Endoscopic Identification Research Articles

The Value of Endoscopic Exposure of Round Window in Cochlear Implant via Facial Recess

Abstract Background Cochlear implantation (CI) has been considered as the best treatment in patients with severe to profound hearing loss unaidable with hearing aids. The main value of endoscope-assisted cochlear implantation is improved visibility of the RW and reduced risk of complications, especially in patients with congenital malformations. Aim of the Work The aim of the present study was to assess the value of endoscopic assisted CI surgery via facial recess approach without elevating tympanic anulus. Methods This Prospective case series study non-randomized sample was performed on a total of 50 patients with severe to profound hearing loss unaidable with hearing aids undergoing unilateral endoscopic assisted cochlear implant surgery with round window electrode insertion at Ain Shams University hospitals from April 2020 till March 2022 and approved 94/2020MD by an ethical committee before the start of the recruitment. Results 50 CI operations with endoscopic assistance were prospectively evaluated. There were 23 male (46%) and 27 female (54%) patients. The majority of the cases were children (41 cases, 82%). Of those 50 patients, 39 (78%) were prelingually hearing impaired. Four cases (8%) had various inner ear abnormalities. The standard mastoidectomy and Posterior Tympanotomy (PT) approach were used for all cases. Med-El (FLEX28TM; FLEX24TM Med-El, Innsbruck, Australia), Advanced Bionics (AB)Mid-Scala Electrode and CochlearTM (NUCLEUS® CI422 WITH SLIM STRAIGHT, NUCLEUS® CI24RE CONTOUR ADVANCETM, Sydney, Australia) implants were used. Endoscope-assisted visualization of the RW region through the PT was performed in all cases Rigid endoscopes with 0o and 30o (1.9, 2.7 and 4 mm in width, 11 and 6 cm in length) and an HD (high-definition) camera system (Karl Storz, Tuttlingen, Germany) were used for the endoscopic evaluation. The RW region was visualized and identified by inserting the endoscope through the PT without elevating the tympanic annulus. Then we switched to the microscope and drilled the RW niche (tegmen) using both hands until a circumferential view of the RW membrane is obtained. After proper exposure, the electrode was inserted through an opening in the RW membrane mostly under microscopical vision. Endoscopic identification of the RW through the PT enabled us to perform regular surgery in all cases. The current study concludes the difference between microscopic exposure and endoscopic exposure represented by Saint Tomas classification regarding round window exposure found that endoscopic exposure of round window classification is better represented by downgrading in the classification of round window exposure as type I 29(58%), type IIa 18(36%) type IIb 3 (6%) Non were type III by endoscopic exposure compared to microscopic exposure of round window is a type I 7(14%), type II 14(28%), type IIb 22(44%) and type III 7 (14%). Conclusion As evident from the current study, Endoscopy proved a great value in exposure and identification of RW in CI surgery through posterior tympanotomy approach, especially in difficult cases. Endoscopic-assisted CI through facial recess could be easily done using 1.9mm 0-degree endoscopy even in cases e narrow PT or rotated cochlea

Complications after Heller myotomy in children: a national multicenter study on the impact of prior endoscopic dilatation and identification of risk factors.

Although esophageal achalasia has been historically treated by Heller myotomy, endoscopic esophageal dilatations are nowadays often the first-line treatment in children. The aim was to assess whether performing an endoscopic dilatation before a Heller myotomy is associated with higher risks of esophageal perforation in children. A retrospective multicentric study was performed, including children that underwent a Heller myotomy (2000-2022, 10 centers). Two groups were compared based on the history of previous dilatation before myotomy. Outcomes esophageal perforation (intra-operative or secondary) and post-operative complications requiring surgery (Clavien-Dindo III). Statistics Comparisons using contingency tables or Kruskal-Wallis when appropriate. Statistical significance: p-value < 0.05. A Heller myotomy was performed in 77 children (median age: 11.8 years), with prior endoscopic dilatation in 53% (n = 41). A laparoscopic approach was used in 90%, with associated fundoplication in 95%. Esophageal perforation occurred in 19% of children (n = 15), including 12 patients with intra-operative mucosal tear and 3 with post-operative complications related to an unnoticed esophageal perforation. Previous endoscopic dilatation did not increase the risk of esophageal perforation (22% vs 17%, OR: 1.4, 95%CI: 0.43-4.69). Post-operative complications occurred in 8% (n = 6), with similar rates regardless of prior endoscopic dilatation. Intra-operative mucosal tear was the only risk factor for post-operative complications, increasing the risk of complications from 5 to 25% (OR: 6.89, 95%CI: 1.38-31.87). Prior endoscopic dilatations did not increase the risk of esophageal perforation or postoperative complications of Heller myotomy in this cohort of children with achalasia. Mucosal tear was identified as a risk factor for post-operative complications.

