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

Automatic detection of early gastric cancer in endoscopy based on Mask region-based convolutional neural networks (Mask R-CNN)(with video)

The artificial intelligence (AI)-assisted endoscopic detection of early gastric cancer (EGC) has been preliminarily developed. The currently used algorithms still exhibit limitations of large calculation and low-precision expression. The present study aimed to develop an endoscopic automatic detection system in EGC based on a mask region-based convolutional neural network (Mask R-CNN) and to evaluate the performance in controlled trials. For this purpose, a total of 4,471 white light images (WLIs) and 2,662 narrow band images (NBIs) of EGC were obtained for training and testing. In total, 10 of the WLIs (videos) were obtained prospectively to examine the performance of the RCNN system. Furthermore, 400 WLIs were randomly selected for comparison between the Mask R-CNN system and doctors. The evaluation criteria included accuracy, sensitivity, specificity, positive predictive value and negative predictive value. The results revealed that there were no significant differences between the pathological diagnosis with the Mask R-CNN system in the WLI test (χ² = 0.189, P=0.664; accuracy, 90.25%; sensitivity, 91.06%; specificity, 89.01%) and in the NBI test (χ² = 0.063, P=0.802; accuracy, 95.12%; sensitivity, 97.59%). Among 10 WLI real-time videos, the speed of the test videos was up to 35 frames/sec, with an accuracy of 90.27%. In a controlled experiment of 400 WLIs, the sensitivity of the Mask R-CNN system was significantly higher than that of experts (χ² = 7.059, P=0.000; 93.00% VS 80.20%), and the specificity was higher than that of the juniors (χ² = 9.955, P=0.000, 82.67% VS 71.87%), and the overall accuracy rate was higher than that of the seniors (χ² = 7.009, P=0.000, 85.25% VS 78.00%). On the whole, the present study demonstrates that the Mask R-CNN system exhibited an excellent performance status for the detection of EGC, particularly for the real-time analysis of WLIs. It may thus be effectively applied to clinical settings.

Assigning a different endoscopist for each annual follow-up may contribute to improved gastric cancer detection rates

Background and study aims  Esophagogastroduodenoscopy (EGD) is an effective and important diagnostic tool to detect gastric cancer (GC). Although previous studies show that examiner, patient, and instrumental factors influence the detection of GC, we analyzed whether assigning a different examiner to surveillance EGD would improve the detection of GC compared to assigning the same examiner as in the previous endoscopy. Patients and methods  We retrospectively reviewed patients who underwent two or more consecutive surveillance EGDs at a single center between 2017 and 2019. We identified factors associated with GC detection using multivariable regression analysis and propensity-score matching. Results  Among 7794 patients, 99 GC lesions in 93 patients were detected by surveillance EGD (detection rate; 1.2 %), with a mean surveillance interval of 11.2 months. Among the detected 99 lesions, 87 (87.9 %) were curatively treated with endoscopy. There were no differences in the clinicopathologic characteristics of GC detected by the same or different endoscopists. GC detection in the group examined by different endoscopists was more statistically significant than in the group examined by the same endoscopist, even after propensity-score matching (1.6 % and 0.7 %; P  < 0.05). Endoscopic experience and other factors were not statistically significant between the two groups. Conclusions  In surveillance EGD, having a different endoscopist for each exam may improve GC detection rates, regardless of the endoscopist's experience.

Performance of an artificial intelligence-based diagnostic support tool for early gastric cancers: Retrospective study

OBJECTIVES

Endoscopists' abilities to diagnose early gastric cancers (EGCs) vary, especially between specialists and nonspecialists. We developed an artificial intelligence (AI)-based diagnostic support tool "Tango" to differentiate EGCs and compared its performance with that of endoscopists.

METHODS

The diagnostic performances of Tango and endoscopists (34 specialists, 42 nonspecialists) were compared using still images of 150 neoplastic and 165 non-neoplastic lesions. Neoplastic lesions included EGCs and adenomas. The primary outcome was to show the noninferiority of Tango (based on sensitivity) over specialists. The secondary outcomes were the noninferiority of Tango (based on accuracy) over specialists and the superiority of Tango (based on sensitivity and accuracy) over nonspecialists. The lower limit of the 95% confidence interval (CI) of the difference between Tango and the specialists for sensitivity was calculated, with >-10% defined as noninferiority and >0% defined as superiority in the primary outcome. The comparable differences between Tango and the endoscopists for each performance were calculated, with >10% defined as superiority and >0% defined as noninferiority in the secondary outcomes.

RESULTS

Tango achieved superiority over the specialists based on sensitivity (84.7% vs. 65.8%, difference 18.9%, 95% CI 12.3-25.3%) and demonstrated noninferiority based on accuracy (70.8% vs. 67.4%). Tango achieved superiority over the nonspecialists based on sensitivity (84.7% vs. 51.0%) and accuracy (70.8% vs. 58.4%).

CONCLUSIONS

The AI-based diagnostic support tool for EGCs demonstrated a robust performance and may be useful to reduce misdiagnosis.

Demarcation Line Determination for Diagnosis of Gastric Cancer Disease Range Using Unsupervised Machine Learning in Magnifying Narrow-Band Imaging

Background and Aims: It is important to determine an accurate demarcation line (DL) between the cancerous lesions and background mucosa in magnifying narrow-band imaging (M-NBI)-based diagnosis. However, it is difficult for novice endoscopists. We aimed to automatically determine the accurate DL using a machine learning method. Methods: We used an unsupervised machine learning approach to determine the DLs. Our method consists of the following four steps: (1) an M-NBI image is segmented into superpixels using simple linear iterative clustering; (2) the image features are extracted for each superpixel; (3) the superpixels are grouped into several clusters using the k-means method; and (4) the boundaries of the clusters are extracted as DL candidates. The 23 M-NBI images of 11 cases were used for performance evaluation. The evaluation investigated the similarity of the DLs identified by endoscopists and our method, and the Euclidean distance between the two DLs was calculated. For the single case of 11 cases, the histopathological examination was also conducted to evaluate the proposed system. Results: The average Euclidean distances for the 11 cases were 10.65, 11.97, 7.82, 8.46, 8.59, 9.72, 12.20, 9.06, 22.86, 8.45, and 25.36. The results indicated that the proposed method could identify similar DLs to those identified by experienced doctors. Additionally, it was confirmed that the proposed system could generate pathologically valid DLs by increasing the number of clusters. Conclusions: Our proposed system can support the training of inexperienced doctors as well as enrich the knowledge of experienced doctors in endoscopy.

Gastric Cancer: Emerging Trends in Prevention, Diagnosis, and Treatment

Gastric adenocarcinoma (GC) is the fourth leading cause of global cancer mortality, and the leading infection-associated cancer. Helicobacter pylori is the dominant risk factor for GC and classified as an IARC class I carcinogen. Surveillance of gastric premalignant conditions is now indicated in high-risk patients. Upper endoscopy is the gold standard for GC diagnosis, and image-enhanced endoscopy increases the detection of gastric premalignant conditions and early gastric cancer (EGC). Clinical staging is crucial for treatment approach, defining early gastric cancer, operable locoregional disease, and advanced GC. Endoscopic submucosal dissection is the treatment of choice for most EGC. Targeted therapies are rapidly evolving, based on biomarkers including MSI/dMMR, HER2, and PD-L1. These advancements in surveillance, diagnostic and therapeutic strategies are expected to improve GC survival rates in the near term.

Clinical outcome of the delineation-without-negative-biopsy strategy in magnifying image-enhanced endoscopy for identifying the extent of differentiated-type early gastric cancer

BACKGROUND

The histologic evaluation of biopsy samples collected from the surrounding mucosa has conventionally been used to determine the horizontal extent of early gastric cancer. Recently, optical delineation using magnifying image-enhanced endoscopy (IEE) has been considered an alternative method to histologic evaluation. This study aimed to assess the clinical outcome and efficacy of this method in identifying cancer margins.

