Influence of artificial intelligence on the diagnostic performance of endoscopists in the assessment of Barrett's esophagus: a tandem randomized and video trial.

BACKGROUND: This study evaluated the effect of an artificial intelligence (AI)-based clinical decision support system on the performance and diagnostic confidence of endoscopists in their assessment of Barrett's esophagus (BE). METHODS: 96 standardized endoscopy videos were assessed by 22 endoscopists with varying degrees of BE experience from 12 centers. Assessment was randomized into two video sets: group A (review first without AI and second with AI) and group B (review first with AI and second without AI). Endoscopists were required to evaluate each video for the presence of Barrett's esophagus-related neoplasia (BERN) and then decide on a spot for a targeted biopsy. After the second assessment, they were allowed to change their clinical decision and confidence level. RESULTS: AI had a stand-alone sensitivity, specificity, and accuracy of 92.2%, 68.9%, and 81.3%, respectively. Without AI, BE experts had an overall sensitivity, specificity, and accuracy of 83.3%, 58.1%, and 71.5%, respectively. With AI, BE nonexperts showed a significant improvement in sensitivity and specificity when videos were assessed a second time with AI (sensitivity 69.8% [95%CI 65.2%-74.2%] to 78.0% [95%CI 74.0%-82.0%]; specificity 67.3% [95%CI 62.5%-72.2%] to 72.7% [95%CI 68.2%-77.3%]). In addition, the diagnostic confidence of BE nonexperts improved significantly with AI. CONCLUSION: BE nonexperts benefitted significantly from additional AI. BE experts and nonexperts remained significantly below the stand-alone performance of AI, suggesting that there may be other factors influencing endoscopists' decisions to follow or discard AI advice.

EUS-guided Gastroenterostomy: A Multicenter International Study Comparing Benign and Malignant Diseases.

BACKGROUND: Endoscopic ultrasound (EUS)-guided gastroenterostomy (EUS-GE) is a minimally invasive therapy for patients with gastric outlet obstruction without the risks of surgical bypass and the limited long-term efficacy of enteral self-expanding metal stent placement. However, due to its novelty, there is a lack of significant data comparing long-term outcomes of patients with EUS-GE, based on the underlying disease. In this study, we compare outcomes of EUS-GE on benign versus malignant indications. METHODS: Consecutive patients from 12 international, tertiary care centers who underwent EUS-GE over 3 years were extracted in a retrospective registry. Demographic characteristics, procedure-related information and follow-up data was collected. Primary outcome was the rate of adverse events associated with EUS-GE and the comparison of the rate of adverse events in benign versus malignant diseases. Secondary outcomes included technical and clinical success as well as hospitalization admission. RESULTS: A total of 103 patients were included: 72 malignant and 31 benign. The characteristics of the patients undergoing EUS-GE is shown in Table 1. The mean age of the cohort was 68 years and 58 years for malignant and benign etiology. Gender distribution was 57% and 39% being females in malignant and benign etiology group, respectively. Clinical success, technical success, average procedure time, and hospital length of stay were similar in both groups. Patients with benign underlying etiology had significantly higher number of surgically altered midgut anatomy (P=0.0379). CONCLUSION: EUS-GE is equally efficient regardless of the underlying etiology (malignant vs. benign), and the adverse events both groups were comparable.

Application of Deep Learning Models to Improve Ulcerative Colitis Endoscopic Disease Activity Scoring Under Multiple Scoring Systems.

BACKGROUND AND AIMS: Lack of clinical validation and inter-observer variability are two limitations of endoscopic assessment and scoring of disease severity in patients with ulcerative colitis [UC]. We developed a deep learning [DL] model to improve, accelerate and automate UC detection, and predict the Mayo Endoscopic Subscore [MES] and the Ulcerative Colitis Endoscopic Index of Severity [UCEIS]. METHODS: A total of 134 prospective videos [1550 030 frames] were collected and those with poor quality were excluded. The frames were labelled by experts based on MES and UCEIS scores. The scored frames were used to create a preprocessing pipeline and train multiple convolutional neural networks [CNNs] with proprietary algorithms in order to filter, detect and assess all frames. These frames served as the input for the DL model, with the output being continuous scores for MES and UCEIS [and its components]. A graphical user interface was developed to support both labelling video sections and displaying the predicted disease severity assessment by the artificial intelligence from endoscopic recordings. RESULTS: Mean absolute error [MAE] and mean bias were used to evaluate the distance of the continuous model's predictions from ground truth, and its possible tendency to over/under-predict were excellent for MES and UCEIS. The quadratic weighted kappa used to compare the inter-rater agreement between experts' labels and the model's predictions showed strong agreement [0.87, 0.88 at frame-level, 0.88, 0.90 at section-level and 0.90, 0.78 at video-level, for MES and UCEIS, respectively]. CONCLUSIONS: We present the first fully automated tool that improves the accuracy of the MES and UCEIS, reduces the time between video collection and review, and improves subsequent quality assurance and scoring.