Visualization of the nervus intermedius during microvascular decompression in hemifacial spasm: anatomical study.

The surgical anatomy of the nervus intermedius (NI) is highly variable. The aim of this study was to describe the anatomy of the NI during endoscope-assisted microvascular decompression (MVD) in hemifacial spasm (HFS), and the involvement of the nerve in the vascular conflict. The authors reviewed a prospectively maintained database for MVDs performed between 2002 and 2022 and extracted clinical data including patient demographics, symptoms, and offending vessel(s). Operative videos and photographs were analyzed retrospectively in an attempt to identify the NI. Endoscopic identification of the NI was possible in 139 of 435 MVDs. The anatomy is very variable. In 79 (56.8%) patients, a single-bundle pattern was detected, whereas a multiple-bundle pattern was identified in 60 (43.2%) patients. Overall the most common pattern was a single-bundle type A (49.7%). In 20.1%, a multiple-bundles type A was identified. In 4.3%, a single-bundle type B was detected. In 2.9% a single-bundle type C was found, and in just 0.7% a multiple-bundles type C was detected. A multiple-origin pattern (type D) was found in 31 patients (22.3%). The NI was frequently involved in the neurovascular conflict (approximately 85%). The type of NI or vascular compression pattern did not affect the results regarding the outcome or recurrence of HFS. The anatomy of the NI is for the first time evaluated endoscopically in MVD for HFS. The nerve had various anatomical patterns that were clearly identified. Further studies to evaluate the compression patterns in relation to NI neuralgia are warranted.

Surveillance Endoscopy in Pediatric Esophageal Atresia: Toward an Evidence-Based Algorithm.

Individuals with esophageal atresia (EA) have lifelong increased risk for mucosal and structural pathology of the esophagus. The use of surveillance endoscopy to detect clinically meaningful pathology has been underexplored in pediatric EA. We hypothesized that surveillance endoscopy in pediatric EA has high clinical yield, even in the absence of symptoms. The medical records of all patients with EA who underwent at least 1 surveillance endoscopy between March 2004 and March 2023 at an international EA referral center were retrospectively reviewed. The primary outcomes were endoscopic identification of pathology leading to an escalation in medical, endoscopic, or surgical management. Logistic regression analysis examined predictors of actionable findings. Nelson-Aalen analysis estimated optimal endoscopic surveillance intervals. Five hundred forty-six children with EA underwent 1,473 surveillance endoscopies spanning 3,687 person-years of follow-up time. A total of 770 endoscopies (52.2%) in 394 unique patients (72.2%) had actionable pathology. Esophagitis leading to escalation of therapy was the most frequently encountered finding (484 endoscopies, 32.9%), with most esophagitis attributed to acid reflux. Barrett's esophagus (intestinal metaplasia) was identified in 7 unique patients (1.3%) at a median age of 11.3 years. No dysplastic lesions were identified. Actionable findings leading to surgical intervention were found in 55 children (30 refractory reflux and 25 tracheoesophageal fistulas). Significant predictors of actionable pathology included increasing age, long gap atresia, and hiatal hernia. Symptoms were not predictive of actionable findings, except dysphagia, which was associated with stricture. Nelson-Aalen analysis predicted occurrence of an actionable finding every 5 years. Surveillance endoscopy uncovers high rates of actionable pathology even in asymptomatic children with EA. Based on the findings of the current study, a pediatric EA surveillance endoscopy algorithm is proposed.

