Single Shot Multibox Detector Automatic Polyp Detection Network Based on Gastrointestinal Endoscopic Images.

PURPOSE: In order to resolve the situation of high missed diagnosis rate and high misdiagnosis rate of the pathological analysis of the gastrointestinal endoscopic images by experts, we propose an automatic polyp detection algorithm based on Single Shot Multibox Detector (SSD). METHOD: In the paper, SSD is based on VGG-16, the fully connected layer is changed to a convolutional layer, and four convolutional layers with successively decreasing scales are added as a new network structure. In order to verify the practicability, it is not only compared with manual polyp detection but also with Mask R-CNN. RESULTS: Multiple experimental results show that the mean Average Precision (mAP) of the SSD network is 95.74%, which is 12.4% higher than the manual detection and 5.7% higher than the Mask R-CNN. When detecting a single frame of image, the detection speed of SSD is 8.41 times that of manual detection. CONCLUSION: Based on the traditional pattern recognition algorithm and the target detection algorithm using deep learning, we select a variety of algorithms to identify and classify polyps to achieve efficient detection results. Our research demonstrates that deep learning has a lot of room for development in the field of gastrointestinal image recognition.

Classification of stomach infections: A paradigm of convolutional neural network along with classical features fusion and selection.

Automated detection and classification of gastric infections (i.e., ulcer, polyp, esophagitis, and bleeding) through wireless capsule endoscopy (WCE) is still a key challenge. Doctors can identify these endoscopic diseases by using the computer-aided diagnostic (CAD) systems. In this article, a new fully automated system is proposed for the recognition of gastric infections through multi-type features extraction, fusion, and robust features selection. Five key steps are performed-database creation, handcrafted and convolutional neural network (CNN) deep features extraction, a fusion of extracted features, selection of best features using a genetic algorithm (GA), and recognition. In the features extraction step, discrete cosine transform, discrete wavelet transform strong color feature, and VGG16-based CNN features are extracted. Later, these features are fused by simple array concatenation and GA is performed through which best features are selected based on K-Nearest Neighbor fitness function. In the last, best selected features are provided to Ensemble classifier for recognition of gastric diseases. A database is prepared using four datasets-Kvasir, CVC-ClinicDB, Private, and ETIS-LaribPolypDB with four types of gastric infections such as ulcer, polyp, esophagitis, and bleeding. Using this database, proposed technique performs better as compared to existing methods and achieves an accuracy of 96.5%.

New endoscopic classification of cascade stomach, a risk factor for reflux esophagitis.

BACKGROUND: We recently demonstrated that cascade stomach detected by barium studies was correlated with upper gastrointestinal symptoms. We developed a new endoscopic classification of cascade stomach and examined its relationship with reflux esophagitis. METHODS: Study 1: the classification (grades 0-3) was based on detecting a ridge that runs from the cardia toward the anterior wall crossing the greater curvature. Inter-observer variation was evaluated by kappa statistics when ten experienced endoscopists used this classification three times each. Study 2: in 710 consecutive subjects (500 men and 210 women) undergoing endoscopic screening, the grade of cascade stomach and incidence of reflux esophagitis were compared. RESULTS: In study 1, the kappa values at the third assessment were 0.85, 0.58, 0.50, and 0.78 for each grade, respectively, while overall agreement was 0.68. In study 2, the incidence of reflux esophagitis in men was 20 % in grade 0, 17 % in grade 1, 25 % in grade 2, and 30 % in grade 3, showing significant differences. Among women, the incidence of reflux esophagitis in each grade was 9, 3, 6, and 35 %, respectively, also showing significant differences. Multivariate analysis showed that independent risk factors for reflux esophagitis were cascade stomach (odds ratio = 2.20), body mass index, and hiatus hernia in men, as well as cascade stomach (odds ratio = 9.01) and smoking tobacco in women. CONCLUSIONS: This endoscopic classification of cascade stomach showed acceptable inter-observer variation. Cascade stomach is a risk factor for reflux esophagitis.

Classification for esophagogastric junction (EGJ) complex based on physiology.