Antireflux band mucosectomy: a novel minimally invasive approach for the treatment of refractory gastroesophageal reflux disease.

Background and Aims: Endoscopic therapies in the treatment of refractory GERD have largely been regarded as inferior as surgical intervention. Procedures such as the transoral endoscopic incisionless fundoplication (TIF), Stretta, and antireflux mucosectomy (ARMS) are less invasive but produce outcomes that are middling to lackluster, with many patients having to continue proton pump therapy without resolution of symptoms. Antireflux band mucosectomy (ARBM), in which the cardia is banded, may provide more effective relief. We present 4 patients with refractory GERD who did not opt for surgical intervention and were successfully treated with the ARBM procedure. Methods: Four patients with GERD refractory to medical therapy underwent ARBM. Three patients had nonerosive esophagitis and 1 patient had erosive esophagitis. Two patients had hiatal hernias ≤2 cm. All underwent preprocedure and postprocedure pH bravo testing as well as follow-up EGD. Results: All cases were completed successfully. All patients underwent a decrease in DeMeester score and acid exposure time. Procedure time ranged from 6 to 15 minutes. There were no adverse events. All patients were taken off proton pump inhibitor (PPI) therapy by 4 weeks postprocedure. Conclusions: Four patients with refractory GERD who declined surgery underwent the ARBM procedure with technical success. All experienced significant decrease to complete resolution of symptoms. All patients were weaned off PPIs by 4 weeks postprocedure.

POEM, GPOEM, and ZPOEM.

Our tripartite narrative review discusses Peroral Endoscopic Myotomy (POEM), gastric POEM (GPOEM) and POEM for Zenker's diverticula (ZPOEM). POEM is the prototypical procedure that launched the novel "3rd space endoscopy" field of advanced endoscopy. It revolutionized achalasia therapy by offering a much less invasive version of the prior gold standard, the laparoscopic Heller myotomy (HM). We review in detail indications, outcomes, technique variations and comparative data between POEM and HM particularly with regard to the hotly debated issue of GERD. We then proceed to discuss two less illustrious but nevertheless important offshoots of the iconic POEM procedure: GPOEM for gastroparesis and ZPOEM for the treatment of hypopharyngeal diverticula. For GPOEM, we discuss the rationale of pylorus-directed therapies, briefly touch on GPOEM technique variations and then focus on the importance of proper patient selection and emerging data in this area. On the third and final part of our review, we discuss ZPOEM and expound on technique variations including our "ultra-short tunnel technique". Our review emphasizes that, despite the superiority of endoscopy over surgery for the treatment of hypopharyngeal diverticula, there is no clear evidence yet of the superiority of the newfangled ZPOEM technique compared to the conventional endoscopic myotomy technique practiced for over two decades prior to the advent of ZPOEM.

Morphologic Severity of Acute Pancreatitis on Imaging Is Independently Associated with Opioid Dose Requirements in Hospitalized Patients.

BACKGROUND: Prior studies have evaluated clinical characteristics associated with opioid dose requirements in hospitalized patients with acute pancreatitis (AP) but did not incorporate morphologic findings on CT imaging. AIMS: We sought to determine whether morphologic severity on imaging is independently associated with opioid dose requirements in AP. METHODS: Adult inpatients with a diagnosis of AP from 2006 to 2017 were reviewed. The highest modified CT severity index (MCTSI) score and the daily oral morphine equivalent (OME) for each patient over the first 7 days of hospitalization were used to grade the morphologic severity of AP and calculate mean OME per day(s) of treatment (MOME), respectively. Multiple regression analysis was used to evaluate the association of MOME with MCSTI. RESULTS: There were 249 patients with AP, of whom 196 underwent contrast-enhanced CT. The mean age was 46 ± 13.6 years, 57.9% were male, and 60% were black. The mean MOME for the patient cohort was 60 ± 52.8 mg/day. MCTSI (ß = 3.5 [95% CI 0.3, 6.7], p = 0.03), early hemoconcentration (ß = 21 [95% CI 4.6, 39], p = 0.01) and first episode of AP (ß = - 17 [95% CI - 32, - 2.7], p = 0.027) were independently associated with MOME. Among the 19 patients undergoing ≥ 2 CT scans, no significant differences in MOME were seen between those whose MCTSI score increased (n = 12) versus decreased/remained the same (n = 7). CONCLUSION: The morphologic severity of AP positively correlated with opioid dose requirements. No difference in opioid dose requirements were seen between those who did versus those who did not experience changes in their morphologic severity.