An artificial intelligence system for chronic atrophic gastritis diagnosis and risk stratification under white light endoscopy

Background and AimsThe diagnosis and stratification of gastric atrophy (GA) predict patients’ gastric cancer progression risk and determine endoscopy surveillance interval. We aimed to construct an artificial intelligence (AI) system for GA endoscopic identification and risk stratification based on the Kimura-Takemoto classification. MethodsWe constructed the system using two trained models and verified its performance. First, we retrospectively collected 869 images and 119 videos to compare its performance with that of endoscopists in identifying GA. Then, we included original image cases of 102 patients to validate the system for stratifying GA and comparing it with endoscopists with different experiences. ResultsThe sensitivity of model 1 was higher than that of endoscopists (92.72% vs. 76.85 %) at image level and also higher than that of experts (94.87% vs. 85.90 %) at video level. The system outperformed experts in stratifying GA (overall accuracy: 81.37 %, 73.04 %, p = 0.045). The accuracy of this system in classifying non-GA, mild GA, moderate GA, and severe GA was 80.00 %, 77.42 %, 83.33 %, and 85.71 %, comparable to that of experts and better than that of seniors and novices. ConclusionsWe established an expert-level system for GA endoscopic identification and risk stratification. It has great potential for endoscopic assessment and surveillance determinations.

The Value of Endoscopic Exposure of Round Window in Cochlear Implant via Facial Recess.

Introduction Cochlear implantation has been considered as the best treatment in patients with severe to profound hearing loss unaidable with hearing aids. The main value of endoscope-assisted cochlear implantation is improved visibility of the RW Objective to assess the value of endoscopic assisted CI surgery via facial recess approach without elevating tympanic anulus. Methods This Prospective case series study non-randomized sample was performed on 50 patients with severe to profound hearing loss unaidable with hearing aids undergoing unilateral endoscopic assisted cochlear implant surgery with round window electrode insertion Results There were 23 male and 27 female patients. Most of the cases were children (41 cases). Of those 50 patients, 39 were prelingually hearing impaired. Four cases had various inner ear abnormalities. The standard mastoidectomy and Posterior Tympanotomy approach were used for all cases. Endoscopic identification of the RW through the PT enabled us to perform regular surgery in all cases. The current study concludes the difference between microscopic exposure and endoscopic exposure represented by Saint Tomas classification found that endoscopic exposure of round window classification is better represented by downgrading in the classification of round window exposure as type I 29(58%), type IIa 18(36%) type IIb 3 (6%) Non were type III by endoscopic exposure compared to microscopic exposure of round window is a type I 7(14%), type II 14(28%), type IIb 22(44%) and type III 7 (14%). Conclusion Endoscopy proved a great value in exposure and identification of RW in CI surgery through posterior tympanotomy approach.

Endoscopic Stone Composition Identification: Is Accuracy Improved by Stone Appearance During Laser Lithotripsy?

To evaluate if videos during laser lithotripsy increase accuracy and confidence of stone identification by urologists compared to still pictures. We obtained representative pictures and videos of 4 major stone types from 8 different patients during ureteroscopy with holmium laser lithotripsy. A REDCap survey was created and emailed to members of the Endourological Society. The survey included a picture followed by the corresponding video of each stone undergoing laser lithotripsy and additional clinical information. Each picture and video included multiple-choice questions about stone composition and response confidence level. Accuracy, confidence levels, and rates of rectification (change from incorrect to correct answer) or confounding (correct to incorrect) after watching videos were analyzed. One hundred eighty-seven urologists responded to the survey. The accuracy rate of stone identification with pictures was 43.8% vs 46.1% with videos (P=.27). Accuracy for individual stones was low and highly variable. Video only improved accuracy for 1 cystine stone. After viewing videos, participants were more likely to rectify vs confound their answers. Urologists were more likely to be confident with videos than pictures alone (65.4% vs 53.7%, respectively; P<.001); however, confident answers were not more likely to yield accurate predictions with videos vs still pictures. Stone identification by urologists is marginally improved with videos vs pictures alone. Overall, accuracy in stone identification is low irrespective of confidence level, picture, and lithotripsy video visualization. Urologists should be cautious in using endoscopic stone appearance to direct metabolic management.