We propose a new classification for esophagogastric junction (EGJ) incorporating both physiologic and morphologic characteristics. Additionally, we contrast it with the Chicago v 3.0 EGJ classification. With Institutional Review Board (IRB) approval, prospectively maintained database was queried to identify patients who underwent high-resolution manometry (HRM) and pH-study between October 2011 and October 2015. Patients with prior foregut intervention, pH study on acid suppression, esophageal dysmotility, or lower esophageal sphincter-crural diaphragm separation of >5 cm were excluded. We classified patients into three groups-Type-A: Complete overlap of lower esophageal sphincter-crural diaphragm (single high-pressure zone); Type-B: Double high-pressure zone with pressure inversion point (PIP) at or above lower esophageal sphincter; Type-C: Double high-pressure zone with PIP below lower esophageal sphincter. A total of 214 included patients were divided into Type-A (n = 101), Type-B (n = 32), and Type-C (n = 81). Abdominal lower esophageal sphincter length (AL), lower esophageal sphincter pressure (LESP), and lower esophageal sphincter pressure integral (LESPI) were significantly lower in Type-C than both Type-A and Type-B [AL(cm): 0.2 vs. 2(P < 0.001) vs. 1.6(P <0.001); LESP(mmHg): 20.1 vs. 32.1(P < 0.001) vs. 29.2(P < 0.001); LESPI(mmHg.cm.s): 187 vs. 412(P < 0.001) vs. 343(P < 0.05)] while overall lower esophageal sphincter length(OL) and Integrated Relaxation Pressure (IRP) were significantly lower in Type-C than Type-A [OL(cm): 2.9 vs. 3.6(P < 0.001); IRP(mmHg): 8.2 vs. 9.6(P < 0.05)]. Type-C patients had significantly higher positive pH score (>14.7) than Type-A and Type-B [72% vs. 47% (P < 0.05) vs. 41% (P < 0.001)]. In Type-C morphology, there is both anatomical and physiological deterioration, weakest lower esophageal sphincter function (abdominal length, lower esophageal sphincter pressure, and lower esophageal sphincter pressure integral) and is most likely to be associated with pathological reflux. This proposed classification incorporates both physiological and morphological derangements in a graded fashion.

New Classification of Gastric Pit Patterns and Vessel Architecture Using Probe-based Confocal Laser Endomicroscopy.

GOALS: To propose a new probe-based confocal laser endomicroscopy (pCLE) classification of gastric pit patterns and vessel architecture, and to assess the accuracy and interobserver agreement. BACKGROUND: pCLE is a newly developed endoscopic device that allows the application of laser microscopy with any conventional endoscope and mosaic imaging. STUDY: A total of 291 pCLE videos from 32 patients were recruited in phase I to establish the new pCLE image classification in the stomach. Eligible patients were then prospectively investigated by pCLE using the newly established classification system. All patients were examined first with high-definition endoscopy followed by pCLE at 7 standardized locations and endoscopic-suspected lesions. Targeted biopsies were performed with precise matching of pCLE recordings. RESULTS: The sensitivity and specificity of type 2b pit pattern for predicting atrophic gastritis were 88.51% and 99.19%, respectively. The sensitivity and specificity of type 2c pit pattern for predicting intestinal metaplasia were 92.34% and 99.34%, respectively. The overall sensitivity and specificity of type 3 pit pattern or vessel architecture for predicting neoplasia were 89.89% and 99.44%, respectively. The interobserver agreement was "substantial" (kappa=0.70) for the differentiation of neoplasia versus non-neoplasia. CONCLUSIONS: The new pCLE classification system in the stomach correlates well with specific pathologic conditions and is reproducible by multiple investigators. Multicenter researches are warranted to further validate its value in clinical practice.

Gastric foregut cystic developmental malformation: case series and literature review.

Foregut cystic developmental malformation (FCDM) is a very rare lesion of the alimentary tract, especially in the stomach. We discuss the concepts of gastric duplication cyst, bronchogenic cysts, and FCDM. Nomenclature has been inconsistent and confusing, but, by some definitions, gastric duplication cysts involve gastric mucosa and submucosal glands, bronchogenic cysts involve respiratory mucosa with underlying cartilage and glands, and FCDM lacks gastric mucosa or underlying glands or cartilage but has pseudostratified ciliated columnar epithelium (PCCE). We searched our departmental case files from the past 15 years and identified 12 cases of FCDM in the alimentary tract. We summarize the features of these 12 cases including a report in detail on a 52-year-old man with a submucosal cyst lined with simple PCCE and irregular and stratified circular muscle layers that merged with gastric smooth muscle bundles near the lesser curvature of the gastric cardia. A literature review of cases with this histology yielded 25 cases. We propose the term gastric-FCDM for such cases. Our own series of 12 cases confirms that preoperative recognition of the entity is infrequent and problematic. The rarity of this developmental disorder, as well as a lack of understanding of its embryologic origins, may contribute to missing the diagnosis. Not appreciating the diagnosis preoperatively can lead to an inappropriate surgical approach. In contrast, presurgical recognition of the entity will contribute to a good outcome and reduced risk of complications.

