a prospective single-center study involved 150 patients with endoscopic criteria of gastric lesions with upper gastrointestinal symptoms referred for upper GI endoscopy met the endoscopic criteria and classified according to HP of biopsies from targeted gastric lesions into chronic gastritis (GI), AG(GII) or IM(GIII). We correlated the endoscopic criteria of the 3 groups with the HP results.

Results

(73 males & 75 females) with ages ranged17-75 years and mean± SD was 41.96 ± 15.95. GI, GII & GIII were [42 patients (28%),82 patients (54.7%) and 26 patients (17.3%)], respectively. Diffuse gastric mottling was more common in GI (74.3%, P<0.001), visible submucosal vessels, gastric atrophy predominated in GII (75.6, 82.3 & 73.1% (P 0.005,0.4 & <0.01)), respectively. Whitish raised lesions were more specific in GIII (85.7%) (P<0.001). The sensitivity and specificity of endoscopic suspicion of chronic gastritis were (86&88% in GI), (87&85% in GII) and (54% & 100% in GIII) (p-0.001). The logistic regression model for risk factors was χ2= 25.74 and 49.32, p < 0.001.

Conclusion

Conventional endoscopy has high sensitivity and specificity for suspicion of chronic gastritis and AG, but low sensitivity and very high specificity for IM. Targeted biopsies may be valuable with image enhanced techniques.

Does previous biopsy lead to cancer overdiagnosis of superficial non-ampullary duodenal epithelial tumors?

Background and study aims  We aimed to evaluate the diagnostic performance of magnifying endoscopy with narrow-band imaging (M-NBI) in superficial non-ampullary duodenal epithelial tumors (SNADETs) regarding the absence or presence of biopsy before M-NBI diagnosis. Patients and methods  Clinicopathological data were retrospectively reviewed for 99 SNADETs from 99 patients who underwent endoscopic resection. The 99 tumors were divided into the non-biopsy group (32 lesions not undergoing biopsy before M-NBI examination) and the biopsy group (67 lesions undergoing biopsy before M-NBI examination). We investigated the correlation between the M-NBI diagnosis and the histopathological diagnosis of the SNADETs in both groups. Results  According to the modified revised Vienna classification, 31 tumors were classified as category 3 (C3) (low-grade adenoma) and 68 as category 4/5 (C4/5) (high-grade adenoma/cancer). The accuracy, sensitivity, and specificity of preoperative M-NBI diagnoses in the non-biopsy group vs the biopsy group were 88 % (95 % confidence interval: 71.0 - 96.5) vs 66 % (51.5 - 75.5), P  = 0.02; 95 % (77.2 - 99.9) vs 89 % (76.4 - 96.4), P  = 0.39; and 70 % (34.8 - 93.3) vs 14 % (3.0 - 36.3), P  < 0.01, respectively. Notably, in the biopsy group, the specificity of M-NBI in SNADETs was low at only 14 % because we over-diagnosed most C3 lesions as C4/5. M-NBI findings might have been compromised by the previous biopsy procedure itself. Conclusions  In the non-biopsy group, the accuracy of M-NBI in SNADETs was excellent in distinguishing C4/5 lesions from C3. The M-NBI findings in SNADETs should be evaluated while carefully considering the influence of a previous biopsy.

Artificial intelligence for cancer detection of the upper gastrointestinal tract

In recent years, artificial intelligence (AI) has been found to be useful to physicians in the field of image recognition due to three elements: deep learning (that is, CNN, convolutional neural network), a high-performance computer, and a large amount of digitized data. In the field of gastrointestinal endoscopy, Japanese endoscopists have produced the world's first achievements of CNN-based AI system for detecting gastric and esophageal cancers. This study reviews papers on CNN-based AI for gastrointestinal cancers, and discusses the future of this technology in clinical practice. Employing AI-based endoscopes would enable early cancer detection. The better diagnostic abilities of AI technology may be beneficial in early gastrointestinal cancers in which endoscopists have variable diagnostic abilities and accuracy. AI coupled with the expertise of endoscopists would increase the accuracy of endoscopic diagnosis.

Early gastric cancer detection in high-risk patients: a multicentre randomised controlled trial on the effect of second-generation narrow band imaging

OBJECTIVE

Early detection of gastric cancer has been the topic of major efforts in high prevalence areas. Whether advanced imaging methods, such as second-generation narrow band imaging (2G-NBI) can improve early detection, is unknown.

DESIGN

This open-label, randomised, controlled tandem trial was conducted in 13 hospitals. Patients at increased risk for gastric cancer were randomly assigned to primary white light imaging (WLI) followed by secondary 2G-NBI (WLI group: n=2258) and primary 2G-NBI followed by secondary WLI (2G-NBI group: n=2265) performed by the same examiner. Suspected early gastric cancer (EGC) lesions in both groups were biopsied. Primary endpoint was the rate of EGC patients in the primary examination. The main secondary endpoint was the positive predictive value (PPV) for EGC in suspicious lesions detected (primary examination).

RESULTS

EGCs were found in 44 (1.9%) and 53 (2.3%; p=0.412) patients in the WLI and 2G-NBI groups, respectively, during primary EGD. In a post hoc analysis, the overall rate of lesions detected at the second examination was 25% (n=36/145), with no significant differences between groups. PPV for EGC in suspicious lesions was 13.5% and 20.9% in the WLI (50/371 target lesions) and 2G-NBI groups (59/282 target lesions), respectively (p=0.015).

CONCLUSION

The overall sensitivity of primary endoscopy for the detection of EGC in high-risk patients was only 75% and should be improved. 2G-NBI did not increase EGC detection rate over conventional WLI. The impact of a slightly better PPV of 2G-NBI has to be evaluated further.

TRIAL REGISTRATION NUMBER

UMIN000014503.

Linked Color Imaging Focused on Neoplasm Detection in the Upper Gastrointestinal Tract : A Randomized Trial

BACKGROUND

Linked color imaging (LCI) is a new image-enhanced endoscopy technique that allows users to recognize slight differences in mucosal color.

OBJECTIVE

To compare the performance of LCI with white light imaging (WLI) in detecting neoplastic lesions in the upper gastrointestinal tract.

DESIGN

A controlled, multicenter trial with randomization using minimization. (University Hospital Medical Information Network Clinical Trials Registry: UMIN000023863).

SETTING

16 university hospitals and 3 tertiary care hospitals in Japan.

PATIENTS

1502 patients with known previous or current cancer of the gastrointestinal tract and undergoing surveillance for gastrointestinal cancer.

INTERVENTION

WLI followed by LCI examination (WLI group) or LCI followed by WLI examination (LCI group).

MEASUREMENTS

Diagnosis of 1 or more neoplastic lesions in the pharynx, esophagus, or stomach in the first examination (primary outcome) and 1 or more neoplastic lesions overlooked in the first examination (secondary outcome).

RESULTS

752 patients were assigned to the WLI group and 750 to the LCI group. The percentage of patients with 1 or more neoplastic lesions diagnosed in the first examination was higher with LCI than with WLI (60 of 750 patients or 8.0% [95% CI, 6.2% to 10.2%] vs. 36 of 752 patients or 4.8% [CI, 3.4% to 6.6%]; risk ratio, 1.67 [CI, 1.12 to 2.50; P = 0.011]). The proportion with overlooked neoplasms was lower in the LCI group than in the WLI group (5 of 750 patients or 0.67% [CI, 0.2% to 1.6%] vs. 26 of 752 patients or 3.5% [CI, 2.3% to 5.0%]; risk ratio, 0.19 [CI, 0.07 to 0.50]).

LIMITATION

Endoscopists were not blinded.

CONCLUSION

LCI is more effective than WLI for detecting neoplastic lesions in the pharynx, esophagus, and stomach.

PRIMARY FUNDING SOURCE

Fujifilm Corporation.