METHODS

Overall, 921 patients with 1018 differentiated-type early gastric tumors who underwent endoscopic submucosal dissection (ESD) were examined. Before ESD, the lesions were classified based on whether they have clear or unclear margins on magnifying IEE. When the lesions had clear margins, the marking dots were placed outside the margins without a negative biopsy. Successful delineation was defined as lesions with clear margins and accurate delineation based on histopathological examination. The primary outcome was the accuracy of optical delineation without a negative biopsy compared with histopathological diagnosis. Moreover, the clinicopathological factors associated with an unsuccessful delineation were assessed.

RESULTS

Of 1018 lesions, 820 had a clear margin and 198 an unclear margin. Of 820 lesions with a clear margin, 817 and 3 had an accurate and inaccurate delineation, respectively, according to the histological examination. Accordingly, the accuracy rate of optical delineation was 99.6% (817/820). The significant independent factors associated with an unsuccessful delineation were absence of Helicobacter pylori infection after eradication, tumor size > 20 mm, and moderate differentiation.

CONCLUSIONS

Optical delineation may be an alternative method to histological evaluation in lesions with a clear margin on magnifying IEE.

Which factors make Barrett's esophagus lesions difficult to diagnose?

Background and study aims  Although the Japan Esophageal Society's magnifying endoscopic classification for Barrett's epithelium (JES-BE) offers high diagnostic accuracy, some cases are challenging to diagnose as dysplastic or non-dysplastic in daily clinical practice. Therefore, we investigated the diagnostic accuracy of this classification and the clinicopathological features of Barrett's esophagus cases that are difficult to diagnose correctly. Patients and methods  Five endoscopists with experience with fewer than 10 cases of magnifying observation for superficial Barrett's esophageal carcinoma reviewed 132 images of Barrett's mucosa or carcinoma (75 dysplastic and 57 non-dysplastic cases) obtained using high-definition magnification endoscopy with narrow-band imaging (ME-NBI). They diagnosed each image as dysplastic or non-dysplastic according to the JES-BE classification, and the diagnostic accuracy was calculated. To identify risk factors for misdiagnosed images, images with a correct rate of less than 40 % were defined as difficult-to-diagnose, and those with 60 % or more were defined as easy-to-diagnose. Logistic regression analysis was performed to identify risk factors for difficult-to-diagnose images. Results  The sensitivity, specificity and overall accuracy were 67 %, 80 % and 73 %, respectively. Of the 132 ME-NBI images, 34 (26 %) were difficult-to-diagnose and 99 (74 %) were easy-to-diagnose. Logistic regression analysis showed low-grade dysplasia (LGD) and high-power magnification images were each significant risk factors for difficult-to-diagnose images (OR: 6.80, P  = 0.0017 and OR: 3.31, P  = 0.0125, respectively). Conclusions  This image assessment study suggested feasibility of the JES-BE classification for diagnosis of Barrett's esophagus by non-expert endoscopists and risk factors for difficult diagnosis as high-power magnification and LGD histology. For non-experts, high-power magnification images are better evaluated in combination with low-power magnification images.

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

Background and aim

Magnifying image-enhanced endoscopy was demonstrated to have higher diagnostic accuracy than white-light endoscopy. However, differentiating early gastric cancers (EGCs) from benign lesions is difficult for beginners. We aimed to determine whether the computer-aided model for the diagnosis of gastric lesions can be applied to videos rather than still images.

Methods

A total of 719 magnifying optical enhancement images of EGCs, 1,490 optical enhancement images of the benign gastric lesions, and 1,514 images of background mucosa were retrospectively collected to train and develop a computer-aided diagnostic model. Subsequently, 101 video segments and 671 independent images were used for validation, and error frames were labeled to retrain the model. Finally, a total of 117 unaltered full-length videos were utilized to test the model and compared with those diagnostic results made by independent endoscopists.

Results

Except for atrophy combined with intestinal metaplasia (IM) and low-grade neoplasia, the diagnostic accuracy was 0.90 (85/94). The sensitivity, specificity, PLR, NLR, and overall accuracy of the model to distinguish EGC from non-cancerous lesions were 0.91 (48/53), 0.78 (50/64), 4.14, 0.12, and 0.84 (98/117), respectively. No significant difference was observed in the overall diagnostic accuracy between the computer-aided model and experts. A good level of kappa values was found between the model and experts, which meant that the kappa value was 0.63.

Conclusions

The performance of the computer-aided model for the diagnosis of EGC is comparable to that of experts. Magnifying the optical enhancement model alone may not be able to deal with all lesions in the stomach, especially when near the focus on severe atrophy with IM. These results warrant further validation in prospective studies with more patients. A ClinicalTrials.gov registration was obtained (identifier number: NCT04563416).

Clinical Trial Registration

ClinicalTrials.gov, identifier NCT04563416.

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

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

Real-time automated diagnosis of colorectal cancer invasion depth using a deep learning model with multimodal data (with video)

BACKGROUND AND AIMS

The optical diagnosis of colorectal cancer (CRC) invasion depth with white light (WL) and image-enhanced endoscopy (IEE) remains challenging. We aimed to construct and validate a 2-modal deep learning-based system, incorporated with both WL and IEE images (named Endo-CRC) in estimating the invasion depth of CRC.

METHODS

Samples were retrospectively obtained from 3 hospitals in China. We combined WL and IEE images into image pairs. Altogether, 337,278 image pairs from 268 noninvasive and superficial CRC and 181,934 image pairs from 82 deep CRC were used for training. A total of 296,644 and 4528 image pairs were used for internal and external tests and for comparison with endoscopists. Thirty-five videos were used for evaluating the real-time performance of the Endo-CRC system. Two deep learning models, solely using either WL (model W) or IEE images (model I), were constructed to compare with Endo-CRC.

RESULTS

The accuracies of Endo-CRC in internal image tests with and without advanced CRC were 91.61% and 93.78%, respectively, and 88.65% in the external test, which did not include advanced CRC. In an endoscopist-machine competition, Endo-CRC achieved an expert comparable accuracy of 88.11% and the highest sensitivity compared with all endoscopists. In a video test, Endo-CRC achieved an accuracy of 100.00%. Compared with model W and model I, Endo-CRC had a higher accuracy (per image pair: 91.61% vs 88.27% compared with model I and 91.61% vs 81.32% compared with model W).

CONCLUSIONS

The Endo-CRC system has great potential for assisting in CRC invasion depth diagnosis and may be well applied in clinical practice.

Magnifying endoscopy with narrow-band imaging is useful in differentiating gastric cancer from matched adenoma in white light imaging

This study assessed the effect of magnifying endoscopy with narrow-band imaging (M-NBI) on the endoscopic differential diagnosis between intramucosal gastric carcinomas and adenomas with matched characteristics. Associations between magnified endoscopic findings and pathological high-grade cellular and architectural atypia were also investigated. In total, the records of 50 adenomas and 50 intramucosal well-differentiated adenocarcinomas matched by tumor size (≥ 20 mm or < 20 mm), shape (depression or non-depression), and color (red or non-red) were extracted. Fourteen endoscopists diagnosed adenoma or cancer in the 100 cases with conventional white light imaging (C-WLI), then did the same with C-WLI + M-NBI.The cancer diagnostic sensitivity, specificity, and accuracy were assessed. The sensitivity of C-WLI + M-NBI for cancer diagnosis was 79.9% compared to 71.6% with C-WLI (p < 0.001). There were no significant differences in specificity (40.1% vs. 36.3%, p = 0.296) and accuracy (55.9% vs. 58.1%, p = 0.163). High-grade cytological or architectural atypia was diagnosed more often with irregular microvascular pattern (IMVP) or microsurface pattern (IMSP), respectively, than the low-grade forms. In conclusion, IMVP and IMSP correlate with high-grade cytological and architectural atypia. M-NBI is useful in differentiating intramucosal carcinoma from adenoma and can reduce underdiagnosis of cancer.