Peroral endoscopic myotomy for management of cricopharyngeal bars (CP-POEM): a retrospective evaluation.

BACKGROUND: Cricopharyngeal bars (CPBs) are a unique etiology of oropharyngeal dysphagia. Symptomatic patients are managed with endoscopic dilation or surgical myotomy. Cricopharyngeal peroral endoscopic myotomy (CP-POEM) is an emerging technique for the management of dysphagia due to CPBs. This study evaluated technical success, clinical success, adverse events, and long-term recurrence following CP-POEM. METHODS: Consecutive patients who underwent POEM for management of CPBs between May 2015 and December 2020 at four tertiary care centers were included. Primary outcome was clinical success (defined as improvement of dysphagia score to ≤ 1). Secondary outcomes were technical success, rate and severity of adverse events, procedure duration, and symptom recurrence. RESULTS: 27 patients (mean age 69 years; 10 female) underwent CP-POEM during the study period. The most common presenting symptoms at the time of index procedure were dysphagia (26; 96.3 %) and regurgitation (20; 74.1 %). Clinical and technical success were achieved in all patients. Mild/moderate adverse events occurred in two patients (7.4 %). CP-POEM significantly reduced the median dysphagia score. CONCLUSIONS: CP-POEM was a safe and effective treatment for symptomatic CPBs. Although symptom recurrence was low, long-term outcome data are needed. CP-POEM should be considered as a management option for symptomatic CPBs at centers with POEM expertise.

Endoscopic Ultrasound-Guided Liver Biopsy.

Endoscopic ultrasound-guided liver biopsy (EUS-LB) has emerged as a safe and effective alternative to percutaneous and trans-jugular approaches for hepatic tissue acquisition. It has shown superior diagnostic yield for the targeted approach of focal lesions, less sampling variability, improved patient comfort, and safety profile. These advantages have contributed to the increased use of EUS-LB as a technique for obtaining liver tissue. In this review, we provide an update on the recent evidence of EUS-LB for the evaluation of liver disease.

Successful removal of a giant esophageal lipoma with hybrid endoscopic submucosal dissection.

Video 1Endoscopic submucosal dissection of a giant esophageal lipoma.

Artificial intelligence for identification and characterization of colonic polyps.

Colonoscopy remains the gold standard exam for colorectal cancer screening due to its ability to detect and resect pre-cancerous lesions in the colon. However, its performance is greatly operator dependent. Studies have shown that up to one-quarter of colorectal polyps can be missed on a single colonoscopy, leading to high rates of interval colorectal cancer. In addition, the American Society for Gastrointestinal Endoscopy has proposed the "resect-and-discard" and "diagnose-and-leave" strategies for diminutive colorectal polyps to reduce the costs of unnecessary polyp resection and pathology evaluation. However, the performance of optical biopsy has been suboptimal in community practice. With recent improvements in machine-learning techniques, artificial intelligence-assisted computer-aided detection and diagnosis have been increasingly utilized by endoscopists. The application of computer-aided design on real-time colonoscopy has been shown to increase the adenoma detection rate while decreasing the withdrawal time and improve endoscopists' optical biopsy accuracy, while reducing the time to make the diagnosis. These are promising steps toward standardization and improvement of colonoscopy quality, and implementation of "resect-and-discard" and "diagnose-and-leave" strategies. Yet, issues such as real-world applications and regulatory approval need to be addressed before artificial intelligence models can be successfully implemented in clinical practice. In this review, we summarize the recent literature on the application of artificial intelligence for detection and characterization of colorectal polyps and review the limitation of existing artificial intelligence technologies and future directions for this field.

Colorectal polyp characterization with standard endoscopy: Will Artificial Intelligence succeed where human eyes failed?