NBI utility in oncologic surgery: An organ by organ review

The main aims of the oncologic surgeon should be an early tumor diagnosis, complete surgical resection, and a careful post-treatment follow-up to ensure a prompt diagnosis of recurrence. Radiologic and endoscopic methods have been traditionally used for these purposes, but their accuracy might sometimes be suboptimal. Technological improvements could help the clinician during the diagnostic and therapeutic management of tumors. Narrow band imaging (NBI) belongs to optical image techniques, and uses light characteristics to enhance tissue vascularization. Because neoangiogenesis is a fundamental step during carcinogenesis, NBI could be useful in the diagnostic and therapeutic workup of tumors. Since its introduction in 2001, NBI use has rapidly spread in different oncologic specialties with clear advantages. There is an active interest in this topic as demonstrated by the thriving literature. It is unavoidable for clinicians to gain in-depth knowledge about the application of NBI to their specific field, losing the overall view on the topic. However, by looking at other fields of application, clinicians could find ideas to improve NBI use in their own specialty. The aim of this review is to summarize the existing literature on NBI use in oncology, with the aim of providing the state of the art: we present an overview on NBI fields of application, results, and possible future improvements in the different specialties.

Near-focus magnification and second-generation narrow-band imaging for early gastric cancer in a randomized trial

BACKGROUND

Magnifying endoscopy with narrow-band imaging (NBI) is effective for the diagnosis of early gastric cancer (EGC). However, magnifying endoscopy is not yet popular globally because of the required level of skill and lack of availability. To overcome these problems, dual-focus endoscopy (standard- and near-focus (NF) modes) has been developed. In this study, we evaluated the diagnostic performance of NF with second-generation (2G)-NBI (NF-NBI) for the diagnosis of EGC.

METHODS

This was a secondary analysis of a multicenter randomized controlled trial of 4523 high-risk patients who underwent gastroscopies at 13 institutions in Japan. Patients were randomly assigned to white-light imaging (WLI) followed by 2G-NBI or to 2G-NBI followed by WLI. Lesions suspicious for EGC, newly detected by non-magnifying WLI or 2G-NBI, were subsequently observed with NF-NBI. All detected lesions were biopsied or resected. The diagnostic performance of NF-NBI was compared with the final histology.

RESULTS

A total of 870 detected lesions (145 EGC, 725 non-EGC) were analyzed. Overall diagnostic performance for EGC using NF-NBI was accuracy 87.7%, sensitivity 60.7%, specificity 93.1%, positive predictive value 63.8%, and negative predictive value 92.2%. There were no significant differences in diagnostic performance between lesions detected by WLI or 2G-NBI. For lesions diagnosed with high (333 lesions) and low (537 lesions) confidences, accuracy was 92.2% and 84.9%, sensitivity was 64.7% and 58.5%, and specificity was 90.5% and 88.8%, respectively.

CONCLUSION

The diagnostic performance of NF-NBI is good and acceptable for diagnosis of EGC in combination with either WLI or 2G-NBI.

Image-enhanced endoscopy for gastric preneoplastic conditions and neoplastic lesions: a systematic review and meta-analysis

BACKGROUND : Image-enhanced endoscopy (IEE) improves the accuracy of endoscopic diagnosis. We aimed to assess the value of IEE for gastric preneoplastic conditions and neoplastic lesions. METHODS : Medline and Embase were searched until December 2018. Studies allowing calculation of diagnostic measures were included. Risk of bias and applicability were assessed using QUADAS-2. Subgroup analysis was performed to explore heterogeneity. RESULTS : 44 studies met the inclusion criteria. For gastric intestinal metaplasia (GIM), narrow-band imaging (NBI) obtained a pooled sensitivity and specificity of 0.79 (95 %CI 0.72-0.85) and 0.91 (95 %CI 0.88-0.94) on per-patient basis; on per-biopsy basis, it was 0.84 (95 %CI 0.81-0.86) and 0.95 (95 %CI 0.94-0.96), respectively. Tubulovillous pattern was the most accurate marker to detect GIM and it was effectively assessed without high magnification. For dysplasia, NBI showed a pooled sensitivity and specificity of 0.87 (95 %CI 0.84-0.89) and 0.97 (95 %CI 0.97-0.98) on per-biopsy basis. The use of magnification improved the performance of NBI to characterize early gastric cancer (EGC), especially when the vessel plus surface (VS) classification was applied. Regarding other technologies, trimodal imaging also obtained a high accuracy for dysplasia (sensitivity 0.93 [95 %CI 0.85-0.98], specificity 0.98 [95 %CI 0.92-1.00]). For atrophic gastritis, no specific pattern was noted and none of the technologies reached good diagnostic yield. CONCLUSION : NBI is highly accurate for GIM and dysplasia. The presence of tubulovillous pattern and the VS classification seem to be useful to detect GIM and characterize EGC, respectively. These features should be used in current practice and to standardize endoscopic criteria for other technologies.

Development and validation of a real-time artificial intelligence-assisted system for detecting early gastric cancer: A multicentre retrospective diagnostic study

Background

We aimed to develop and validate a real-time deep convolutional neural networks (DCNNs) system for detecting early gastric cancer (EGC).

Methods

All 45,240 endoscopic images from 1364 patients were divided into a training dataset (35823 images from 1085 patients) and a validation dataset (9417 images from 279 patients). Another 1514 images from three other hospitals were used as external validation. We compared the diagnostic performance of the DCNN system with endoscopists, and then evaluated the performance of endoscopists with or without referring to the system. Thereafter, we evaluated the diagnostic ability of the DCNN system in video streams. The accuracy, sensitivity, specificity, positive predictive value, negative predictive value and Cohen's kappa coefficient were measured to assess the detection performance.

Finding

The DCNN system showed good performance in EGC detection in validation datasets, with accuracy (85.1%–91.2%), sensitivity (85.9%–95.5%), specificity (81.7%–90.3%), and AUC (0.887–0.940). The DCNN system showed better diagnostic performance than endoscopists and improved the performance of endoscopists. The DCNN system was able to process oesophagogastroduodenoscopy (OGD) video streams to detect EGC lesions in real time.

Interpretation

We developed a real-time DCNN system for EGC detection with high accuracy and stability. Multicentre prospective validation is needed to acquire high-level evidence for its clinical application.

Funding

This work was supported by the National Natural Science Foundation of China (grant nos. 81672935 and 81871947), Jiangsu Clinical Medical Center of Digestive System Diseases and Gastrointestinal Cancer (grant no. YXZXB2016002), and Nanjing Science and Technology Development Foundation (grant no. 2017sb332019).

Histological Architecture of Gastric Epithelial Neoplasias That Showed Absent Microsurface Patterns, Visualized by Magnifying Endoscopy with Narrow-Band Imaging

BACKGROUND/AIMS

The objective of this study was to elucidate the histological structure of the absent microsurface patterns (MSPs) that were visualized by magnifying endoscopy with narrow-band imaging (M-NBI).

METHODS

The study included consecutive gastric epithelial neoplasias for which M-NBI findings and histological findings could be compared on a one-to-one basis. The lesions were classified as absent MSPs and present MSPs based on the findings obtained using M-NBI. Of the histopathological findings for each lesion that corresponded to M-NBI findings, crypt opening densities, crypt lengths, crypt opening diameters, intercrypt distances, and crypt angles were measured and compared.

RESULTS

Thirty-six lesions were included in the analysis; of these, 17 lesions exhibited absent MSP and 19 lesions exhibited present MSP. Comparing the histological measurements for absent MSPs vs. present MSPs, median crypt opening density was 0.9 crypt openings/mm vs. 4.8 crypt openings/mm (p<0.001), respectively. The median crypt length, median crypt opening diameter, median intercrypt distance, and median crypt angle were 80.0 μm vs. 160 μm (p<0.001), 40.0 μm vs. 44.2 μm (p=0.09), 572.5 μm vs. 166.7 μm (p<0.001), and 21.6 degrees vs. 15.5 degrees (p<0.001), respectively.