The Feasibility of Applying Artificial Intelligence to Gastrointestinal Endoscopy to Improve the Detection Rate of Early Gastric Cancer Screening

Convolutional neural networks in the field of artificial intelligence show great potential in image recognition. It assisted endoscopy to improve the detection rate of early gastric cancer. The 5-year survival rate for advanced gastric cancer is less than 30%, while the 5-year survival rate for early gastric cancer is more than 90%. Therefore, earlier screening for gastric cancer can lead to a better prognosis. However, the detection rate of early gastric cancer in China has been extremely low due to many factors, such as the presence of gastric cancer without obvious symptoms, difficulty identifying lesions by the naked eye, and a lack of experience among endoscopists. The introduction of artificial intelligence can help mitigate these shortcomings and greatly improve the accuracy of screening. According to relevant reports, the sensitivity and accuracy of artificial intelligence trained on deep cirrocumulus neural networks are better than those of endoscopists, and evaluations also take less time, which can greatly reduce the burden on endoscopists. In addition, artificial intelligence can also perform real-time detection and feedback on the inspection process of the endoscopist to standardize the operation of the endoscopist. AI has also shown great potential in training novice endoscopists. With the maturity of AI technology, AI has the ability to improve the detection rate of early gastric cancer in China and reduce the death rate of gastric cancer related diseases in China.

Convolutional neural network-based system for endocytoscopic diagnosis of early gastric cancer

Background

Endocytoscopy (ECS) aids early gastric cancer (EGC) diagnosis by visualization of cells. However, it is difficult for non-experts to accurately diagnose EGC using ECS. In this study, we developed and evaluated a convolutional neural network (CNN)-based system for ECS-aided EGC diagnosis.

Methods

We constructed a CNN based on a residual neural network with a training dataset comprising 906 images from 61 EGC cases and 717 images from 65 noncancerous gastric mucosa (NGM) cases. To evaluate diagnostic ability, we used an independent test dataset comprising 313 images from 39 EGC cases and 235 images from 33 NGM cases. The test dataset was further evaluated by three endoscopists, and their findings were compared with CNN-based results.

Results

The trained CNN required 7.0 s to analyze the test dataset. The area under the curve of the total ECS images was 0.93. The CNN produced 18 false positives from 7 NGM lesions and 74 false negatives from 28 EGC lesions. In the per-image analysis, the accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were 83.2%, 76.4%, 92.3%, 93.0%, and 74.6%, respectively, with the CNN and 76.8%, 73.4%, 81.3%, 83.9%, and 69.6%, respectively, for the endoscopist-derived values. The CNN-based findings had significantly higher specificity than the findings determined by all endoscopists. In the per-lesion analysis, the accuracy, sensitivity, specificity, PPV, and NPV of the CNN-based findings were 86.1%, 82.1%, 90.9%, 91.4%, and 81.1%, respectively, and those of the results calculated by the endoscopists were 82.4%, 79.5%, 85.9%, 86.9%, and 78.0%, respectively.

Conclusions

Compared with three endoscopists, our CNN for ECS demonstrated higher specificity for EGC diagnosis. Using the CNN in ECS-based EGC diagnosis may improve the diagnostic performance of endoscopists.

Photodynamic Diagnosis for the Identification of Intestinal-Type Gastric Cancers and High-Grade Adenomas

Gastric cancer is the second most common cancer in Japan. The incidence of gastric cancer remains high owing to the increase in the elderly population. Endoscopy outperforms radiography in identifying early gastric cancer (EGC). Furthermore, image-enhanced endoscopy (IEE) has been developed and implemented worldwide in clinical practice. Magnifying IEE images can help to visualize the microvascular pattern and microstructure architecture, which is used for the characterization of EGC. However, accurate endoscopic diagnosis requires the experience and skill of endoscopists, making an objective and simple diagnostic method desirable. In this retrospective study, we investigated the diagnostic yield of 5-aminolevulinic acid (5-ALA)-mediated photodynamic diagnosis (PDD) for identifying gastric cancers and high-grade adenomas. In total, 52 lesions from 43 patients were ultimately included in the study. We detected 5-ALA-mediated protoporphyrin IX fluorescence in 45 of the 52 lesions that were initially intended for PDD, resulting in a detection rate of 86.5%, whereas each signet ring cell carcinoma was negative using 5-ALA PDD. In eight of the patients with multiple lesions, 17 lesions were identified using 5-ALA PDD. Again, we took biopsies from six areas that we suspected as new lesions. While 4 lesions were gastric neoplasms resected by endoscopic submucosal dissection, two other lesions were normal. Preoperative 5-ALA-PDD could provide additional diagnostic yields to detect such multiple lesions simultaneously. No severe adverse events were observed. Prospective multicenter studies are warranted to confirm the usefulness of 5-ALA PDD for EGC identification.

Usefulness of artificial intelligence in early gastric cancer

Gastric cancer (GC) is a major cancer worldwide, with high mortality and morbidity. Endoscopy, important for the early detection of GC, requires trained skills, high-quality technologies, surveillance and screening programs. Early diagnosis allows a better prognosis, through surgical or curative endoscopic therapy. Magnified endoscopy with virtual chromoendoscopy remarkably improve the detection of early gastric cancer (EGC) when endoscopy is performed by expert endoscopists. Artificial intelligence (AI) has also been introduced to GC diagnostics to increase diagnostic efficiency. AI improves the early detection of gastric lesions because it supports the non-expert and experienced endoscopist in defining the margins of the tumor and the depth of infiltration. AI increases the detection rate of EGC, reduces the rate of missing tumors, and characterizes EGCs, allowing clinicians to make the best therapeutic decision, that is, one that ensures curability. AI has had a remarkable evolution in medicine in recent years, moving from the research phase to clinical practice. In addition, the diagnosis of GC has markedly progressed. We predict that AI will allow great evolution in the diagnosis and treatment of EGC by overcoming the variability in performance that is currently a limitation of chromoendoscopy.

Diagnosis of histological type of early gastric cancer by magnifying narrow‐band imaging: A multicenter prospective study

Objectives

Distinguishing undifferentiated‐type from differentiated‐type early gastric cancers (EGC) is crucial for determining the indication of endoscopic resection. We aimed to investigate the diagnostic performance of white‐light endoscopy (WLE) and magnifying narrow‐band imaging (M‐NBI) for the histological type of EGC.

Methods

In this multicenter prospective study, patients with histologically proven cT1 EGC, macroscopically depressed or flat type, size ≥5 mm, and without erosion/ulcer, were recruited. The diagnostic criterion of WLE for undifferentiated‐type EGC was pale color. The M‐NBI algorithm was created based on microsurface and microvascular patterns, and lesions with absent microsurface pattern and opened‐loop microvascular patterns were diagnosed as undifferentiated‐type. The center of the lesion was defined as the evaluation point and was initially evaluated by WLE, then by M‐NBI, and a biopsy specimen was taken as a reference standard. The primary and key secondary endpoints were overall diagnostic accuracy and specificity, respectively.

Results

In total, 167 lesions (122 differentiated‐type and 45 undifferentiated‐type EGCs) in 167 patients were analyzed. The overall accuracy, sensitivity, specificity, and positive likelihood ratio of WLE for undifferentiated‐type cancer were 80%, 69%, 84%, and 4.4, respectively, and those of M‐NBI were 82%, 53%, 93%, and 7.2, respectively. There was no significant difference in overall accuracy (p = 0.755), but specificity was significantly higher in M‐NBI (p = 0.041).

Conclusions

The use of M‐NBI did not improve the accuracy of WLE for the diagnosis of depressed/flat undifferentiated‐type EGCs but improved the specificity. It may reduce surgical overtreatment by preventing misdiagnosis of differentiated‐type EGC as undifferentiated‐type.

Proposal of minimum elements for screening and diagnosis of gastric cancer by an international Delphi consensus

The World Endoscopy Organization Stomach and Duodenal Diseases Committee extracted minimum elements for screening and diagnosis of gastric cancer (GC) in aim to support countries that do not have national guidelines on screening and diagnosis of GC. Current national or international guidelines were collected worldwide and recommendations were classified according to the quality of evidence and were finalized through a modified Delphi method. The minimum elements consist of seven categories: [1] Extraction of high-risk patients of GC before esophagogastroduodenoscopy (EGD), [2] Patients who need surveillance of GC, [3] Method to ensure quality of EGD for detection of GC, [4] Individual GC risk assessment by EGD, [5] Extraction of high-risk patients of GC after EGD [6] Qualitative or differential diagnosis of GC by EGD, and [7] Endoscopic assessment to choose the therapeutic strategy for GC. These minimum elements will be a guide to promote the elimination of GC among countries with a high incidence of GC who lack national guidelines or screening programs.