Correlation between magnifying narrow band imaging and histopathology in gastric protruding/or polypoid lesions: a pilot feasibility trial.

BACKGROUND: Several study showed usefulness of microscopic capillaries, seen by magnifying narrow band imaging (NBI) endoscopy for predicting histopathology among superficial depressed or flat elevated gastric neoplasia (GN). Here we assessed the diagnostic efficacy of magnifying NBI for predicting histopathology among gastric protruding/or polypoid lesions. METHODS: Using endoscopic pictures of magnifying NBI from 95 protruding/or polypoid lesions (19 fundic gland polyps: FGP, 47 hyperplastic polyps: HP, and 29 GN), fine mucosal patterns were classified into four categories: small round, prolonged, villous or ridge, and unclear patterns, and micro vascular patterns were classified into five categories: honey comb, dense vascular, fine net work, core vascular, and unclear patterns. RESULTS: Most suggestive micro vascular patterns for predicting FGP, and HP were honeycomb (sensitivity 94.7%, specificity 97.4%), and dense vascular patterns (sensitivity 93.6%, specificity 91.6%), respectively. Fine net work, core vascular, and unclear patterns presented higher specificity (97%, 100%, and 100%) for predicting GN, and diagnostic efficacy of combined of those patterns was favorable (sensitivity 86.2%, specificity 97.0%). CONCLUSION: Micro vascular patterns by using magnifying NBI provides meaningful information for predicting the histopathology of gastric protruding/or polypoid lesions.

[Portal gastropathy: diagnosis, classification].

Portal hypertensive gastropathy (PHG) is a complex of secondary macroscopic and microscopic changes in the mucous layer of the stomach, resulting from portal hypertension of any origin. The overall prevalence of PHG ranges from 9.1 to 80%. PHG is a potential cause of an acute or chronic gastric bleeding. The presence of PHG is associated with prognosis deterioration that dictates the need for improved diagnosis and treatment strategy. The article summarizes literature on the pathological changes, diagnosis and classification of portal hypertensive gastropathy.

Miami classification for probe-based confocal laser endomicroscopy.

An essential element for any new advanced imaging technology is standardization of indications, terminology, categorization of images, and research priorities. In this review, we propose a state-of-the-art classification system for normal and pathological states in gastrointestinal disease using probe-based confocal laser endomicroscopy (pCLE). The Miami classification system is based on a consensus of pCLE users reached during a meeting held in Miami, Florida, in February 2009.

Clinical validation and risk factors for delayed gastric emptying based on the International Study Group of Pancreatic Surgery (ISGPS) Classification.

BACKGROUND: Delayed gastric emptying (DGE) is one of the most common complications after pancreatoduodenectomy (PD). Because an objective, universally accepted definition of DGE does not yet exist, it is impossible to compare complication rates and outcomes of new operative approaches, operative techniques, and clinical trials. The International Study Group of Pancreatic Surgery (ISGPS) has proposed a universal classification for DGE based on clinical outcomes, but this classification has not been tested rigorously and applied to clinical data. Therefore, the aim of this study was to analyze our experience and to identify predictive factors for DGE by applying the ISGPS classification at a high-volume hospital. METHODS: From October 2002 to December 2007, 129 consecutive patients underwent PD at the Department of Surgery, Yonsei University Medical Center. The severity of DGE was determined according to the ISGPS classification, and risk factors were evaluated retrospectively. RESULTS: The overall incidence of DGE was 33.3%, with 16 (12.4%) patients having grade A, 14 (10.9%) grade B, and 13 (10.1%) grade C. Clinical outcomes worsened progressively as clinical relevant DGE increased. In multivariate analysis, clinically relevant pancreatic fistula (grade B/C) and patients with benign pathology were identified as independent factors for DGE. CONCLUSION: Pancreatic leakage is a serious complication after PD and is also associated with DGE. The ISGPS classification is a clear and useful tool to assess clinical outcomes.

Development of a rating system for digestive system impairments: Korean Academy of Medical Sciences Guideline.