CONCLUSION

Histological findings showed that lesions exhibiting absent MSPs had lower crypt opening density, shorter crypt length, greater intercrypt distance, and larger crypt angle.

Narrowband imaging with near-focus magnification for discriminating the gastric tumor margin before endoscopic resection: A prospective randomized multicenter trial

BACKGROUND AND AIM

This study investigated the usefulness of near-focus narrowband imaging (NF-NBI) for determining gastric tumor margins compared with indigo carmine chromoendoscopy (ICC) before endoscopic submucosal dissection (ESD).

METHODS

This prospective randomized controlled trial was conducted at seven teaching hospitals in Korea. Patients with gastric adenoma or differentiated adenocarcinoma undergoing ESD were enrolled and randomly assigned to the NF-NBI or ICC group. A marking dot was placed on the most proximal margin of the tumor before ESD. The primary endpoint was delineation accuracy, which was defined as presence of marking dots within 1 mm of the tumor margin under microscopic observation.

RESULTS

A total of 200 patients in the NF-NBI group and 195 patients in the ICC group were included. The delineation accuracy rate was 84.5% in the NF-NBI group and 81.0% in the ICC group (P = 0.44). However, the distance from the marking dot to the margin of the tumor was significantly shorter in the NF-NBI group than in the ICC group (0.8 ± 0.8 vs 1.2 ± 1.3 mm, P < 0.01). Even after adjustment of other clinicopathological factors that are associated with difficulty of tumor delineation, NF-NBI did not show significant association with accurate delineation (odds ratio of 0.86, P = 0.60).

CONCLUSIONS

This prospective multicenter study showed that NF-NBI is not superior to ICC in terms of accurately delineating gastric tumors (NCT02661945).

Three-dimensional flexible endoscopy enables more accurate endoscopic recognition and endoscopic submucosal dissection marking for superficial gastric neoplasia: a pilot study to compare two- and three-dimensional imaging

BACKGROUND

Three-dimensional (3D) visualization offers better depth recognition than two-dimensional (2D) imaging, thus helping to provide more useful information. We compared 3D and 2D endoscopy with regard to endoscopic recognition and endoscopic submucosal dissection (ESD) marking for superficial gastric neoplasia.

METHODS

ESD marking was performed on half of a neoplasia margin under 2D observation and the on other half under 3D observation for 28 gastric lesions (26 early gastric cancers and 2 adenomas). The accuracy of ESD marking was evaluated based on the distance between the pathological and endoscopic neoplasia margins measured on histology sections of ESD specimens. The technical ease of ESD marking and endoscopic lesion recognition (lesion morphology, lesion extent, and comprehensive endoscopic cognition) were assessed using visual analog scale (VAS) questionnaires.

RESULTS

The mean distance between the pathological and endoscopic margins under 3D observation (1.03 ± 0.80 mm) was significantly (p = 0.002) shorter than that under 2D observation (1.94 ± 1.96 mm). The VAS for technical ease of ESD marking under 3D observation was significantly better (p < 0.01) than that under 2D observation. The VAS for all aspects of endoscopic recognition under 3D observation was significantly better (p < 0.01) than under 2D observation.

CONCLUSIONS

3D flexible endoscopy achieved more accurate endoscopic recognition and ESD marking for superficial gastric neoplasia than a 2D approach in a clinical setting of ESD.

Diagnosis of early gastric cancer using image enhanced endoscopy: a systematic approach

This paper provides an overview of the principles of a vessel plus surface (VS) classification system to explain the diagnostic system of early gastric cancer using image-enhanced magnifying endoscopy. Furthermore, this paper introduces the magnifying endoscopy simple diagnostic algorithm for gastric cancer (MEADA-G) developed according to the VS classification system, with a description of the procedures performed for diagnosis. In addition to the diagnostic system, white opaque substance (WOS), light blue crest (LBC), white globe appearance (WGA), and vessels within epithelial circle (VEC) patterns, which are representative findings that can be observed in the gastric mucosa by image-enhanced magnifying endoscopy, are also described. Image-enhanced magnifying endoscopy is particularly useful in the diagnosis of differentiated-type early gastric cancer. It is important to use the appropriate clinical strategies based on a comprehensive understanding of the usefulness and limitations of the diagnostic system described in this paper.

Diagnostic limitations of magnifying endoscopy with narrow-band imaging in early gastric cancer

Background and study aims  Magnifying endoscopy with narrow band imaging (M-NBI) has made a huge contribution to endoscopic diagnosis of early gastric cancer (EGC). However, we sometimes encountered false-negative cases with M-NBI diagnosis (i. e., M-NBI diagnostic limitation lesion: M-NBI-DLL). However, clinicopathological features of M-NBI-DLLs have not been well elucidated. We aimed to clarify the clinicopathological features and histological reasons of M-NBI-DLLs. Patients and methods  In this single-center retrospective study, M-NBI-DLLs were extracted from 456 EGCs resected endoscopically at our hospital. We defined histological types of M-NBI-DLLs and analyzed clinicopathologically to clarify histological reasons of M-NBI-DLLs. Results  Of 456 EGCs, 48 lesions (10.5 %) of M-NBI-DLLs were enrolled. M-NBI-DLLs was classified into four histological types as follows: gastric adenocarcinoma of fundic-gland type (GA-FG, n = 25), gastric adenocarcinoma of fundic-gland mucosal type (GA-FGM, n = 1), differentiated adenocarcinoma (n = 14), and undifferentiated adenocarcinoma (n = 8). Thirty-nine lesions of M-NBI-DLLs were H. pylori -negative gastric cancers (39/47, 82.9 %). Histological reasons for M-NBI-DLLs were as follows: 1) completely covered with non-neoplastic mucosa (25/25 GA-FG, 8/8 undifferentiated adenocarcinoma); 2) well-differentiated adenocarcinoma with low-grade atypia (1/1 GA-FGM, 14/14 differentiated adenocarcinoma); 3) similarity of surface structure (10/14 differentiated adenocarcinoma); and 4) partially covered and/or mixed with a non-neoplastic mucosa (1/1 GA-FGM, 6/14 differentiated adenocarcinoma). Conclusions  Diagnostic limitations of M-NBI depend on four distinct histological characteristics. For accurate diagnosis of M-NBI-DLLs, it may be necessary to fully understand endoscopic features of these lesions using white light imaging and M-NBI based on these histological characteristics and to take a precise biopsy.

Performance of a computer-aided diagnosis system in diagnosing early gastric cancer using magnifying endoscopy videos with narrow-band imaging (with videos)

BACKGROUND AND AIMS

The performance of magnifying endoscopy with narrow-band imaging (ME-NBI) using a computer-aided diagnosis (CAD) system in diagnosing early gastric cancer (EGC) is unclear. Here, we aimed to clarify the differences in the diagnostic performance between expert endoscopists and the CAD system using ME-NBI.

METHODS

The CAD system was pretrained using 1492 cancerous and 1078 noncancerous images obtained using ME-NBI. One hundred seventy-four videos (87 cancerous and 87 noncancerous videos) were used to evaluate the diagnostic performance of the CAD system using the area under the curve (AUC), accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). For each item, comparisons were made between the CAD system and 11 experts who were skilled in diagnosing EGC using ME-NBI with clinical experience of more than 1 year at our hospital.

RESULTS

The CAD system demonstrated an AUC of 0.8684. The accuracy, sensitivity, specificity, PPV, and NPV were 85.1% (95% confidence interval [95% CI], 79.0-89.6), 87.4% (95% CI, 78.8-92.8), 82.8% (95% CI, 73.5-89.3), 83.5% (95% CI, 74.6-89.7), and 86.7% (95% CI, 77.8-92.4), respectively. The CAD system was significantly more accurate than 2 experts, significantly less accurate than 1 expert, and not significantly different from the remaining 8 experts.