Multi-label recognition of cancer-related lesions with clinical priors on white-light endoscopy

Deep learning-based computer-aided diagnosis techniques have demonstrated encouraging performance in endoscopic lesion identification and detection, and have reduced the rate of missed and false detections of disease during endoscopy. However, the interpretability of the model-based results has not been adequately addressed by existing methods. This phenomenon is directly manifested by a significant bias in the representation of feature localization. Good recognition models experience severe feature localization errors, particularly for lesions with subtle morphological features, and such unsatisfactory performance hinders the clinical deployment of models. To effectively alleviate this problem, we proposed a solution to optimize the localization bias in feature representations of cancer-related recognition models that is difficult to accurately label and identify in clinical practice. Optimization was performed in the training phase of the model through the proposed data augmentation method and auxiliary loss function based on clinical priors. The data augmentation method, called partial jigsaw, can "break" the spatial structure of lesion-independent image blocks and enrich the data feature space to decouple the interference of background features on the space and focus on fine-grained lesion features. The annotation-based auxiliary loss function used class activation maps for sample distribution correction and led the model to present localization representation converging on the gold standard annotation of visualization maps. The results show that with the improvement of our method, the precision of model recognition reached an average of 92.79%, an F1-score of 92.61%, and accuracy of 95.56% based on a dataset constructed from 23 hospitals. In addition, we quantified the evaluation representation of visualization feature maps. The improved model yielded significant offset correction results for visualized feature maps compared with the baseline model. The average visualization-weighted positive coverage improved from 51.85% to 83.76%. The proposed approach did not change the deployment capability and inference speed of the original model and can be incorporated into any state-of-the-art neural network. It also shows the potential to provide more accurate localization inference results and assist in clinical examinations during endoscopies.

Real-time use of artificial intelligence for diagnosing early gastric cancer by magnifying image-enhanced endoscopy: a multicenter diagnostic study (with videos)

BACKGROUND AND AIMS

Endoscopy is a pivotal method for detecting early gastric cancer (EGC). However, skill among endoscopists varies greatly. Here, we proposed a deep learning-based system named ENDOANGEL-ME to diagnose EGC in magnifying image-enhanced endoscopy (M-IEE).

METHODS

M-IEE images were retrospectively obtained from 6 hospitals in China, including 4667 images for training and validation, 1324 images for internal tests, and 4702 images for external tests. One hundred eighty-seven stored videos from 2 hospitals were used to evaluate the performance of ENDOANGEL-ME and endoscopists and to assess the effect of ENDOANGEL-ME on improving the performance of endoscopists. Prospective consecutive patients undergoing M-IEE were enrolled from August 17, 2020 to August 2, 2021 in Renmin Hospital of Wuhan University to assess the applicability of ENDOANGEL-ME in clinical practice.

RESULTS

A total of 3099 patients undergoing M-IEE were enrolled in this study. The diagnostic accuracy of ENDOANGEL-ME for diagnosing EGC was 88.44% and 90.49% in internal and external images, respectively. In 93 internal videos, ENDOANGEL-ME achieved an accuracy of 90.32% for diagnosing EGC, significantly superior to that of senior endoscopists (70.16% ± 8.78%). In 94 external videos, with the assistance of ENDOANGEL-ME, endoscopists showed improved accuracy and sensitivity (85.64% vs 80.32% and 82.03% vs 67.19%, respectively). In 194 prospective consecutive patients with 251 lesions, ENDOANGEL-ME achieved a sensitivity of 92.59% (25/27) and an accuracy of 83.67% (210/251) in real clinical practice.

CONCLUSIONS

This multicenter diagnostic study showed that ENDOANGEL-ME can be well applied in the clinical setting. (Clinical trial registration number: ChiCTR2000035116.).

Additive effect of evaluating microsurface and microvascular patterns using magnifying endoscopy with narrow-band imaging in gastric cancer: a post-hoc analysis of a single-center observational study

Background

No studies have compared the performance of microvascular and microsurface patterns alone with their combination in patients undergoing magnifying endoscopy with narrow-band imaging for diagnosing gastric cancer. This study aimed to clarify the differences in diagnostic performance among these methods.

Methods

Thirty-three participating endoscopists who had received specialized training in magnifying endoscopy evaluated the microvascular and microsurface patterns of images of 106 cancerous and 106 non-cancerous lesions. If classified as “irregular,” the lesion was diagnosed as gastric cancer. To evaluate diagnostic performance, we compared the diagnostic accuracy, sensitivity, and specificity of these methods.

Results

Performance-related items did not differ significantly between microvascular and microsurface patterns. However, the diagnostic accuracy and sensitivity were significantly higher when using a combination of these methods than when using microvascular (82.1% [76.4–86.7] vs. 76.4% [70.3–81.6] and 69.8% [60.5–77.8] vs. 63.2% [53.7–71.8]; P < 0.001 and P = 0.008, respectively) or microsurface (82.1% [76.4–86.7] vs. 73.6% [67.3–79.1] and 69.8% [60.5–77.8] vs. 52.8% [43.4‒62.1]; both, P < 0.001) patterns alone. The additive effect on diagnostic accuracy and sensitivity was 5.7‒8.6% and 6.6‒17.0%, respectively.

Conclusions

We demonstrate the superiority of the combination of microvascular and microsurface patterns over microvascular or microsurface patterns alone for diagnosing gastric cancer. Our data support the use of the former method in clinical practice. Although a major limitation of this study was its retrospective, single-center design, our findings may help to improve the diagnosis of gastric cancer.

Differential diagnosis of superficial duodenal epithelial tumor and non-neoplastic lesion in duodenum by magnified endoscopic examination with image-enhanced endoscopy

BACKGROUND

Differential diagnosis of superficial duodenal epithelial tumors (SDETs) and non-neoplastic lesions (NNLs) in duodenum by endoscopy is difficult. Here, we attempted to distinguish them by magnified endoscopic examination with image-enhanced endoscopy (IEE-ME).

METHODS

Various IEE-ME findings of 95 SDETs who underwent endoscopic resection and 58 NNLs who underwent biopsy were retrospectively reviewed.

RESULTS

When we compared the IEE-ME findings of SDETs and NNLs, the presence of demarcation line (DL) (97.9% vs. 79.3%, P = 0.0002), white opaque substance (WOS) (84.2% vs. 1.7%, P < 0.0001) and light blue crest (LBC) (93.7% vs. 32.8%, P < 0.0001) and the absence of enlarged marginal epithelium (EME) (98.9% vs. 62.1%, P < 0.0001) were significantly more frequent in SDETs than NNLs. When divided into each superficial structure, it was the most effective to evaluate the combination of WOS and LBC as SDET with open-loop structure (OLS), and the combination of DL and EME as SDET with closed-loop structure (CLS). However, LBC was excluded because of low inter- and intra-observer agreements. Finally, the sensitivity, specificity and accuracy for the diagnosis of SDETs were 88.4%, 98.3% and 92.2%, respectively, and we developed an algorithm for the differential diagnosis of duodenal lesions.

CONCLUSION

We could distinguish SDET from NNL, diagnosed SDET as presence of WOS indicated OLS of superficial structure, and presence of DL and absence of EME indicated CLS of superficial structure.

Diagnosis of Esophageal Lesions by Multi-Classification and Segmentation Using an Improved Multi-Task Deep Learning Model

It is challenging for endoscopists to accurately detect esophageal lesions during gastrointestinal endoscopic screening due to visual similarities among different lesions in terms of shape, size, and texture among patients. Additionally, endoscopists are busy fighting esophageal lesions every day, hence the need to develop a computer-aided diagnostic tool to classify and segment the lesions at endoscopic images to reduce their burden. Therefore, we propose a multi-task classification and segmentation (MTCS) model, including the Esophageal Lesions Classification Network (ELCNet) and Esophageal Lesions Segmentation Network (ELSNet). The ELCNet was used to classify types of esophageal lesions, and the ELSNet was used to identify lesion regions. We created a dataset by collecting 805 esophageal images from 255 patients and 198 images from 64 patients to train and evaluate the MTCS model. Compared with other methods, the proposed not only achieved a high accuracy (93.43%) in classification but achieved a dice similarity coefficient (77.84%) in segmentation. In conclusion, the MTCS model can boost the performance of endoscopists in the detection of esophageal lesions as it can accurately multi-classify and segment the lesions and is a potential assistant for endoscopists to reduce the risk of oversight.

