Expert Endoscopists Research Articles

Comparison of performances of artificial intelligence versus expert endoscopists for real-time assisted diagnosis of esophageal squamous cell carcinoma (with video)

Narrow-band imaging (NBI) is currently regarded as the standard modality for diagnosing esophageal squamous cell carcinoma (SCC). We developed a computerized image-analysis system for diagnosing esophageal SCC by NBI and estimated its performance with video images. Altogether, 23,746 images from 1544 pathologically proven superficial esophageal SCCs and 4587 images from 458 noncancerous and normal tissue were used to construct an artificial intelligence (AI) system. Five- to 9-second video clips from 144 patients captured by NBI or blue-light imaging were used as the validation dataset. These video images were diagnosed by the AI system and 13 board-certified specialists (experts). The diagnostic process was divided into 2 parts: detection (identify suspicious lesions) and characterization (differentiate cancer from noncancer). The sensitivities, specificities, and accuracies for the detection of SCC were, respectively, 91%, 51%, and 63% for the AI system and 79%, 72%, and 75% for the experts. The sensitivity of the AI system was significantly higher than that of the experts, but its specificity was significantly lower. Sensitivities, specificities, and accuracy for the characterization of SCC were, respectively, 86%, 89%, and 88% for the AI system and 74%, 76%, and 75% for the experts. The receiver operating characteristic curve showed that the AI system had significantly better diagnostic performance than the experts. Our AI system showed significantly higher sensitivity for detecting SCC and higher accuracy for characterizing SCC from noncancerous tissue than endoscopic experts.

Mo1637 NOVEL METHOD TO DETERMINE THE DEPTH OF INVASION IN COLORECTAL NEOPLASMS WITH ULTRASOUND ELASTOGRAPHY USING A FORWARD-VIEWING RADIAL-ARRAY ECHOENDOSCOPE

With the increasing use of endoscopy for treating colorectal cancer, accurate diagnosis of the depth of invasion in colorectal neoplasms has become an important clinical issue. Our study aimed to compare the diagnostic ability and clinical efficacy of endoscopic ultrasound elastography (EUS-EG) with those of magnifying chromoendoscopy (MCE) and EUS in diagnosing the depth of invasion in colorectal neoplasms. Patients with superficial colorectal neoplasms evaluated by MCE, EUS, and EUS-EG were enrolled. The primary clinical endpoint was the diagnostic yield differentiating intramucosal and shallow submucosal cancers from deep submucosal and advanced colorectal neoplasms, using the elastic score of colorectal neoplasms (ES-CRN). In addition, validation of ES-CRN was performed using inter- and intra-observer agreement from two expert and two non-expert endoscopists. We examined the diagnostic yield of additional patients prospectively due to these results. Thirty-one patients (33 lesions) with colorectal neoplasms were enrolled. The diagnostic yields (sensitivity, specificity, positive predictive value, negative predictive value and accuracy) of deeply submucosal cancers or advanced colorectal neoplasms, respectively, were as follows: EUS-EG:100%/88.2%/86.7%/100%/93.3%, MCE: 66.7%/94.4%/90.9%/77.3%/81.8% and EUS: 93.3%/77.8%/77.8%/93.3%/84.8%. Interobserver agreement for the ES-CRN was 0.70/0.80 and 0.45/0.40 for the first/second assessment, for the two expert endoscopists and two non-expert endoscopists. Furthermore, we examined additional 33 patients (34 lesions) prospectively. The diagnostic yields of prospective study were as follows, respectively: EUS-EG:83.3%/81.5%/50.0%/95.7%/81.8%, MCE: 75.0%/92.0%/75.0%/92.0%/87.9% and EUS: 85.7%/61.5%/37.5%/94.1%/66.7%. The diagnostic yield was from equal to higher for EUS-EG than for MCE and EUS, with moderate inter and intra-observer agreement, even for non-expert endoscopists. Therefore, EUS-EG may be a useful modality to determine the depth of invasion in colorectal neoplasms.