CONCLUSIONS

The overall performance of the CAD system using ME-NBI videos in diagnosing EGC was considered good and was equivalent to or better than that of several experts. The CAD system may prove useful in the diagnosis of EGC in clinical practice.

A new exploration of white globe appearance (WGA) in ulcerative lesions

AIM

 White globe appearance (WGA), a small white lesion with a globular shape that can be clearly visualized by magnifying endoscopy with narrow-band imaging (ME-NBI), was reported to be a reliable marker of early gastric cancer (EGC). However, we found that this endoscopic presentation could also be seen in non-cancerous tissues, especially in ulcerative lesions. This study aimed to further investigate the diagnostic value of WGA in differentiating non-cancerous lesions from EGC in ulcer-type cases.

MATERIALS AND METHODS

 We retrospectively reviewed 54 cases of EGC and 155 cases of non-cancerous lesions in this study, all of which had endoscopic imaging data of ME-NBI scanning and pathological data of biopsy or resected specimens. The correlation of the prevalence of WGA and ulcerative lesions, as well as the characteristics of WGA between the 2 groups were analyzed in this study.

RESULTS

 WGA was more common in ulcerative lesions (27.6 %, 21/76) than in non-ulcerative lesions (3.8 %, 5/133) (p < 0.001) in our study. In the ulcerative cases, no significant difference in prevalence of WGA was observed between EGC and non-cancerous lesions (p = 0.532). Compared with WGA in EGC, WGA in non-cancerous lesions tended to show the characteristic of tree-branch-like vessels on globular shape (p < 0.001).

CONCLUSIONS

 WGA is more likely to occur in ulcerative lesions, and the presence of WGA alone cannot distinguish EGC from non-cancerous lesions in ulcer-type cases. In WGA-positive tissue, tree-branch-like vessels of globular shape may provide a certain clinical value in diagnosis of non-cancerous lesions or EGC.

Recent advances in the detection and management of early gastric cancer and its precursors

Despite declines in incidence, gastric cancer remains a disease with a poor prognosis and limited treatment options due to its often late stage of diagnosis. In contrast, early gastric cancer has a good to excellent prognosis, with 5-year survival rates as high as 92.6% after endoscopic resection. There remains an East-West divide for this disease, with high incidence countries such as Japan seeing earlier diagnoses and reduced mortality, in part thanks to the success of a national screening programme. With missed cancers still prevalent at upper endoscopy in the West, and variable approaches to assessment of the high-risk stomach, the quality of endoscopy we provide must be a focus for improvement, with particular attention paid to the minority of patients at increased cancer risk. High-definition endoscopy with virtual chromoendoscopy is superior to white light endoscopy alone. These enhanced imaging modalities allow the experienced endoscopist to accurately and robustly detect high-risk lesions in the stomach. An endoscopy-led staging strategy would mean biopsies could be targeted to histologically confirm the endoscopic impression of premalignant lesions including atrophic gastritis, gastric intestinal metaplasia, dysplasia and early cancer. This approach to quality improvement will reduce missed diagnoses and, combined with the latest endoscopic resection techniques performed at expert centres, will improve early detection and ultimately patient outcomes. In this review, we outline the latest evidence relating to diagnosis, staging and treatment of early gastric cancer and its precursor lesions.

Clinical applicability of gastroscopy with narrow-band imaging for the diagnosis of Helicobacter pylori gastritis, precancerous gastric lesion, and neoplasia

Premalignant gastric lesions such as atrophic gastritis and intestinal metaplasia frequently occur in subjects with long-term Helicobacter pylori (H. pylori) infection. The regular arrangement of collecting venules (RAC) is seen in the normal gastric corpus, whereas mucosal swelling and redness without RAC are observed in H. pylori-infected mucosa. Despite successful H. pylori eradication, the presence of atrophic gastritis and/or gastric intestinal metaplasia (GIM) is a risk factor for gastric cancer. With the development of advanced imaging technologies, recent studies have reported the usefulness of narrow-band imaging (NBI) for endoscopic diagnosis of atrophic gastritis and GIM. Using NBI endoscopy with magnification (M-NBI), atrophic gastritis is presented as irregular coiled microvessels and loss of gastric pits. Typical M-NBI endoscopic findings of GIM are a light blue crest and a white opaque substance. Based on the microvascular patterns, fine network, core vascular, and unclear patterns are useful for predicting gastric dysplasia in polypoid lesions. For diagnosis of early gastric cancer (EGC), a systematic classification using M-NBI endoscopy has been proposed on the basis of the presence of a demarcation line and an irregular microvascular/microsurface pattern. Furthermore, M-NBI endoscopy has been found to be more accurate for determining the horizontal margin of EGC compared to conventional endoscopy. In this review, we present up-to-date results on the clinical usefulness of gastroscopy with NBI for the diagnosis of H. pylori gastritis, precancerous gastric lesion, and neoplasia.

Helicobacter pylori-induced inflammation masks the underlying presence of low-grade dysplasia on gastric lesions

BACKGROUND

Helicobacter pylori (H. pylori) infection has been associated with a long-term risk of precancerous gastric conditions (PGC) even after H. pylori eradication.

AIM

To investigate the efficacy of High-Resolution White-Light Endoscopy with Narrow-Band Imaging in detecting PGC, before/after H. pylori eradication.

METHODS

We studied 85 consecutive patients with H. pylori-related gastritis with/without PGC before and 6 mo after proven H. pylori eradication. Kimura-Takemoto modified and endoscopic grading of gastric intestinal metaplasia classifications, were applied to assess the endoscopic extension of atrophy and intestinal metaplasia. The histological result was considered to be the gold standard. The Sydney System, the Operative-Link on Gastritis-Assessment, and the Operative-Link on Gastric-Intestinal Metaplasia were used for defining histological gastritis, atrophy and intestinal metaplasia, whereas dysplasia was graded according to World Health Organization classification. Serum anti-parietal cell antibody and anti-intrinsic factor were measured when autoimmune atrophic gastritis was suspected.

RESULTS

After H. pylori eradication histological signs of mononuclear/polymorphonuclear cell infiltration and Mucosal Associated Lymphoid Tissue-hyperplasia, disappeared or decreased in 100% and 96.5% of patients respectively, whereas the Operative-Link on Gastritis-Assessment and Operative-Link on Gastric-Intestinal Metaplasia stages did not change. Low-Grade Dysplasia prevalence was similar on random biopsies before and after H. pylori eradication (17.6% vs 10.6%, P = 0.19), but increased in patients with visible lesions (0% vs 22.4%, P < 0.0001). At a multivariate analysis, the probability for detecting dysplasia after resolution of H. pylori-related active inflammation was higher in patients with regression or reduction of Mucosal Associated Lymphoid Tissue hyperplasia, greater alcohol consumption, and anti-parietal cell antibody and/or anti-intrinsic factor positivity [odds ratio (OR) = 3.88, 95% confidence interval (CI): 1.31-11.49, P = 0.01; OR = 3.10, 95%CI: 1.05-9.12, P = 0.04 and OR = 5.47, 95%CI: 1.33-22.39, P < 0.04, respectively].

CONCLUSION

High-Resolution White-Light Endoscopy with Narrow-Band Imaging allows an accurate diagnosis of Low-Grade Dysplasia on visible lesions after regression of H. pylori-induced chronic gastritis. Patients with an overlap between autoimmune/H. pylori-induced gastritis may require more extensive gastric mapping.

Guidelines for endoscopic diagnosis of early gastric cancer

The Japan Gastroenterological Endoscopy Society developed the Guideline for Endoscopic Diagnosis of Early Gastric Cancer based on scientific methods. Endoscopy for the diagnosis of early gastric cancer has been acknowledged as a useful and highly precise examination, and its use has become increasingly more common in recent years. However, the level of evidence in this field is low, and it is often necessary to determine recommendations based on expert consensus only. This clinical practice guideline consists of the following sections to provide the current guideline: [I] Risk stratification of gastric cancer before endoscopic examination, [II] Detection of early gastric cancer, [III] Qualitative diagnosis of early gastric cancer, [IV] Diagnosis to choose the therapeutic strategy for gastric cancer, [V] Risk stratification after endoscopic examination, and [VI] Surveillance of early gastric cancer.