Evaluation of Microvascular Patterns Alone Using Endocytoscopy with Narrow-Band Imaging for Diagnosing Gastric Cancer

INTRODUCTION

Although endocytoscopy (EC) with narrow-band imaging (NBI) is effective in diagnosing gastric cancer, no diagnostic system has been validated. We explored a specific diagnostic system for gastric cancer using EC with NBI.

METHODS

Equal numbers of images from cancerous and noncancerous areas (114 images each) were assessed by endoscopists with (development group: 33) and without (validation group: 28) specific training in magnifying endoscopy with NBI. Microvascular and microsurface patterns (MS) in each image were evaluated. Lesions were diagnosed as cancerous when patterns were deemed "irregular." The accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of a diagnosis according to patterns on EC with NBI (microvascular pattern [MV] alone, MS alone, and both) were evaluated and compared between groups to determine the diagnostic performance.

RESULTS

In the development and validation groups, diagnoses based on the MV alone had significantly higher accuracy (91.7% vs. 76.3%, p < 0.0001 and 92.5% vs. 67.5%, p < 0.0001, respectively) and sensitivity (88.6% vs. 68.3%, p < 0.0001 and 89.5% vs. 38.6%, p < 0.0001, respectively) than those based on the MS alone. In both groups, there were no significant differences in diagnostic accuracy between using the MV alone and both patterns.

DISCUSSION/CONCLUSION

Evaluation of the MV alone is a simple and accurate diagnostic method for gastric cancer. This system could find widespread applications in clinical practice.

Deep learning system compared with expert endoscopists in predicting early gastric cancer and its invasion depth and differentiation status (with videos)

BACKGROUND AND AIMS

We aimed to develop and validate a deep learning-based system that covers various aspects of early gastric cancer (EGC) diagnosis, including detecting gastric neoplasm, identifying EGC, and predicting EGC invasion depth and differentiation status. Herein, we provide a state-of-the-art comparison of the system with endoscopists using real-time videos in a nationwide human-machine competition.

METHODS

This multicenter, prospective, real-time, competitive comparative, diagnostic study enrolled consecutive patients who received magnifying narrow-band imaging endoscopy at the Peking University Cancer Hospital from June 9, 2020 to November 17, 2020. The offline competition was conducted in Wuhan, China, and the endoscopists and the system simultaneously read patients' videos and made diagnoses. The primary outcomes were sensitivity in detecting neoplasms and diagnosing EGCs.

RESULTS

One hundred videos, including 37 EGCs and 63 noncancerous lesions, were enrolled; 46 endoscopists from 44 hospitals in 19 provinces in China participated in the competition. The sensitivity rates of the system for detecting neoplasms and diagnosing EGCs were 87.81% and 100%, respectively, significantly higher than those of endoscopists (83.51% [95% confidence interval [CI], 81.23-85.79] and 87.13% [95% CI, 83.75-90.51], respectively). Accuracy rates of the system for predicting EGC invasion depth and differentiation status were 78.57% and 71.43%, respectively, slightly higher than those of endoscopists (63.75% [95% CI, 61.12-66.39] and 64.41% [95% CI, 60.65-68.16], respectively).

CONCLUSIONS

The system outperformed endoscopists in identifying EGCs and was comparable with endoscopists in predicting EGC invasion depth and differentiation status in videos. This deep learning-based system could be a powerful tool to assist endoscopists in EGC diagnosis in clinical practice.

Multi-Task Model for Esophageal Lesion Analysis Using Endoscopic Images: Classification with Image Retrieval and Segmentation with Attention

The automatic analysis of endoscopic images to assist endoscopists in accurately identifying the types and locations of esophageal lesions remains a challenge. In this paper, we propose a novel multi-task deep learning model for automatic diagnosis, which does not simply replace the role of endoscopists in decision making, because endoscopists are expected to correct the false results predicted by the diagnosis system if more supporting information is provided. In order to help endoscopists improve the diagnosis accuracy in identifying the types of lesions, an image retrieval module is added in the classification task to provide an additional confidence level of the predicted types of esophageal lesions. In addition, a mutual attention module is added in the segmentation task to improve its performance in determining the locations of esophageal lesions. The proposed model is evaluated and compared with other deep learning models using a dataset of 1003 endoscopic images, including 290 esophageal cancer, 473 esophagitis, and 240 normal. The experimental results show the promising performance of our model with a high accuracy of 96.76% for the classification and a Dice coefficient of 82.47% for the segmentation. Consequently, the proposed multi-task deep learning model can be an effective tool to help endoscopists in judging esophageal lesions.

Endoscopic Classifications of Early Gastric Cancer: A Literature Review

Endoscopic technologies have been continuously advancing throughout the years to facilitate improvement in the detection and diagnosis of gastric lesions. With the development of different endoscopic diagnostic modalities for EGC, several classifications have been advocated for the evaluation of gastric lesions, aiming for an early detection and diagnosis. Sufficient knowledge on the appearance of EGC on white light endoscopy is fundamental for early detection and management. On the other hand, those superficial EGC with subtle morphological changes that are challenging to be detected with white light endoscopy may now be clearly defined by means of image-enhanced endoscopy (IEE). By combining magnifying endoscopy and IEE, irregularities in the surface structures can be evaluated and highlighted, leading to improvements in EGC diagnostic accuracy. The main scope of this review article is to offer a closer look at the different classifications of EGC based on several endoscopic diagnostic modalities, as well as to introduce readers to newer and novel classifications, specifically developed for the stomach, for the assessment and diagnosis of gastric lesions.

"Pink Pattern" Visualized in Magnifying Endoscopy With Narrow-Band Imaging Is a Novel Feature of Early Differentiated Gastric Cancer: A Bridge Between Endoscopic Images and Histopathological Changes

Background: A pink color change occasionally found by us under magnifying endoscopy with narrow-band imaging (ME-NBI) may be a special feature of early gastric cancer (EGC), and was designated the “pink pattern”. The purposes of this study were to determine the relationship between the pink pattern and the cytopathological changes in gastric cancer cells and whether the pink pattern is useful for the diagnosis of EGC.

Methods: The color features of ME-NBI images and pathological images of cancerous gastric mucosal surfaces were extracted and quantified. The cosine similarity was calculated to evaluate the correlation between the pink pattern and the nucleus-to-cytoplasm ratio of cancerous epithelial cells. Two diagnostic tests were performed by 12 endoscopists using stored ME-NBI images of 185 gastric lesions to investigate the diagnostic efficacy of the pink pattern for EGC. The diagnostic values, such as the area under the curve (AUC), the accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), of test 1 and test 2 were compared.

Results: The cosine similarity between the color values of ME-NBI images and pathological images of 20 lesions was at least 0.744. The median AUC, accuracy, sensitivity, specificity, PPV, and NPV of test 2 were significantly better than those of test 1 for all endoscopists and for the junior and experienced groups.

Conclusions: The pink pattern observed in ME-NBI images correlated strongly with the change in the nucleus-to-cytoplasm ratio of gastric epithelial cells, and could be considered a useful marker for the diagnosis of differentiated EGC.

Feasibility of standardized procedures of white light gastroscopy for clinical practice: A multicenter study in China

OBJECTIVE

We aimed to establish a standardized procedure for white light gastroscopy (WLG) to screen gastric lesions including early gastric cancer (EGC) in China and to verify its efficacy and feasibility in clinical practice.

METHODS

A standardized WLG procedure for outpatients at nine tertiary hospitals in Beijing was established. Clinical information of the participants and details of the endoscopic procedures were recorded.