The American Society for Gastrointestinal Endoscopy (ASGE) has proposed the "resect-and-discard" and "diagnose-and-leave" strategies for diminutive colorectal polyps to reduce the costs of unnecessary polyp resection and pathology evaluation. However, the diagnostic thresholds set by these guidelines are not always met in community practice. To overcome this sub-optimal performance, artificial intelligence (AI) has been applied to the field of endoscopy. The incorporation of deep learning algorithms with AI models resulted in highly accurate systems that match the expert endoscopists' optical biopsy and exceed the ASGE recommended thresholds. Recent studies have demonstrated that the integration of AI in clinical practice results in significant improvement in endoscopists' diagnostic accuracy while reducing the time to make a diagnosis. Yet, several points need to be addressed before AI models can be successfully implemented in clinical practice. In this review, we summarize the recent literature on the application of AI for characterization of colorectal polyps, and review the current limitation and future directions for this field.

Multicenter experience with digital single-operator cholangioscopy in pregnant patients.

Background and study aims The use of fluoroscopy during pregnancy should be minimized given that a clear-cut safe radiation dose in pregnancy is unknown. The role of digital single-operator cholangioscopy (D-SOC) as an alternative to standard enodoscopic retrograde cholangiopancreatography (ERCP) in pregnant patients has not been comprehensively studied. This study assessed 1 Technical success defined as performance of ERCP with D-SOC without the use of fluoroscopy in pregnant patients; 2 safety of D-SOC in pregnancy; and 3 maternal and neonatal outcomes after D-SOC during/after pregnancy. Patients and methods This was an international, multicenter, retrospective study at 6 tertiary centers. Pregnant patients who underwent D-SOC for the treatment of bile duct stones and/or strictures were included. Results A total of 10 patients underwent D-SOC. Indications for ERCP were choledocholithiasis, strictures, previous stent removal, and choledocholithiasis/stent removal. Bile duct cannulation without fluoroscopy was achieved in 10 of 10 patients (100 %). Moreover, 50 % of patients (5/10) completed a fluoroless ERCP with D-SOC. Mean fluoroscopy dose and fluoroscopy time were 3.4 ±â€Š7.2 mGy and 0.5 ±â€Š0.8 min, respectively. One case of mild bleeding and one case of moderate post-ERCP pancreatitis occurred. The mean gestational age at delivery was 36.2 ±â€Š2.6 weeks. Median birth weight was 2.5 kg [IQR: 2.2-2.8]. No birth defects were noted. Conclusion ERCP guided by D-SOC appears to be a feasible and effective alternative to standard ERCP in pregnant patients. It enables avoidance of radiation in half of cases.

Assessment of the learning curve for EUS-guided gastroenterostomy for a single operator.

BACKGROUND AND AIMS: EUS-guided gastroenterostomy (EUS-GE) is increasingly used as an alternative to surgery and enteral stent placement to manage gastric outlet obstruction (GOO). However, no data are available on the learning curve (LC) for EUS-GE. Defining the LC is necessary to create adequate subspecialty training programs and quality assurance. METHODS: This study is a retrospective analysis of a prospectively maintained dataset of patients who underwent EUS-GE at 1 tertiary referral center. Primary outcome was the LC for EUS-GE defined by the number of cases needed to achieve proficiency and mastery using cumulative sum (CUSUM) analysis. Moving average graphs and sequential time-block analysis were also performed to assess procedural time. Secondary outcomes included efficacy and safety of EUS-GE. RESULTS: Eighty-seven consecutive patients underwent EUS-GE, mostly for malignant GOO. For consistency, 14 patients were excluded from analysis (noncautery-assisted EUS-GE, 11; surgical anatomy, 3). The same endoscopist performed all procedures using the same freehand technique. Technical success was achieved in 68 of 73 patients (93%). Immediate adverse events occurred in 4 patients (5.5%), whereas late adverse events occurred only in 1 patient (1%), all managed conservatively or endoscopically. All immediate adverse events occurred during the first 39 cases. Clinical success (defined as resuming at least an oral liquid diet within a week) was achieved in 97% of patients. The mean procedural time was 36 minutes (standard deviation, 24). Evaluation of the CUSUM curve revealed that 25 cases were needed to achieve proficiency and 40 cases to achieve mastery. These results were confirmed with the average moving curve and sequential time-block analysis. CONCLUSIONS: We report, for the first time, data on the LC for EUS-GE. About 25 procedures can be considered as the threshold to achieve proficiency and about 40 cases are needed to reach mastery of the technique.

Long-term outcomes of per-oral endoscopic myotomy in achalasia patients with a minimum follow-up of 4†years: a multicenter study.