Diagnostics of stomach cancer in Russia: fi rst results of the multicenter study “RADIUS” (Early diagnosis of stomach cancer in dyspepsia)

The aim of the study was to evaluate the timeliness and eff ectiveness of the diagnosis of precancerous diseases and early forms of gastric cancer in primary health care. Materials and methods. A multicenter retrospective descriptive study was conducted with an analysis of the medical records of 426 patients who were diagnosed with malignant neoplasm of the stomach at the time of going to outpatient facilities. We used data from 125 outpatient facilities in 7 centers for the period from 2009 to 2019. Results. The average age of patients with a fi rst established diagnosis of malignant neoplasia was 61,7 (95% CI 58.6 ÷ 64.8) years. The prevailing form of malignant neoplastic disease was the option “Adenocarcinoma, intestinal cancer” — 77,7%, diff use type cancer — 12,7%, other histological types — 9,6%. The diagnosis of malignant neoplastic disease at the III and IV stages of the disease was fi rst established in 67,4% of patients. The number of years lived by patients with a diagnosis of gastric gastric cancer was 2,0 years. Over a 10-year follow-up, 75% of patients died. Risk factors for gastric cancer were identifi ed in 41% of respondents. Signs of dyspepsia syndrome were observed in 31,5% of patients, and they appeared on average 4,6 (95% CI 4,4 ÷ 4,8) years before the diagnosis of gastric malignancy. Among people with gastric cancer, dyspepsia symptoms occur every day in every third patient. The features of the manifestations of dyspepsia in the observed patients were: daily manifestations of symptoms (33,3%), symptoms at night — 9,4%, an increase in the intensity of dyspepsia (21,3%), the lack of a clinical response to the therapy (proton pump inhibitors, prokinetics, eradication of Helicobacter pylori infection) in 58% of cases. Anxiety symptoms were recorded 2,4 years before a diagnosis of stomach cancer.

Clinical Practice Guideline for Endoscopic Resection of Early Gastrointestinal Cancer

Although surgery was the standard treatment for early gastrointestinal cancers, endoscopic resection is now a standard treatment for early gastrointestinal cancers without regional lymph node metastasis. High-definition white light endoscopy, chromoendoscopy, and image-enhanced endoscopy such as narrow band imaging are performed to assess the edge and depth of early gastrointestinal cancers for delineation of resection boundaries and prediction of the possibility of lymph node metastasis before the decision of endoscopic resection. Endoscopic mucosal resection and/or endoscopic submucosal dissection can be performed to remove early gastrointestinal cancers completely by en bloc fashion. Histopathological evaluation should be carefully made to investigate the presence of risk factors for lymph node metastasis such as depth of cancer invasion and lymphovascular invasion. Additional treatment such as radical surgery with regional lymphadenectomy should be considered if the endoscopically resected specimen shows risk factors for lymph node metastasis. This is the first Korean clinical practice guideline for endoscopic resection of early gastrointestinal cancer. This guideline was developed by using mainly de novo methods and encompasses endoscopic management of superficial esophageal squamous cell carcinoma, early gastric cancer, and early colorectal cancer. This guideline will be revised as new data on early gastrointestinal cancer are collected.

Convolutional Neural Network for Differentiating Gastric Cancer from Gastritis Using Magnified Endoscopy with Narrow Band Imaging

BACKGROUND

Early detection of early gastric cancer (EGC) allows for less invasive cancer treatment. However, differentiating EGC from gastritis remains challenging. Although magnifying endoscopy with narrow band imaging (ME-NBI) is useful for differentiating EGC from gastritis, this skill takes substantial effort. Since the development of the ability to convolve the image while maintaining the characteristics of the input image (convolution neural network: CNN), allowing the classification of the input image (CNN system), the image recognition ability of CNN has dramatically improved.

AIMS

To explore the diagnostic ability of the CNN system with ME-NBI for differentiating between EGC and gastritis.

METHODS

A 22-layer CNN system was pre-trained using 1492 EGC and 1078 gastritis images from ME-NBI. A separate test data set (151 EGC and 107 gastritis images based on ME-NBI) was used to evaluate the diagnostic ability [accuracy, sensitivity, positive predictive value (PPV), and negative predictive value (NPV)] of the CNN system.

RESULTS

The accuracy of the CNN system with ME-NBI images was 85.3%, with 220 of the 258 images being correctly diagnosed. The method's sensitivity, specificity, PPV, and NPV were 95.4%, 71.0%, 82.3%, and 91.7%, respectively. Seven of the 151 EGC images were recognized as gastritis, whereas 31 of the 107 gastritis images were recognized as EGC. The overall test speed was 51.83 images/s (0.02 s/image).

CONCLUSIONS

The CNN system with ME-NBI can differentiate between EGC and gastritis in a short time with high sensitivity and NPV. Thus, the CNN system may complement current clinical practice of diagnosis with ME-NBI.

Clinical Practice Guideline for Endoscopic Resection of Early Gastrointestinal Cancer

Although surgery was the standard treatment for early gastrointestinal cancers, endoscopic resection is now a standard treatment for early gastrointestinal cancers without regional lymph node metastasis. High-definition white light endoscopy, chromoendoscopy, and image-enhanced endoscopy such as narrow band imaging are performed to assess the edge and depth of early gastrointestinal cancers for delineation of resection boundaries and prediction of the possibility of lymph node metastasis before the decision of endoscopic resection. Endoscopic mucosal resection and/or endoscopic submucosal dissection can be performed to remove early gastrointestinal cancers completely by <i>en bloc</i> fashion. Histopathological evaluation should be carefully made to investigate the presence of risk factors for lymph node metastasis such as depth of cancer invasion and lymphovascular invasion. Additional treatment such as radical surgery with regional lymphadenectomy should be considered if the endoscopically resected specimen shows risk factors for lymph node metastasis. This is the first Korean clinical practice guideline for endoscopic resection of early gastrointestinal cancer. This guideline was developed by using mainly de novo methods and encompasses endoscopic management of superficial esophageal squamous cell carcinoma, early gastric cancer, and early colorectal cancer. This guideline will be revised as new data on early gastrointestinal cancer are collected.

Esophageal regenerative therapy using cell sheet technology

We have been conducting research on esophageal regenerative therapy using cell sheet technology. In particular, in the endoscopic field, we have pushed forward clinical research on endoscopic transplantation of cultured autologous oral mucosal epithelial cell sheets to esophageal ulcer after endoscopic submucosal dissection (ESD). We started research in this direction using animal models in 2004 and performed clinical research in 2012 in collaboration with Nagasaki University and Karolinska Institute. Although in full-circumferential cases it was difficult to prevent esophageal stricture after ESD, there were no complications and stricture could be suppressed. The cell sheet technology is still in its infancy. However, we are convinced that it has a high potential for application in various areas of gastrointestinal science. In this review, we focus on the pre-clinical and clinical trial results obtained and on the theoretical aspects of (1) stricture prevention, (2) esophageal tissue engineering research, and (3) endoscopic transplantation, and review the esophageal regenerative therapy by cell sheet technology.

Additive Effect of Magnifying Endoscopy with Narrow-Band Imaging for Diagnosing Mixed-Type Early Gastric Cancers

BACKGROUND

Pretreatment biopsy may not correctly diagnose mixed-type early gastric cancers. Despite reports on the usefulness of magnifying endoscopy with narrow-band imaging in diagnosing early gastric cancers, no reports exist on differences in magnifying endoscopy with narrow-band imaging findings between differentiated-type-predominant mixed-type and undifferentiated-type-predominant mixed-type early gastric cancers.

AIM

This study aimed to clarify differences in magnifying endoscopy with narrow-band imaging findings and investigate the additive effect of combining magnifying endoscopy with narrow-band imaging and biopsy findings for pretreatment histological-type diagnosis.