RESULTS

A total of 1051 participants were enrolled in a baseline conventional endoscopic survey between March 2014 and December 2015, while 2156 patients were enrolled in the standardized WLG operation from January 2016 to June 2017. The procedure time of the standardized procedure was significantly longer than that of the baseline conventional procedure (P = 0.003). More images were obtained during the standardized procedure compared with the baseline conventional procedure (P < 0.001). The overall detection rate of gastric lesions in the standardized procedure group was significantly higher than that in the baseline procedure group (52.5% vs 38.4%, P < 0.01). The satisfaction scores of both participants and endoscopists in the standardized procedure group were significantly higher than in the baseline procedure group.

CONCLUSIONS

Compared with the conventional procedure, standardized WLG procedure significantly improves the detection rate of gastric lesions as well as the satisfaction score of participants and endoscopists despite its longer procedure time. It is effective and feasible in clinical practice in China for the use of currently available endoscopic equipment.

Complete one-to-one correspondence between magnifying endoscopic and histopathologic images: the KOTO method II

Recent advances in magnifying endoscopy with narrow-band imaging/blue laser imaging have aided in the diagnosis of gastrointestinal lesions. However, it requires knowledge of the relationship between magnifying endoscopic and histopathological images. We propose a novel method which makes possible a complete correspondence between magnifying endoscopic and histopathological images at the single glandular duct level. The KOTO method II enables three-dimensional visualization of the correlation between the endoscopic surface pattern of the mucosa and histopathological images. This method may be helpful in the development of diagnosis using magnifying endoscopy.

Current Status of Photodynamic Diagnosis for Gastric Tumors

Although the recent development and widespread use of image-enhanced endoscopy and magnifying endoscopy have improved endoscopic diagnosis of gastric cancer, it is somewhat complicated, requires a higher level of expertise, and is still subjective. Photodynamic endoscopic diagnosis (PDED) is based on the fluorescence of photosensitizers that accumulate in tumors, which enables objective evaluation independent of the endoscopist’s experience, and is useful for tumor detection. The objective of this work was to perform a narrative review of PDED for gastric tumors and to introduce our approach to PDED in gastric tumors in our hospital. In our review there have been case reports of PDED for gastric cancer, but its usefulness has not been established because no prospective studies evaluating its usefulness have been performed. In our previous study, 85.7% (42/49) of gastric tumors exhibited fluorescence in PDED. PDED may be useful in the diagnosis of early gastric cancer. Our previous studies were pilot studies in cancer patients; therefore, future prospective studies are required to verify the usefulness of PDED.

Artificial Intelligence in Gastric Cancer: Identifying Gastric Cancer Using Endoscopic Images with Convolutional Neural Network

Simple Summary

Gastric cancer (GC) is one of the most newly diagnosed cancers and the fifth leading cause of death globally. Previous studies reported that the detection rate of gastric cancer (EGC) at an earlier stage is low, and the overall false-negative rate with esophagogastroduodenoscopy (EGD) is up to 25.8%, which often leads to inappropriate treatment. Accurate diagnosis of EGC can reduce unnecessary interventions and benefits treatment planning. Convolutional neural network (CNN) models have recently shown promising performance in analyzing medical images, including endoscopy. This study shows that an automated tool based on the CNN model could improve EGC diagnosis and treatment decision.

Abstract

Gastric cancer (GC) is one of the most newly diagnosed cancers and the fifth leading cause of death globally. Identification of early gastric cancer (EGC) can ensure quick treatment and reduce significant mortality. Therefore, we aimed to conduct a systematic review with a meta-analysis of current literature to evaluate the performance of the CNN model in detecting EGC. We conducted a systematic search in the online databases (e.g., PubMed, Embase, and Web of Science) for all relevant original studies on the subject of CNN in EGC published between 1 January 2010, and 26 March 2021. The Quality Assessment of Diagnostic Accuracy Studies-2 was used to assess the risk of bias. Pooled sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio were calculated. Moreover, a summary receiver operating characteristic curve (SROC) was plotted. Of the 171 studies retrieved, 15 studies met inclusion criteria. The application of the CNN model in the diagnosis of EGC achieved a SROC of 0.95, with corresponding sensitivity of 0.89 (0.88–0.89), and specificity of 0.89 (0.89–0.90). Pooled sensitivity and specificity for experts endoscopists were 0.77 (0.76–0.78), and 0.92 (0.91–0.93), respectively. However, the overall SROC for the CNN model and expert endoscopists was 0.95 and 0.90. The findings of this comprehensive study show that CNN model exhibited comparable performance to endoscopists in the diagnosis of EGC using digital endoscopy images. Given its scalability, the CNN model could enhance the performance of endoscopists to correctly stratify EGC patients and reduce work load.

Unified magnifying endoscopic classification for esophageal, gastric and colonic lesions: a feasibility pilot study

Background and study aims  Image-enhanced magnifying endoscopy allows optimization of the detection and diagnosis of lesions found in the gastrointestinal tract. Current organ-specific classifications are well-accepted by specialized endoscopists but may pose confusion for general gastroenterologists. To address this, our group proposed the Unified Magnifying Endoscopic Classification (UMEC) which can be applied either in esophagus, stomach, or colon. The aim of this study was to evaluate the diagnostic performance and clinical applicability of UMEC. Patients and methods  A single-center, feasibility pilot study was conducted. Two endoscopists with experience in magnifying narrow band imaging (NBI), blinded to white-light and non-magnifying NBI findings as well as histopathological diagnosis, independently reviewed and diagnosed all images based on UMEC. In brief, UMEC is divided into three categories: non-neoplasia, intramucosal neoplasia, and deep submucosal invasive cancer. The diagnostic performance of UMEC was assessed while using the gold standard histopathology as a reference. Results  A total of 303 gastrointestinal lesions (88 esophageal squamous lesions, 90 gastric lesions, 125 colonic lesions) were assessed. The overall accuracy for both endoscopists in the diagnosis of esophageal squamous cell cancer, gastric cancer, and colorectal cancer were 84.7 %, 89.5 %, and 83.2 %, respectively. The interobserver agreement for each organ, Kappa statistics of 0.51, 0.73, and 0.63, was good. Conclusions  UMEC appears to be a simple and practically acceptable classification, particularly to general gastroenterologists, due to its good diagnostic accuracy, and deserves further evaluation in future studies.

Endoscopic diagnosis and treatment of gastric dysplasia and early cancer: Current evidence and what the future may hold

Gastric cancer accounts for a significant proportion of worldwide cancer-related morbidity and mortality. The well documented precancerous cascade provides an opportunity for clinicians to detect and treat gastric cancers at an endoscopically curable stage. In high prevalence regions such as Japan and Korea, this has led to the implementation of population screening programs. However, guidelines remain ambiguous in lower prevalence regions. In recent years, there have been many advances in the endoscopic diagnosis and treatment of early gastric cancer and precancerous lesions. More advanced endoscopic imaging has led to improved detection and characterization of gastric lesions as well as superior accuracy for delineation of margins prior to resection. In addition, promising early data on artificial intelligence in gastroscopy suggests a future role for this technology in maximizing the yield of advanced endoscopic imaging. Data on endoscopic resection (ER) are particularly robust in Japan and Korea, with high rates of curative ER and markedly reduced procedural morbidity. However, there is a shortage of data in other regions to support the applicability of protocols from these high prevalence countries. Future advances in endoscopic therapeutics will likely lead to further expansion of the current indications for ER, as both technology and proceduralist expertise continue to grow.

Application of Current Image-Enhanced Endoscopy in Gastric Diseases

Image-enhanced endoscopy (IEE) plays an integral role in endoscopic diagnosis and treatment. IEE enables an early and accurate detection of cancer and characterization of lesions prior to therapeutic decisions. Ideal IEE can serve as an optical or digital chromoscopic endoscopy, as well as an optical biopsy that predicts exact histopathology. Several IEE modalities have recently been developed and are used in the clinical field. The stomach is a challenging organ for imaging because of its complex secretion function and status of Helicobacter pylori infection. Therefore, understanding the current IEE modalities for their clinical applicability in an evidence-based approach is warranted. Along with technology refinements, the new paradigm will be available for the diagnosis of gastric cancer or other conditions in the stomach in the near future.