Background and study aims Per-oral endoscopic myotomy (POEM) is associated with a short-term clinical response of 82 % to 100 % in treatment of patients with achalasia. Data are limited on the long-term durability of the clinical response in these patients. The aim of this study was to determine the long-term outcomes of patients undergoing POEM for management of achalasia. Methods This was a retrospective multicenter cohort study of consecutive patients who underwent POEM for management of achalasia. Patients had a minimum of 4 years follow-up. Clinical response was defined by an Eckardt score ≤ 3. Results A total of 146 patients were included from 11 academic medical centers. Mean (±â€ŠSD) age was 49.8 (±â€Š16) years and 79 (54 %) were female. The most common type of achalasia was type II, seen in 70 (47.9 %) patients, followed by type I seen in 41 (28.1 %) patients. Prior treatments included: pneumatic dilation in 29 (19.9 %), botulinum toxin injection in 13 (8.9 %) and Heller myotomy in seven patients (4.8 %). Eight adverse events occurred (6 mucosotomies, 2 pneumothorax) in eight patients (5.5 %). Median follow-up duration was 55 months (IQR 49.9-60.6). Clinical response was observed in 139 (95.2 %) patients at follow-up of ≥ 48 months. Symptomatic reflux after POEM was seen in 45 (32.1 %) patients, while 35.3 % of patients were using daily PPI at 48 months post POEM. Reflux esophagitis was noted in 16.8 % of patients who underwent endoscopy. Conclusion POEM is a durable and safe procedure with an acceptably low adverse event rate and an excellent long-term clinical response.

Room temperature water infusion during colonoscopy insertion induces rectosigmoid colon mucus production.

BACKGROUND: Water filling during colonoscopy improves several colonoscopy outcomes. We evaluated an anecdotal observation that room temperature water filling during colonoscope insertion results in mucus production in the left colon, which may impair mucosal visualization during withdrawal. METHODS: We performed 55 colonoscopies with either water or saline filling during insertion, and video recorded the examinations. Three blinded observers scored the amount of mucus visible on the video recordings. RESULTS: 29 patients had water filling and 26 patients had saline filling during insertion. Demographic features, procedure indications, volume of infused fluid, and insertion time to the cecum were similar in the two groups. All three blinded observers rated the mucus as greater after water filling than after saline (median 3 out of 5 vs. 1 out of 5; P < 0.001), with a kappa value for interobserver agreement of 0.364 (P < 0.001). CONCLUSION: Room temperature water filling is associated with mucus production by the rectosigmoid colon, requiring additional cleansing during withdrawal.

EUS-guided 22-gauge fine needle biopsy versus single-incision with needle knife for the diagnosis of upper gastrointestinal subepithelial lesions: a randomized controlled trial.

Background and study aims EUS-FNA has suboptimal accuracy in diagnosing gastrointestinal subepithelial tumors (SETs). EUS-guided 22-gauge fine needle biopsy (EUS-FNB) and single-incision with needle knife (SINK) were proposed to increase accuracy of diagnosis. This study aimed to prospectively compare the diagnostic accuracy and safety of EUS-FNB with SINK in patients with upper gastrointestinal SETs. Patients and methods All adult patients referred for EUS evaluation of upper gastrointestinal SETs ≥ 15 mm in size were eligible for inclusion. Patients were randomized to undergo EUS-FNB or SINK. Lesions were sampled with a 22-gauge reverse beveled core needle in the EUS-FNB group and by a conventional needle-knife sphincterotome and biopsy forceps in the SINK group. Patients were blinded to the technique used. The primary outcome was diagnostic accuracy. Secondary outcomes included adverse events, histological yield and procedure duration. Study enrollment was terminated early due to poor recruitment. Results A total of 56 patients (31 male (55.37 %); mean age, 67.41 ±â€Š12.70 years) were randomized to either EUS-FNB (n = 26) or SINK (n = 30). Technical success was 96.15 % and 96.66 %, respectively. The majority of lesions were gastrointestinal stromal tumors (51.78 %). No significant difference was found between EUS-FNB and SINK in terms of diagnostic accuracy for a malignant or benign disease (76 % vs. 89.28 %, respectively; P  = 0.278). The rate of adverse events (none severe) was also comparable (7.69 % vs. 10 %, respectively; P =  1.0) including two abdominal pain episodes in the EUS-FNB group compared to two delayed bleeding (one requiring hospitalization and radiologic embolization) and 1 abdominal pain in the SINK group. Conclusion EUS-FNB and SINK are equally effective techniques for upper gastrointestinal SETs sampling. SINK can be associated with mild to moderate delayed bleeding.