METHODS

Patients undergoing endoscopic submucosal dissection as initial treatment between April 2005 and March 2017 participated in this retrospective study. There were 156 differentiated-type-predominant mixed-type and 36 undifferentiated-type-predominant mixed-type lesions. We extracted the most significant magnifying endoscopy with narrow-band imaging findings of differentiated-type-predominant mixed-type and undifferentiated-type-predominant mixed-type lesions using multivariate analysis and compared the accuracy, sensitivity, and specificity between pretreatment biopsy alone and a combination of biopsy and magnifying endoscopy with narrow-band imaging findings.

RESULTS

Significant magnifying endoscopy with narrow-band imaging findings was fine network pattern in differentiated-type-predominant and corkscrew pattern in undifferentiated-type-predominant mixed-type lesions. Accuracy, sensitivity, and specificity were significantly higher with combined biopsy and magnifying endoscopy with narrow-band imaging findings than with pretreatment biopsy alone.

CONCLUSIONS

The study results demonstrated the additive effect of magnifying endoscopy with narrow-band imaging with biopsy for diagnosing mixed-type early gastric cancers. This study may be beneficial in routine practice because it indicates a possibility of reducing additional surgery after endoscopic submucosal resection because of incorrect diagnosis of histological type.

Objective Assessment of the Utility of Chromoendoscopy with a Support Vector Machine

Purpose

The utility of chromoendoscopy for early gastric cancer (GC) was determined by machine learning using data of color differences.

Methods

Eighteen histopathologically confirmed early GC lesions were examined. We prepared images from white light endoscopy (WL), indigo carmine (Indigo), and acetic acid-indigo carmine chromoendoscopy (AIM). A border between cancerous and non-cancerous areas on endoscopic images was established from post-treatment pathological findings, and 2000 pixels with equivalent luminance values were randomly extracted from each image of cancerous and non-cancerous areas. Each pixel was represented as a three-dimensional vector with RGB values and defined as a sample. We evaluated the Mahalanobis distance using RGB values, indicative of color differences between cancerous and non-cancerous areas. We then conducted diagnosis test using a support vector machine (SVM) for each image. SVM was trained using the 100 training samples per class and determined which area each of 1900 test samples per class came from.

Results

The means of the Mahalanobis distances for WL, Indigo, and AIM were 1.52, 1.32, and 2.53, respectively and there were no significant differences in the three modalities. Diagnosability per endoscopy technique was assessed using the F1 measure. The means of F1 measures for WL, Indigo, and AIM were 0.636, 0.618, and 0.687, respectively. AIM images were better than WL and Indigo images for the diagnosis of GC.

Conclusion

Objective assessment by SVM found AIM to be suitable for diagnosis of early GC based on color differences.

Convolutional neural network for the diagnosis of early gastric cancer based on magnifying narrow band imaging

BACKGROUND

Magnifying endoscopy with narrow band imaging (M-NBI) has been applied to examine early gastric cancer by observing microvascular architecture and microsurface structure of gastric mucosal lesions. However, the diagnostic efficacy of non-experts in differentiating early gastric cancer from non-cancerous lesions by M-NBI remained far from satisfactory. In this study, we developed a new system based on convolutional neural network (CNN) to analyze gastric mucosal lesions observed by M-NBI.

METHODS

A total of 386 images of non-cancerous lesions and 1702 images of early gastric cancer were collected to train and establish a CNN model (Inception-v3). Then a total of 341 endoscopic images (171 non-cancerous lesions and 170 early gastric cancer) were selected to evaluate the diagnostic capabilities of CNN and endoscopists. Primary outcome measures included diagnostic accuracy, sensitivity, specificity, positive predictive value, and negative predictive value.

RESULTS

The sensitivity, specificity, and accuracy of CNN system in the diagnosis of early gastric cancer were 91.18%, 90.64%, and 90.91%, respectively. No significant difference was spotted in the specificity and accuracy of diagnosis between CNN and experts. However, the diagnostic sensitivity of CNN was significantly higher than that of the experts. Furthermore, the diagnostic sensitivity, specificity and accuracy of CNN were significantly higher than those of the non-experts.

CONCLUSIONS

Our CNN system showed high accuracy, sensitivity and specificity in the diagnosis of early gastric cancer. It is anticipated that more progress will be made in optimization of the CNN diagnostic system and further development of artificial intelligence in the medical field.

Optimal number of endoscopic biopsies for diagnosis of early gastric cancer

Background and study aims  No recommendations are available for optimal number of endoscopic biopsies for early gastric cancer (GC), and whether detection of early GC is improved by increasing the number of biopsy is unclear. We therefore evaluated the relationship between number of biopsies and diagnostic accuracy. Materials and methods  We retrospectively evaluated 858 early GCs (623 from endoscopic submucosal dissection and 235 surgical specimens), which we classified as obtained after one, two, or three or more biopsies. We assessed diagnostic accuracy by number of biopsies, and in subgroups by tumor diameter, gross type, and surface color. Results  Almost half the lesions were obtained after one biopsy each, 30 % after two biopsies, and 20 % after three or more biopsies. Although diagnostic accuracy increased with biopsy number, it was significantly greater for the two-biopsy group than the one-biopsy group, (92.5 % vs. 83.9 %, P  = 0.0009), but did not significantly differ between the two- and three or more-biopsy groups. This finding was seen when tumors were evaluated by size, but not by elevated type and surface color, for which more biopsies did not improve diagnostic accuracy. Multivariate analysis demonstrated that two or more biopsies was the independent significant factors for diagnostic accuracy. Conclusions  Two biopsies are the optimal number required to diagnose early GC.

Automatic detection of early gastric cancer in endoscopic images using a transferring convolutional neural network

Endoscopic image diagnosis assisted by machine learning is useful for reducing misdetection and interobserver variability. Although many results have been reported, few effective methods are available to automatically detect early gastric cancer. Early gastric cancer have poor morphological features, which implies that automatic detection methods can be extremely difficult to construct. In this study, we proposed a convolutional neural network-based automatic detection scheme to assist the diagnosis of early gastric cancer in endoscopic images. We performed transfer learning using two classes (cancer and normal) of image datasets that have detailed texture information on lesions derived from a small number of annotated images. The accuracy of our trained network was 87.6%, and the sensitivity and specificity were well balanced, which is important for future practical use. We also succeeded in presenting a candidate region of early gastric cancer as a heat map of unknown images. The detection accuracy was 82.8%. This means that our proposed scheme may offer substantial assistance to endoscopists in decision making.

Reliability and accuracy of blue light imaging for staging of intestinal metaplasia in the stomach

Background and aims: An endoscopic grading system (EGGIM) using narrow-band-imaging (NBI) has shown to accurately identify patients with extensive gastric intestinal metaplasia (GIM). However, description with alternative systems such as blue-light-imaging (BLI) is limited. The aim of this study is to determine the reliability and accuracy of BLI-bright regarding diagnosis and staging of GIM.Methods: Reliability of WLE (white-light-endoscopy) and BLI among 6 observers was assessed using a standard classification based on endoscopic images. Afterward, 37 patients were submitted to gastroscopy using FujifilmEG-760Z and endoscopists had to determine EGGIM score using BLI-bright and to perform gastric biopsies for operative-link-of-gastric-intestinal-metaplasia (OLGIM) calculation. BLI-bright accuracy was determined by comparing results with prior EGGIM scores with NBI and current OLGIM.Results: Compared with WLE, the interobserver reliability between observers was substantially better with BLI (Weighted Kappa: 0.8 vs 0.41). There was an 84% agreement between BLI and NBI assessing EGGIM intervals (EGGIM 0-4vs5-10). The area under the ROC curve was 0.90 (95%CI: 0.79-1.0) using the cut-off of EGGIM > 4 to determine advanced GIM, with a sensitivity of 100% (95%CI: 88-100%).Discussion: BLI-bright is reliable for the diagnosis of gastric intestinal metaplasia and agrees significantly with NBI evaluation. Preliminary data suggests high sensitivity for identifying patients with increased risk of gastric cancer.