Semantic Segmentation of Gastric Polyps in Endoscopic Images Based on Convolutional Neural Networks and an Integrated Evaluation Approach.

Convolutional neural networks (CNNs) have received increased attention in endoscopic images due to their outstanding advantages. Clinically, some gastric polyps are related to gastric cancer, and accurate identification and timely removal are critical. CNN-based semantic segmentation can delineate each polyp region precisely, which is beneficial to endoscopists in the diagnosis and treatment of gastric polyps. At present, just a few studies have used CNN to automatically diagnose gastric polyps, and studies on their semantic segmentation are lacking. Therefore, we contribute pioneering research on gastric polyp segmentation in endoscopic images based on CNN. Seven classical semantic segmentation models, including U-Net, UNet++, DeepLabv3, DeepLabv3+, Pyramid Attention Network (PAN), LinkNet, and Muti-scale Attention Net (MA-Net), with the encoders of ResNet50, MobineNetV2, or EfficientNet-B1, are constructed and compared based on the collected dataset. The integrated evaluation approach to ascertaining the optimal CNN model combining both subjective considerations and objective information is proposed since the selection from several CNN models is difficult in a complex problem with conflicting multiple criteria. UNet++ with the MobineNet v2 encoder obtains the best scores in the proposed integrated evaluation method and is selected to build the automated polyp-segmentation system. This study discovered that the semantic segmentation model has a high clinical value in the diagnosis of gastric polyps, and the integrated evaluation approach can provide an impartial and objective tool for the selection of numerous models. Our study can further advance the development of endoscopic gastrointestinal disease identification techniques, and the proposed evaluation technique has implications for mathematical model-based selection methods for clinical technologies.

Infection with male and female Trichuris trichiura diagnosed in a non‐endemic area

Trichuris trichiura parasitizes only humans through fecal-oral transmission. In non-endemic areas, the frequency of endoscopic identification has been increasing due to the increasing number of immigrants from endemic countries. To prevent infection, it is important to pay attention to sanitary conditions such as soil and water sources.

Colonic diverticular bleeding: An update on pathogenesis and management

Colonic diverticular bleeding is the most common cause of lower gastrointestinal (GI) bleeding, which can be life-threatening and frequently recurrent. In recent years, the prevalence of diverticulosis has increased in developed countries, with a documented incidence of 50% in patients older than 60 years. Based on the evidence, the use of anticoagulants and/or antiplatelets in the elderly population has resulted in an increased incidence of acute diverticular bleeding. According to the literature, about 50% of patients with diverticular bleeding require a blood transfusion, and 18% – 53% need emergency surgery. Although endoscopic identification of the culprit diverticula and appropriate intervention is a challenge, the newer treatment modality, over-the-scope clip method (OTSC) has been demonstrated to be an effective endoscopic hemostatic method in severe diverticular bleeding, especially in cases of rebleeding after first-line conventional endoscopic procedures. In this review, we summarize the pathophysiology of colonic diverticulosis and diverticular bleeding, recent evidence in its management, and existing theories on various preventive strategies to control diverticular bleeding. We also discuss the efficacy and treatment outcome of the OTSC technique in controlling diverticular bleeding.

WEO Newsletter.

WEO Newsletter.

Fourth branchial cleft cyst in an infant. Case report and modified surgical approach

Fourth branchial cleft anomalies are rare anomalies with a wide variety of clinical symptomatology and different surgical approaches. These approaches has been described with variable success and rates of complications with no standard surgical approach described. In this case report we present an infant with left sided fourth branchial cleft cyst treated with a modified surgical approach combining endoscopic identification of the inner opening and insertion of a fine guidewire followed by surgical excision and ligation of the fistula tract.