Dynamic diagnosis of early gastric cancer with microvascular blood flow rate using magnifying endoscopy (with video): A pilot study

BACKGROUND AND AIM

Magnifying endoscopy (ME) diagnostic algorithm for early gastric cancer (EGC) relies on qualitative features such as microvascular (MV) architecture and microsurface structure; however, it is a "static" diagnostic algorithm that uses still images. ME can visualize red blood cell flow within subepithelial microvessels in real time. Here, we evaluated the utility of using the MV blood flow rate in combination with ME for the diagnosis of EGC as a retrospective study.

METHODS

Patients with differentiated-type EGC (n = 10) or patchy redness (n = 10) underwent ME with blue laser imaging. The mean MV blood flow rates of EGC, patchy redness, and background mucosa were calculated by the mean movement distance of one tagging red blood cell using split images of ME with blue laser imaging videos. We compared the mean MV blood flow rate between EGC, patchy redness, and background mucosa and also calculated the MV blood flow imaging ratio (inside lesion/background mucosa) between EGC and patchy redness.

RESULTS

Mean MV blood flow rate was significantly lower in EGC (1481 μm/s; range 1057-1762) than in patchy redness (3859 μm/s; 2435-5899) or background mucosa (4140.6 μm/s; 2820-6247) (P < 0.01). The MV blood flow imaging ratio was significantly lower in EGC (0.39; 0.27-0.62) than in patchy redness (0.90; 0.78-1.1) (P < 0.01).

CONCLUSIONS

Dynamic diagnosis with MV blood flow rate using ME may be useful for the differential diagnosis of EGC and patchy redness. Endoscopic assessment of dynamic processes within the gastric mucosa may facilitate the diagnosis of EGC.

Endoscopic Advances for Gastric Neoplasia Detection

This article explores advances in endoscopic neoplasia detection with supporting clinical evidence and future aims. The ability to detect early gastric neoplastic lesions amenable to curative endoscopic submucosal dissection provides the opportunity to decrease gastric cancer mortality rates. Newer imaging techniques offer enhanced views of mucosal and microvascular structures and show promise in differentiating benign from malignant lesions and improving targeted biopsies. Conventional chromoendoscopy is well studied and validated. Narrow band imaging demonstrates superiority over magnified white light. Autofluorescence imaging, i-scan, flexible spectral imaging color enhancement, and bright image enhanced endoscopy show promise but insufficient evidence to change current clinical practice.

Fundamentals, Diagnostic Capabilities and Perspective of Narrow Band Imaging for Early Gastric Cancer

The development of image-enhanced endoscopy has dramatically improved the qualitative and quantitative diagnosis of gastrointestinal tumors. In particular, narrow band imaging (NBI) has been widely accepted by endoscopists around the world in their daily practice. In 2009, Yao et al. proposed vessel plus surface (VS) classification, a diagnostic algorithm for early gastric cancer using magnifying endoscopy with NBI (ME-NBI), and in 2016, Muto et al. proposed a magnifying endoscopy simple diagnostic algorithm for early gastric cancer (MESDA-G) based on VS classification. In addition, the usefulness of ME-NBI in the differential diagnosis of gastric cancer from gastritis, diagnosis of lesion extent, inference of histopathological type, and diagnosis of depth has also been investigated. In this paper, we narrative review the basic principles, current status, and future prospects of NBI.

Comparison of the diagnostic efficacy of blue laser imaging with narrow band imaging for gastric cancer and precancerous lesions: a meta-analysis

BACKGROUND AND AIMS

novel endoscopic techniques including narrowband imaging (NBI) and blue laser imaging (BLI) have led to the improved detection of early stage gastric cancer and precancerous lesions. However, these techniques are not generally thought to be equivalent at present and BLI is generally considered as superior to NBI. Therefore, this comprehensive meta-analysis aimed to definitively compare the diagnostic efficacy of NBI and BLI for the diagnosis of gastric cancer and precancerous lesions.

METHODS

relevant articles were identified via searches of the PubMed, Web of Science, Embase and Cochrane Library databases from their inception until October 2019. In total, 28 relevant studies were identified and incorporated into the meta-analysis. RevMan5.3 was used to assess the relative diagnostic efficacy of these two imaging modalities in these studies. The threshold was assessed using Meta-DiSc 1.4 and STATA 14.0 for bivariate regression modeling of pooled studies.

RESULTS

the pooled sensitivity of BLI for gastric cancer was 0.89 (0.80, 0.95) and the specificity was 0.92 (0.76, 0.98). The pooled sensitivity of NBI for gastric cancer was 0.83 (0.75, 0.89) and the specificity was 0.95 (0.91, 0.97). The pooled sensitivity of BLI for precancerous lesions was 0.81 (0.71, 0.87) and the specificity was 0.90 (0.80, 0.96). The pooled sensitivity of NBI for precancerous lesions was 0.80 (0.75, 0.85) and the specificity was 0.88 (0.77, 0.94).

CONCLUSIONS

this study showed that both BLI and NBI have a very high diagnostic efficacy for the detection of gastric cancer and precancerous lesions, the sensitivity and specificity of these two approaches were similar.

Systematic Review on Optical Diagnosis of Early Gastrointestinal Neoplasia

BACKGROUND

Meticulous endoscopic characterization of gastrointestinal neoplasias (GN) is crucial to the clinical outcome. Hereby the indication and type of resection (endoscopically, en-bloc or piece-meal, or surgical resection) are determined. By means of established image-enhanced (IEE) and magnification endoscopy (ME) GN can be characterized in terms of malignancy and invasion depth. In this context, the statistical evidence and accuracy of these diagnostic procedures should be elucidated. Here, we present a systematic review of the literature.

RESULTS

21 Studies could be found which met the inclusion criteria. In clinical prospective trials and meta-analyses, the diagnostic accuracy of >90% for characterization of malignant neoplasms could be documented, if ME with IEE was used in squamous cell esophageal cancer, stomach, or colonic GN.

CONCLUSIONS

Currently, by means of optical diagnosis, today's gastrointestinal endoscopy is capable of determining the histological subtype, exact lateral spread, and depth of invasion of a lesion. The prerequisites for this are an exact knowledge of the anatomical structures, the endoscopic classifications based on them, and a systematic learning process, which can be supported by training courses. More prospective clinical studies are required, especially in the field of Barrett's esophagus and duodenal neoplasia.

[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.

Benefits of linked color imaging for recognition of early differentiated-type gastric cancer: in comparison with indigo carmine contrast method and blue laser imaging

BACKGROUND AND AIM

Linked color imaging (LCI) is a novel endoscopy system, which enhances slight differences in mucosal color. However, whether LCI is more useful than other kinds of image-enhanced endoscopy (IEE) in recognizing early gastric cancer remains unclear. This study aimed to evaluate LCI efficacy compared with the indigo carmine contrast method (IC), and blue laser imaging-bright (BLI-brt) in early differentiated-type gastric cancer recognition.

METHODS

We retrospectively analyzed early differentiated-type gastric cancer, which were examined by all four imaging techniques (white light imaging, IC, LCI, BLI-brt) at Asahi University Hospital from June 2014 to November 2018. Both subjective evaluation (using ranking score: RS) and objective evaluation (using color difference score: CDS) were adopted to quantify early differentiated-type gastric cancer recognition.

RESULTS

During this period, 87 lesions were enrolled in this study. Both RS and CDS of LCI were significantly higher (p < 0.01) than those of IC and BLI-brt. Both RS and CDS of BLI-brt had no significant difference compared with those of IC. Subgroup analysis revealed that LCI was especially useful in post-Helicobacter pylori eradication patients and flat or depressed lesions compared with IC and BLI-brt.

CONCLUSIONS

LCI appears to be more beneficial for the recognition of early differentiated-type gastric cancer in endoscopic screenings than IC and BLI-brt from the middle to distant view.

Identifying early gastric cancer under magnifying narrow-band images with deep learning: a multicenter study

BACKGROUND AND AIMS

Narrow-band imaging with magnifying endoscopy (ME-NBI) has shown advantages in the diagnosis of early gastric cancer (EGC). However, proficiency in diagnostic algorithms requires substantial expertise and experience. In this study, we aimed to develop a computer-aided diagnostic model for EGM (EGCM) to analyze and assist in the diagnosis of EGC under ME-NBI.