Comparative study of magnifying narrow-band imaging and conventional white light endoscopy in the diagnosis of Helicobacter pylori status after eradication therapy

Predicting Helicobacter pylori (Hp) status by endoscopic finding would be useful in recent clinical condition that the use of proton-pump inhibitors, anti-platelet, and anti-coagulant have become widespread. We aimed to elucidate the diagnostic accuracy of magnifying narrow-band imaging (M-NBI) endoscopy in distinguishing Hp status in patients with or without history of successful Hp eradication and compare this accuracy to the diagnostic accuracy of conventional white light (WL) endoscopy.Two hundred seven endoscopic examinations before and after Hp eradication were performed in prospective 163 patients. Endoscopic images by using the M-NBI and conventional WL were stored electronically and randomly allocated to 2 readers for evaluation. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and diagnostic accuracy were assessed by reference to Hp status assessed by conventional clinical test.Sensitivity, specificity, PPV, NPV, and accuracy for predicting Hp status for the conventional WL was 72.2%, 75.5%, 72.2%, 75.5%, and 73.9% for the first reader; 86.6%, 57.3%, 64.1%, 82.9%, and 71.0% for the second reader. On the other hand, sensitivity, specificity, PPV, NPV, and accuracy for predicting Hp status for the M-NBI was 96.9%, 93.6%, 93.1%, 97.1%, and 95.2% for the first reader; 92.8%, 93.6%, 92.8%, 93.6%, and 93.2% for the second reader, respectively. The diagnostic accuracy of M-NBI was significantly higher than that of WL (P < .0001 for both readers). Inter-observer agreement of M-NBI (k = 0.83) was also better than that of WL (k = 0.53).M-NBI was capable of distinguishing Hp status before and after eradication therapy.

Analysis of factors related to poor outcome after e-learning training in endoscopic diagnosis of early gastric cancer using magnifying narrow-band imaging

BACKGROUND AND AIMS

An e-learning system teaching endoscopic diagnostic process for early gastric cancer using magnifying endoscopy with narrow-band imaging (M-NBI) was established, and its efficacy in improving the diagnostic performance for early gastric cancer was proven in a multicenter randomized controlled trial. The aim of this study was to clarify the difference in learning effect in each lesion characteristic.

METHODS

Three hundred sixty-five participants diagnosed 40 gastric lesions based on M-NBI findings using the vessel-plus-surface classification system. The diagnosis data collected from each participant were assessed in this study. The accuracy of NBI cancer diagnosis was assessed using area under the receiver operating characteristics curve (AUC/ROC) analysis. AUC/ROCs were separately calculated in each lesion characteristic (shape and size), and the data were compared between tests 1 and 3.

RESULTS

Continuous net reclassification improvement (cNRI) analysis of all lesions revealed significant improvement in reclassification when participants underwent e-learning (cNRI, 1.17; P < .01). The integrated discrimination improvement analysis demonstrated that the e-learning system improved diagnostic ability (.19; P < .01). According to the analysis depending on the lesion's characteristics, high AUC/ROCs were demonstrated in depressed and small lesions (<10 mm; .90 and .93, respectively). The cNRI analysis showed remarkable e-learning improvement in both depressed (cNRI, 1.33; P < .01) and small lesions (cNRI, 1.46; P < .01). However, no significant e-learning improvement was observed in elevated or flat lesions.

CONCLUSIONS

In M-NBI education for endoscopists, a good learning outcome was obtained in depressed and small lesions, but a poor learning outcome was demonstrated in elevated and flat lesions. (Clinical trial registration number: UMIN000008569.).

Influence of Helicobacter pylori Infection on Endoscopic Findings of Gastric Adenocarcinoma of the Fundic Gland Type

PURPOSE

Gastric adenocarcinoma of the fundic gland type (chief cell predominant type) (GA-FG-CCP) was first reported as a rare adenocarcinoma found in the normal fundic mucosa. Recent studies have proposed the possibility that GA-FG-CCPs were also generated in the atrophic mucosa after Helicobacter pylori (HP) eradication therapy. However, little is known on the endoscopic findings of GA-FG-CCP generated in the atrophic mucosa due to its extreme rarity.

MATERIALS AND METHODS

A total of 8 patients who underwent endoscopic submucosal resection and were diagnosed with GA-FG-CCP generated in the HP-uninfected mucosa (4 cases, HP-uninfected group) or HP-eradicated atrophic mucosa (4 cases, HP-eradicated group) were retrospectively analyzed, and their endoscopic findings, including magnifying endoscopy with narrow band imaging (M-NBI), and pathological features were compared.

RESULTS

While GA-FG-CCPs in the 2 groups displayed similar macroscopic appearance, M-NBI demonstrated that characteristic microvessels (tapered microvessels like withered branches) were specifically identified in the HP-eradicated group. Pathological investigation revealed that a decreasing number of fundic glands and thinned foveolar epithelium covering tumor ducts were thought to lower the thickness of the covering layer over tumor ducts in the HP-eradicated group. Moreover, dilation of vessels just under the surface of the lesions contributed to the visualization of microvessels by M-NBI.

CONCLUSIONS

The change in background mucosa due to HP infection influenced the thickness of the covering layer over the tumor ducts and M-NBI finding of GA-FG-CCP.

Gastric foveolar-type adenomas endoscopically showing a raspberry-like appearance in the Helicobacter pylori -uninfected stomach

Background and study aims  Foveolar-type adenoma is described as a very rare tumor that occurs in individuals without Helicobacter pylori (HP) infection and diagnosed as adenocarcinoma in the Japanese Classification of Gastric Carcinoma (JCGC). However, we have frequently encountered patients with foveolar-type adenoma that endoscopically resembles a hyperplastic polyp, suggesting that it has just been overlooked to date. Here, we analyzed clinicopathological characteristics of a special subtype of foveolar-type adenoma showing specific endoscopic findings. Patients and methods  From a total of 212 patients with gastric cancer resected during a 22-month period, we enrolled 14 (6.6 %) diagnosed with foveolar-type adenoma (adenocarcinoma in JCGC). HP infection status was determined by eradication history, HP serum IgG antibody level, urea breath test, and endoscopic and histological findings. All lesions were observed using white-light endoscopy and narrow-band imaging with magnification endoscopy (NBIME). Endoscopically resected lesions were histologically examined. Results  None of 14 patients had a current or past history of HP infection. All lesions were visualized on non-atrophic gastric mucosa as small reddish protrusions with fine granular surface, showing a raspberry-like appearance. NBIME showed papillary or gyrus-like microstructures with irregular capillary. Lesions were histologically diagnosed as foveolar-type adenoma showing MUC5AC-positive gastric mucin phenotype. Ki-67 was overexpressed (median labeling index 69.9 %, range 28.4 - 92.1 %), though all lesions were an intraepithelial tumor without stromal invasion. p53 over-staining was not seen in any. Conclusions  Raspberry-like lesions on non-atrophic gastric mucosa in HP-uninfected individuals should be evaluated for the possibility of a special subtype of foveolar-type adenoma.

Easy Detection and Fast Removal of Gastric Anisakis during Narrow-Band Imaging Endoscopy with L-Menthol Administration

Anisakiasis is a parasitic disease caused by the ingestion of raw or uncooked seafood infected with third-stage larvae of the anisakid nematodes. A 45-year-old Japanese man presented with epigastric pain and itchy skin with rash on his arm, chest, and back after eating vinegar-marinated raw mackerel sushi. He underwent an emergent endoscopic examination using narrow-band imaging (NBI) that revealed two anisakid larvae. NBI showed the larvae more clearly than white light imaging on the cardiac region of the stomach. We sprayed L-menthol on the larvae for stopping their movement and then easily removed them using biopsy forceps. The macroscopic examination and genotype analysis of the specimens revealed the two larvae as belonging to A. simplex sensu stricto. Our case demonstrates the usefulness of endoscopic examination with NBI and of the L-menthol spray in visualizing and immobilizing the larvae for removal.