Endoscopic application of fibrin glue may be a feasible method of treatment for postintubation tracheal lacerations in cats.

To determine the feasibility of endoscopic application of fibrin glue for the treatment of experimentally induced postintubation tracheal laceration (PITL) in feline cadavers. The secondary objective was to determine the optimal technique for application of the fibrin glue. 20 feline cadavers (n = 10 fresh and 10 frozen). An experimentally induced tracheal rupture was created via overinflation of an endotracheal tube cuff. After endoscopic identification of the tracheal tear, fibrin glue was instilled into the tracheal defect in either a bridging or filling fashion. Following the procedure, the airway of each cat was examined and leak tested. Length of tear, volume of glue applied, procedural time, and glue efficacy were recorded. Experimentally induced tracheal lacerations were full thickness with a mean length of 3.27 ± 0.96 cm. A complete seal was attained in 6 of the 9 fresh cadavers when filling the defect with fibrin glue. In the remaining 3 fresh cadavers, air leakage was restricted to the dorsal mediastinum. Bridging the defect with fibrin glue did not attain a seal in fresh or frozen cadavers. The median volume of glue used to fill defects in fresh cadavers was 0.5 mL (range, 0.4 to 2 mL). Procedural time for the application of fibrin glue was 10.5 ± 4.1 minutes for bridging the defect and 7.8 ± 1.5 minutes for filling the defect. Endoscopic application of fibrin glue may be a feasible method of treatment for PITL in cats.

A Makeshift Blue Light Filter for Endoscopic Identification of Traumatic Cerebrospinal Fluid Leak Using Fluorescein

&#x0D; &#x0D; &#x0D; &#x0D; ABSTRACTObjective: To describe a makeshift blue light filter for endoscopic visualization of a traumatic cerebrospinal fluid leak repair using intrathecal fluorescein and its application in one patient.&#x0D; &#x0D; &#x0D; &#x0D; &#x0D; Methods:Study Design:Surgical Instrumentation&#x0D; Setting:Tertiary Government Training Hospital&#x0D; Patient:One&#x0D; &#x0D; &#x0D; &#x0D; &#x0D; Results: Intra-operative endoscopic identification of fistulae sites was achieved using intrathecal injection of fluorescein that fluoresced using our makeshift blue light filter in a 43-year-old man who presented with a 3-month history of rhinorrhea due to skull base fractures along with multiple facial and upper extremity fractures he sustained after a fall from a standing height of 6 feet. He underwent transnasal endoscopic repair of cerebrospinal fluid fistulae in the planum sphenoidale, clivus and sellar floor. Post-operatively, there was complete resolution of rhinorrhea with no complications noted.&#x0D; &#x0D; &#x0D; &#x0D; &#x0D; Conclusion: Our makeshift blue light filter made from readily available materials may be useful for endoscopic identification of CSF leaks using fluorescein in a low- to middle-income country setting like ours.&#x0D; &#x0D; &#x0D; &#x0D; &#x0D; &#x0D; &#x0D; &#x0D; &#x0D; &#x0D; &#x0D; &#x0D; &#x0D; &#x0D; &#x0D; &#x0D;

Development of a minimally invasive endoscopic technique for excisional biopsy of the axillary lymph nodes in dogs.

To develop and describe a minimally invasive technique for excisional biopsy of the axillary lymph nodes in dogs. Descriptive cadaver and clinical case series. Four canine cadavers and three clinical patients. A 3D computed tomographic reconstruction of the canine axilla was used to identify an optimal avenue of approach to the lymph nodes. This approach was refined using endoscopic techniques in four cadavers (six procedures) and potential surgical hazards, landmarks, and the surgical time required for excisional biopsy of the nodes was recorded. The procedure was then performed in three clinical cases. Axillary lymph node removal was achieved using an endoscopic technique with surgical times of 58 and 35 minutes in two of three clinical cases. The third case required conversion to an open approach after endoscopic identification of the node. No major complications were encountered. Excisional biopsy of the axillary lymph nodes can be performed successfully using a minimally invasive technique in the dog. Further investigation in clinical cases is needed to determine the risks and complications of this procedure. Minimally invasive excisional biopsy of the axillary lymph nodes in dogs can be performed and may have a role in assisting with staging and local disease control in oncologic cases.