A systematic and effective welfare system for people with digestive system impairments is required. In Korea, an objective and scientific rating guideline does not exist to judge the digestive system impairments. Whether the impairments exist or not and the degree of it need to be examined. Thus, with these considerations we need a scientific rating guideline for digestive system impairments to fit our cultural and social background. In 2007, a research team, for the development of rating impairment guidelines, was organized under the supervision of Korean Academy of Medical Sciences. The rating guidelines for digestive system impairments was classified into upper and lower gastrointestinal tracts impairments and liver impairment. We developed objective rating guidelines for the upper gastrointestinal tract, the impairment generated after surgery for the stomach, duodenum, esophagus, and for the lower gastrointestinal tract, the impairment generated after construction and surgery for colon, rectum, anus, and intestinal stomas. We tried to make the rating impairment guidelines to include science, objectivity, convenience, rationality, and actuality. We especially emphasized objectivity as the most important value. We worked to make it easy and convenient to use for both the subjects who received the impairment ratings and the doctors who will give the ratings.

Pathogenesis of portal hypertensive gastropathy: translating basic research into clinical practice.

Portal hypertensive gastropathy (PHG) is often seen in patients with portal hypertension, and can lead to transfusion-dependent anemia as well as acute, life-threatening bleeding episodes. This Review focuses on the mechanisms that underlie the pathogenesis of PHG that provide reasonable grounds for the treatment of this condition, and ultimately enable translation of basic research into clinical practice. Increased portal pressure associated with cirrhosis and liver dysfunction is critical for the development of clinically significant PHG, and leads to impaired gastric mucosal defense mechanisms that render the stomach susceptible to mucosal injury. The use of pharmacological agents such as beta-blockers reduces the frequency of bleeding episodes in PHG. As a last resort, surgical decompression of the portal system, transjugular intrahepatic stent placement and liver transplantation can resolve this condition. Elimination of known risk factors for gastric injury such as alcohol, aspirin and traditional NSAIDs is critical. The role of Helicobacter pylori colonization of the gastric mucosa in PHG is not clear. Careful and critical interpretation of human and experimental data can be helpful to establish a rationale for the medical management of this important condition.

Development of a Rating System for Digestive System Impairments: Korean Academy of Medical Sciences Guideline

A systematic and effective welfare system for people with digestive system impairments is required. In Korea, an objective and scientific rating guideline does not exist to judge the digestive system impairments. Whether the impairments exist or not and the degree of it need to be examined. Thus, with these considerations we need a scientific rating guideline for digestive system impairments to fit our cultural and social background. In 2007, a research team, for the development of rating impairment guidelines, was organized under the supervision of Korean Academy of Medical Sciences. The rating guidelines for digestive system impairments was classified into upper and lower gastrointestinal tracts impairments and liver impairment. We developed objective rating guidelines for the upper gastrointestinal tract, the impairment generated after surgery for the stomach, duodenum, esophagus, and for the lower gastrointestinal tract, the impairment generated after construction and surgery for colon, rectum, anus, and intestinal stomas. We tried to make the rating impairment guidelines to include science, objectivity, convenience, rationality, and actuality. We especially emphasized objectivity as the most important value. We worked to make it easy and convenient to use for both the subjects who received the impairment ratings and the doctors who will give the ratings.

Gastric polyps: classification and management.

Gastric polyps can be broadly defined as luminal lesions projecting above the plane of the mucosal surface. They are relatively frequent in routine pathology practice, where the main goal is to rule out the possibility of malignancy. Various subtypes of gastric polyps are recognized and generally divided into nonneoplastic and neoplastic. We will review herein only a limited subset of gastric polyps representing the most common or, sometimes, challenging.

Image-based discriminating morphological features for gastric atrophy assessment: a step to go further.

The aim of this study is to establish a basis for automated assessment of gastric atrophy according to the Updated Sydney System. We sought to minimize inter- and intra-observer variations in the application of the Sydney System. A total of 160 biopsies were examined by three pathologists and graded using the visual scale of the Updated Sydney System. A consensus was reached on 135 biopsies. Digital images were captured for the studied biopsies. Image processing techniques were used to extract four morphological features that uniquely discriminate each atrophy grade. The features are related to gland density and shape. To validate the reproducibility of these features, the K-Means clustering technique was used. We were able to grade the atrophy with an overall precision of 95.6%. Furthermore, the proposed features were able to distinguish four discrete grades without any significant overlap. This has not been achieved by previous studies.