METHODS

A total of 1777 ME-NBI images from 295 cases were collected from 3 centers. These cases were randomly divided into a training cohort (n = 170), an internal test cohort (ITC, n = 73), and an external test cohort (ETC, n = 52). EGCM based on VGG-19 architecture (Visual Geometry Group [VGG], Oxford University, Oxford, UK) with a single fully connected 2-classification layer was developed through fine-tuning and validated on all cohorts. Furthermore, we compared the model with 8 endoscopists with varying experience. Primary comparison measures included accuracy, area under the receiver operating characteristic curve (AUC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).

RESULTS

EGCM acquired AUCs of .808 in the ITC and .813 in the ETC. Moreover, EGCM achieved similar predictive performance as the senior endoscopists (accuracy: .770 vs .755, P = .355; sensitivity: .792 vs .767, P = .183; specificity: .745 vs .742, P = .931) but better than the junior endoscopists (accuracy: .770 vs .728, P < .05). After referring to the results of EGCM, the average diagnostic ability of the endoscopists was significantly improved in terms of accuracy, sensitivity, PPV, and NPV (P < .05).

CONCLUSIONS

EGCM exhibited comparable performance with senior endoscopists in the diagnosis of EGC and showed the potential value in aiding and improving the diagnosis of EGC by endoscopists.

Image-Enhanced Endoscopy and Its Corresponding Histopathology in the Stomach

In recent years, the technological innovation and progress of endoscopic equipment have been remarkable, and various endoscopic observation techniques have been developed. Among them, representative techniques are magnified observation and narrow-band imaging. Magnifying endoscopy with narrow-band imaging (M-NBI) can visualize superficial microanatomies in the stomach. The normal morphology of the microanatomy visualized using M-NBI differs according to the part of the stomach. The vessel plus surface (VS) classification system has been developed as a diagnostic criterion for early gastric cancer using M-NBI, and its usefulness has been proven. Based on the VS classification system, a magnifying endoscopy simple diagnostic algorithm for early gastric cancer (MESDA-G), a simplified algorithm used for early gastric cancer diagnosis, was created. We aimed to describe the anatomic structure of the stomach that can be viewed using M-NBI and outline the principles and clinical application of the VS classification system and MESDA-G. (Gut Liver 2021;15:-337)

Risk Factors and Prediction Model for Non-curative Resection of Early Gastric Cancer With Endoscopic Resection and the Evaluation

Background and Study Aim: EGC, also known as Early Gastric Cancer is known to lack the lymph node metastasis and confined along the mucosa, which is treated through an endoscopic resection procedure that includes ESD (Endoscopic Submucosal dissection) and EMR (Endoscopic Mucosal Resection). However, some cases underwent residual disease, recurrence, or additional gastrectomy because of non-curative resection. The following research aims to delineate the threat factors causing the non-curative resection as well as develop a predictive model. Patient and Methods: Effort was taken to collect all the records about the health history of pathologically diagnosed EGC who experienced endoscopic treatment in the Department of Endoscopy, the Capital Medical University, and Beijing Friendship Hospital from January 2012 to January 2020. Patients were grouped into two categories primarily; a curative resection group and finally a non-curative resection group based on the outcomes of the postoperative pathological and immunohistochemical examination results. The statistical methods used included single factor analysis, a multivariate logistic regression analysis and a chi-square test. A nomogram for the prediction of non-curative resection was constructed, which included information on age, gender, resection method, postoperative pathology, tumor size, ulcer, treatment, and infiltration depth. Receiver operating characteristic (ROC) curve analysis and calibration were performed to present the predictive accuracy of the nomogram. Results: Of 443 patients with 478 lesions who had undergone ESD or EMR for EGCs, 127 were identified as being treated non-curative resection. Older patients (>60 years), a large tumor size (>30 mm), submucosal lesion, piecemeal resection, EMR for treatment and undifferentiated tumor histology were associated with non-curative resection group. Our risk nomogram showed good discriminated performance in internal validation (bootstrap-corrected area under the receiver-operating characteristic curve, 0.881; P < 0.001). Conclusions: A validated prediction model was developed to identify people who were subject to undergoing a non-curative resection for ESD. The predictive model that we formulated is essential in providing reliable information to guide the decision-making process on the treatment for EGC before undertaking an endoscopic resection.

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.

Diagnostic performance in gastric cancer is higher using endocytoscopy with narrow-band imaging than using magnifying endoscopy with narrow-band imaging

BACKGROUND

For diagnosing gastric cancer, differences in the diagnostic performance between endocytoscopy with narrow-band imaging and magnifying endoscopy with narrow-band imaging have not been reported. We aimed to clarify these differences by analyzing diagnoses made by endoscopists in Japan.

METHODS

This single-center retrospective cohort study used 106 cancerous and 106 non-cancerous images obtained via both modalities (total, 424 images) for diagnosis. Sixty-one endoscopists with varying experience levels from 45 institutions were included. Diagnostic accuracy, sensitivity, specificity, and positive and negative predictive values were evaluated to determine the diagnostic performance of each modality and compared using the Mann-Whitney U test.

RESULTS

Among all endoscopists, diagnostic accuracy, sensitivity, positive predictive value, and negative predictive value were higher with endocytoscopy with narrow-band imaging than with magnifying endoscopy with narrow-band imaging (percentage [95% confidence interval]: 78.8% [76.4-83.0%] versus 72.2% [69.3-73.6%], p < 0.0001; 82.1% [78.3-85.9%] versus 64.2% [60.4-69.8%], p < 0.0001; 88.7% [82.6-90.7%] versus 78.5% [75.4-85.1%], p = 0.0023; 79.0% [75.3-80.5%] versus 68.5% [66.4-71.6%], p < 0.0001, respectively). In the magnifying endoscopy with narrow-band imaging-trained group, these values were also higher with endocytoscopy with narrow-band imaging than with magnifying endoscopy with narrow-band imaging (p < 0.0001, p = 0.0001, p = 0.0143, and p < 0.0001, respectively). Diagnostic accuracy, sensitivity, and negative predictive value were higher with endocytoscopy with narrow-band imaging than with magnifying endoscopy with narrow-band imaging in the magnifying endoscopy with narrow-band imaging-untrained group (p = 0.0041, p = 0.0049, and p = 0.0098, respectively).

CONCLUSIONS

Diagnostic performance was higher using endocytoscopy with narrow-band imaging than using magnifying endoscopy with narrow-band imaging. Our results may help change the technique used to diagnose gastric cancer.

Quality in diagnostic upper gastrointestinal endoscopy for the detection and surveillance of gastric cancer precursor lesions: Position paper of AEG, SEED and SEAP

This position paper, sponsored by the Asociación Española de Gastroenterología [Spanish Association of Gastroenterology], the Sociedad Española de Endoscopia Digestiva [Spanish Gastrointestinal Endoscopy Society] and the Sociedad Española de Anatomía Patológica [Spanish Anatomical Pathology Society], aims to establish recommendations for performing an high quality upper gastrointestinal endoscopy for the screening of gastric cancer precursor lesions (GCPL) in low-incidence populations, such as the Spanish population. To establish the quality of the evidence and the levels of recommendation, we used the methodology based on the GRADE system (Grading of Recommendations Assessment, Development and Evaluation). We obtained a consensus among experts using a Delphi method. The document evaluates different measures to improve the quality of upper gastrointestinal endoscopy in this setting and makes recommendations on how to evaluate and treat the identified lesions. We recommend that upper gastrointestinal endoscopy for surveillance of GCPL should be performed by endoscopists with adequate training, administering oral premedication and use of sedation. To improve the identification of GCPL, we recommend the use of high definition endoscopes and conventional or digital chromoendoscopy and, for biopsies, NBI should be used to target the most suspicious areas of intestinal metaplasia. Regarding the evaluation of visible lesions, the risk of submucosal invasion should be evaluated with magnifying endoscopes and endoscopic ultrasound should be reserved for those with suspected deep invasion. In lesions amenable to endoscopic resection, submucosal endoscopic dissection is considered the technique of choice.