S2084 An Unusual Cause of Rectal Bleeding: Colonic Angiolipoma

Introduction: Angiolipomas are rare, benign, subcutaneous tumors composed of adipose tissue and vascular components. Angiolipomas are extremely rare in the colon, accounting for less than 1% of all colonic benign lesions. While these tumors are typically painful, they may also present with rectal bleeding or positive fecal occult blood test in an otherwise asymptomatic patient. Therefore, it is important to recognize the clinical, radiological, and endoscopic findings associated with these masses. Case Description/Methods: A 49-year-old male with hypertension and hyperlipidemia presented for evaluation of 3 days of self-limited rectal bleeding. The patient had no prior colonoscopies and denied associated symptoms of abdominal pain, diarrhea, or constipation. On physical examination, the abdomen was soft, non-tender, and non-distended with no palpable masses. Colonoscopy was remarkable for a malignant-appearing, large, ulcerated, non-bleeding mass in the hepatic flexure (Figure 1). Biopsy of the mass demonstrated features of focal active colitis and erosion without evidence of dysplasia or malignancy. The patient was referred to oncology for further evaluation of the mass. Positron emission tomography-computed tomography (PET-CT) results were remarkable for abnormal thickened appearance of the colon at the hepatic flexure with an adjacent 4.4 cm fatty structure in the proximal transverse colon without fluorodeoxyglucose radiotracer uptake. The patient was evaluated by colorectal surgery and received a robotic-assisted laparoscopic right colectomy with an isoperistaltic ileotransverse colon anastomosis. The specimen was sent to pathology and results demonstrated a 4.4 × 4.0 × 4.0 cm protruding mass with glanular mucosa and submucosal fatty cut surfaces, consistent with angiolipoma. Immunohistochemical (IHC) staining was unremarkable for angiomyolipoma. The patient tolerated the procedure well without significant sequelae. Discussion: Angiolipomas are typically painful yet benign tumors that are rarely found in the gastrointestinal tract. Although abdominal pain, rectal bleeding, and/or positive fecal occult blood testing are commonly reported on presentation, diagnosis is complicated as symptoms tend to be non-specific. Preoperative diagnosis of angiolipoma is possible when aided by CT, endoscopy, and MRI, but surgical pathology remains the gold standard for final diagnosis of the tumor. When angiolipomas are fully excised, the prognosis is excellent.Figure 1.: Endoscopic identification of malignant-appearing mass in hepatic flexure.

Detection and Characterization of Gastric Cancer Using Cascade Deep Learning Model in Endoscopic Images.

Endoscopy is widely applied in the examination of gastric cancer. However, extensive knowledge and experience are required, owing to the need to examine the lesion while manipulating the endoscope. Various diagnostic support techniques have been reported for this examination. In our previous study, segmentation of invasive areas of gastric cancer was performed directly from endoscopic images and the detection sensitivity per case was 0.98. This method has challenges of false positives and computational costs because segmentation was applied to all healthy images that were captured during the examination. In this study, we propose a cascaded deep learning model to perform categorization of endoscopic images and identification of the invasive region to solve the above challenges. Endoscopic images are first classified as normal, showing early gastric cancer and showing advanced gastric cancer using a convolutional neural network. Segmentation on the extent of gastric cancer invasion is performed for the images classified as showing cancer using two separate U-Net models. In an experiment, 1208 endoscopic images collected from healthy subjects, 533 images collected from patients with early stage gastric cancer, and 637 images from patients with advanced gastric cancer were used for evaluation. The sensitivity and specificity of the proposed approach in the detection of gastric cancer via image classification were 97.0% and 99.4%, respectively. Furthermore, both detection sensitivity and specificity reached 100% in a case-based evaluation. The extent of invasion was also identified at an acceptable level, suggesting that the proposed method may be considered useful for the classification of endoscopic images and identification of the extent of cancer invasion.