Risk of multiple early gastric cancers in a patient with precursor lesions and endoscopic surveillance for 7 years: A case report

RATIONALE

Severe mucosal atrophy or intestinal metaplasia is a risk factor for synchronous and metachronous intestinal gastric cancer. Magnifying endoscopy with narrow-band imaging was useful for assessing differentiated early gastric cancer (EGC).

PATIENT CONCERNS

A 62-year-old Chinese female was diagnosed with 5 multiple EGCs or high-grade dysplasia (HGD) with endoscopic surveillance for 7 years.

DIAGNOSES

Synchronous and metachronous multiple EGCs.

INTERVENTIONS

Endoscopic submucosal dissection (ESD) with en bloc resection was performed for all 5 multiple lesions. The ESD specimens were pathologically diagnosed with adenocarcinoma confined to the mucosa or HGD.

OUTCOMES

After endoscopy resection, no residual, recurrent, or synchronous lesions were detected by endoscopic surveillance after ESD.

LESSONS

Long-term, meticulous endoscopic surveillance is needed to monitor risk factors associated with multiple EGCs in patients with severe mucosal atrophy or intestinal metaplasia despite successful Helicobacter pylori eradication.

Utility of a standardized training program for endoscopic diagnosis of early gastrointestinal neoplasia

Background and study aims  Image enhanced endoscopy (IEE) allows endoscopists to improve recognition and characterization of gastrointestinal neoplasia. The Asian Novel Bio-Imaging and Intervention Group (ANBIG) conducted a standardized training program in endoscopic diagnosis and treatment of early gastrointestinal cancers in Asia. We embarked on a study to investigate the effect of this module on endoscopic diagnosis of early gastrointestinal neoplasia. Methods  This prospectively collected database was from workshops conducted on training for endoscopic diagnosis of early gastrointestinal neoplasia. All workshops were conducted in a standardized format, which included a pretest, a learning phase consisting of didactic lectures, case discussion, and live demonstration followed by a post-test to assess knowledge gained. The pretest and post-training tests were standardized questions addressing four domains, including basic knowledge of imaging and diagnosis of esophageal, gastric, and colonic neoplasia. Results  From November 2013 to November 2016, 41 ANBIG workshops were conducted in 13 countries. A total of 1863 delegates and 40 faculty participated in these workshops. Of the delegates, 627 completed both tests. There was a significant improvement after training in all domains of the tests. There was a trend in general lack of knowledge across all domains for delegates from "low" healthcare cost countries before training. All delegates demonstrated significant improvement in knowledge of all domains after the workshop irrespective of whether they were from "high" or "low" healthcare cost per capita countries. Conclusion  A standardized teaching program on IEE improved the diagnostic ability and quality of endoscopists in recognizing early gastrointestinal neoplasia in Asia.

Multicenter prospective study on the histological diagnosis of gastric cancer by narrow band imaging-magnified endoscopy with and without acetic acid

Background and study aims  The usefulness of endoscopy for diagnosing histological type remains unclear. This study aimed to examine the diagnostic accuracy of white light endoscopy (WLE), magnified endoscopy with narrow band imaging (NBI-ME), and NBI-ME with acetic acid enhancement (NBI-AA) for histological type of gastric cancer. Patients and methods  Patients with depressed-type gastric cancers resected by endoscopic submucosal dissection were prospectively enrolled, and 221 cases were analyzed. Histological type was diagnosed by WLE, followed by NBI-ME and NBI-AA. Histological type was classified into differentiated adenocarcinoma and undifferentiated adenocarcinoma. Histological type was diagnosed based on lesion color in WLE, surface patterns (pit, villi, and unclear) and vascular irregularities in NBI-ME, and surface patterns in NBI-AA. Results  Histological types of target areas were differentiated adenocarcinoma and undifferentiated adenocarcinoma in 206 and 15 cases, respectively. Diagnostic accuracy of WLE, NBI-ME, and NBI-AA for the histological type was 96.4 % (213/221), 96.8 % (214/221), and 95.5 % (211/221), respectively. No significant differences were observed among modalities. Positive predictive value based on endoscopic findings in NBI-ME was 98.0 % (149/152) for the villi pattern, 100 % (19/19) for the irregular pit pattern, 100 % (9/9) for the unclear surface pattern with a vascular network, 90.3 % (28/31) for the unclear surface pattern with mild vascular irregularity, and 88.9 % (8/9) for the unclear surface pattern with severe vascular irregularity. Conclusions  NBI-ME and NBI-AA did not show any advantages over WLE for diagnostic accuracy. Villi pattern, irregular pit pattern, and vascular network may be useful for identifying differentiated adenocarcinoma.

An Asian consensus on standards of diagnostic upper endoscopy for neoplasia

BACKGROUND

This is a consensus developed by a group of expert endoscopists aiming to standardise the preparation, process and endoscopic procedural steps for diagnosis of early upper gastrointestinal (GI) cancers.

METHOD

The Delphi method was used to develop consensus statements through identification of clinical questions on diagnostic endoscopy. Three consensus meetings were conducted to consolidate the statements and voting. We conducted a systematic literature search on evidence for each statement. The statements were presented in the second consensus meeting and revised according to comments. The final voting was conducted at the third consensus meeting on the level of evidence and agreement.

RESULTS

Risk stratification should be conducted before endoscopy and high risk endoscopic findings should raise an index of suspicion. The presence of premalignant mucosal changes should be documented and use of sedation is recommended to enhance detection of superficial upper GI neoplasms. The use of antispasmodics and mucolytics enhanced visualisation of the upper GI tract, and systematic endoscopic mapping should be conducted to improve detection. Sufficient examination time and structured training on diagnosis improves detection. Image enhanced endoscopy in addition to white light imaging improves detection of superficial upper GI cancer. Magnifying endoscopy with narrow-band imaging is recommended for characterisation of upper GI superficial neoplasms. Endoscopic characterisation can avoid unnecessary biopsy.

CONCLUSION

This consensus provides guidance for the performance of endoscopic diagnosis and characterisation for early gastric and oesophageal neoplasia based on the evidence. This will enhance the quality of endoscopic diagnosis and improve detection of early upper GI cancers.

Linked color imaging improves the endoscopic visibility of gastric mucosal cancers

Background and study aims  As a newly developed endoscopy technique, linked color imaging (LCI) provides very bright images with enhanced color tones. With the objective of improving the detection rate of gastric mucosal cancers, which are often difficult to detect, we examined the utility of LCI from the viewpoint of visibility. Patients and methods  The current study used 100 consecutive gastric mucosal cancers ≤ 20 mm in diameter. For each lesion, we selected one endoscopic image acquired by white-light imaging (WLI), blue-laser imaging (BLI) -bright, and LCI modes. Four endoscopists interpreted the images; using a previously reported scale, we scored the visibility level on a scale of 1 - 4. Results  The mean (± SD) visibility scores were 2.54 ± 1.10 for WLI, 3.02 ± 1.07 for BLI-bright, and 3.28 ± 0.97 for LCI. The score was significantly higher for BLI-bright compared with WLI ( P  < .001) and again higher for LCI compared with BLI-bright ( P  < .001). For the experts, the scores for BLI-bright and LCI were similar, but both were significantly higher than the score for WLI. For the trainees, there was no significant difference between the WLI and BLI-bright scores, but LCI score was significantly higher than those for WLI and BLI-bright scores. With regard to clinical characteristics, LCI particularly enhanced visibility of normochromic, flat and depressed lesions, which had the lowest visibility scores of all three modalities compared with those of the other lesions. Conclusion  LCI increased visibility and may contribute to early detection of gastric mucosal cancers.