Endoscopic sexing in turtles and tortoises: 467 cases (2007-2017).

This multi-institutional retrospective study evaluated the feasibility and safety of endoscopic sex identification in 467 turtles and tortoises, representing 10 species. Medical records of turtles and tortoises that underwent endoscopic sex identification at the University of Georgia, New England Aquarium and Turtle Conservancy were reviewed for presurgical management, anaesthesia, endoscopic equipment and surgical techniques, endoscopic results and complications. The majority of animals weighed less than 200g, were fasted and anaesthetised using an injectable combination of ketamine, dexmedetomidine and morphine or hydromorphone, supplemented by local lidocaine at the prefemoral site. Anaesthetic reversal using atipamezole alone or in combination with naloxone was routine. For uncomplicated procedures, mean total anaesthesia, surgery and recovery times were 22, 4 and 18minutes, respectively. All animals were placed in lateral recumbency for a prefemoral endoscopic approach to the coelom using a rigid telescope and sterile fluid infusion to visualise the gonads. Sex identification was definitive in 99.4% (n=464) of the animals. Iatrogenic bladder perforation was the most common complication (n=5), which necessitated extended anaesthesia and surgical time for repair. Only a single anaesthetic-related death was reported, which was associated with human error and drug overdose. This is the first large-scale study to retrospectively evaluate endoscopic sex identification in multiple chelonian species. Results suggest that endoscopic sexing is a safe, accurate and practical means for sex identification in turtles and tortoises, and represents a valuable tool in their reproductive management.

Current Status and Future Perspectives of Artificial Intelligence in Colonoscopy.

The early endoscopic identification, resection, and treatment of precancerous adenoma and early-stage cancer has been shown to reduce not only the prevalence of colorectal cancer but also its mortality rate. Recent advances in endoscopic devices and imaging technology have dramatically improved our ability to detect colorectal lesions and predict their pathological diagnosis. In addition to this, rapid advances in artificial intelligence (AI) technology mean that AI-related research and development is now progressing in the diagnostic imaging field, particularly colonoscopy, and AIs (i.e., devices that mimic cognitive abilities, such as learning and problem-solving) already approved as medical devices are now being introduced into everyday clinical practice. Today, there is an increasing expectation that sophisticated AIs will be able to provide high-level diagnostic performance irrespective of the level of skill of the endoscopist. In this paper, we review colonoscopy-related AI research and the AIs that have already been approved and discuss the future prospects of this technology.

Application of Convolutional Neural Networks for Diagnosis of Eosinophilic Esophagitis Based on Endoscopic Imaging.

Subjective symptoms associated with eosinophilic esophagitis (EoE), such as dysphagia, are not specific, thus the endoscopic identification of suggestive EoE findings is quite important for facilitating endoscopic biopsy sampling. However, poor inter-observer agreement among endoscopists regarding diagnosis has become a complicated issue, especially with inexperienced practitioners. Therefore, we constructed a computer-assisted diagnosis (CAD) system using a convolutional neural network (CNN) and evaluated its performance as a diagnostic utility. A CNN-based CAD system was developed based on ResNet50 architecture. The CNN was trained using a total of 1192 characteristic endoscopic images of 108 patients histologically proven to be in an active phase of EoE (≥15 eosinophils per high power field) as well as 1192 normal esophagus images. To evaluate diagnostic accuracy, an independent test set of 756 endoscopic images from 35 patients with EoE and 96 subjects with a normal esophagus was examined with the constructed CNN. The CNN correctly diagnosed EoE in 94.7% using a diagnosis per image analysis, with an overall sensitivity of 90.8% and specificity of 96.6%. For each case, the CNN correctly diagnosed 37 of 39 EoE cases with overall sensitivity and specificity of 94.9% and 99.0%, respectively. These findings indicate the usefulness of CNN for diagnosing EoE, especially for aiding inexperienced endoscopists during medical check-up screening.