Tu1501 ADDITION OF LINKED COLOR IMAGING TO WHITE LIGHT ENDOSCOPY IMPROVES DELINEATION PERFORMANCE OF EARLY GASTRIC CANCER LESIONS BY NON-EXPERT ENDOSCOPISTS

Early gastric cancer lesions are subtle, focally distributed, and poorly visible endoscopically. Therefore, endoscopists have difficulties evaluating these subtle early gastric cancer lesions. The aim of this study was to evaluate the role of linked color imaging (LCI) for the visualization of early gastric cancer in comparison to white light endoscopy (WLE) alone, when assessed by non-expert endoscopists. 40 unique cases of early gastric cancer, visualized in both WLE and LCI, were collected for this study. To establish ground truth, all cases were delineated by three expert endoscopists from Japan. Using a web based module, these cases were delineated by 48 non-expert endoscopist assessors of four different countries (Japan, Netherlands, Portugal and Sweden). Endoscopic expertise was divided into different levels: fellow in training, junior endoscopist and senior endoscopist. The module consists of three assessment phases, with a wash-out period of 2 weeks in between each phase. Assessment 1: WLE alone; Assessment 2: LCI alone; Assessment 3: WLE+LCI in a side-to-side display. The outcomes of this study were: 1) overlap between assessors’ delineation and expert ground truth; 2) assessors’ delineation performance in terms of differentiating between neoplastic tissue and normal tissue; 3) ability to delineate the lesions (VAS-scores 1-10); 4) assessors’ preferred imaging modality (WLE, LCI or no preference). Linear mixed-effect models showed a significant increase in assessors’ delineation performance from 77% (SD 26%) in WLE alone and 75% (SD 28%) in LCI alone to 81% (SD 24%) when combining both imaging modalities (P < 0.001). When differentiating neoplastic from non-dysplastic tissue, the combination WLE+LCI also caused increased performance scores (49% in WLE and 49% in LCI to 53% in WLE+LCI; P < 0.001). Japanese assessors performed significantly better than Dutch, Portuguese and Swedish assessors when delineating early gastric cancer lesions, regardless of imaging modality (87% vs. 71%, 72% and 73% respectively; P < 0.001). There was no distinction between the different levels of endoscopic expertise. Median VAS scores were higher for phase 2 (6; IQR 5-8) and 3 (7; IQR 5-8) compared with phase 1 (5; IQR 3-7) for the ability to delineate the lesion (P < 0.001). Assessors preferred LCI over WLE (72.4% vs. 5.2%, 22.4% no preference) for the appreciation of early gastric cancer lesions. The combined use of WLE and LCI resulted in enhanced delineation of early gastric cancer lesions by non-expert endoscopists. Although assessors appreciated the ability of LCI for delineation better than WLE, there was no difference in separate delineation scores for LCI and WLE.

Su1548 THE IMPACT OF INCREASED BILE DUCT PRESSURE RELATED TO PAPILLARY STENOSIS AND THE ROLE OF VISCERAL HYPERSENSITIVITY IN PATIENTS WITH REFRACTORY RIGHT UPPER QUADRANT (RUQ) PAIN POST CHOLECYSTECTOMY

Functional abdominal pain is commonly related to visceral hypersensitivity (VH) and recent studies have shown success in management of such patients with Amitriptyline. A subgroup of patients with chronic RUQ pain post cholecystectomy routinely fail medical management and are referred for evaluation of Sphincter of Oddi (SOD) dysfunction/papillary stenosis. The association of RUQ pain, dilated common bile duct (CBD) and elevated liver function tests (LFT’s) is strongly suggestive of papillary stenosis as it has been shown to benefit the most from biliary sphincterotomy. The aim of this study was to determine the risks/benefits of endoscopic sphincterotomy in this subgroup of patients with refractory RUQ pain and to delineate the role of Amitriptyline in the presence of VH. Retrospective review, from 2015 to 2018, of treatment outcomes of patients with post cholecystectomy RUQ abdominal pain refractory to medical treatment. All patients had treatment with biliary sphincterotomy by single expert endoscopist. Patients with pancreaticobiliary malignancies or prior history of pancreatitis were excluded from the study. Data including patient demographics, clinical symptoms (pre and post ERCP), elevated LFT’s, CBD size, use of narcotics, prior history of IBS, post procedure complications (pancreatitis, bleeding) and prophylactic pancreatic duct stent placement was collected. Patients with no or partial improvement were treated with low dose (10-25 mg) Amitriptyline daily for presumed VH. 75 patients were included in the study. Mean age was 52 years with 68 (93%) females. 30 (40%) of patients had a dilated CBD. 14 (19%) had elevated LFT’s. 57 (76%) had constant RUQ pain. A total of 67 (89%) patients reported improvement in symptoms post sphincterotomy (49 (65%) patients achieved partial resolution and 18 (24%) complete resolution). A total of 13 (17%) patients did not achieve symptoms resolution and were treated with Amitriptyline. 12 out of 13 patients (92%) experienced a favorable symptom response to Amitriptyline. 6 (8%) patients had minor complications (pain and mild pancreatitis) and 1 patient (1%) had a major complication (necrotizing pancreatitis). A prophylactic pancreatic duct stent was placed in 38 (50.7%) of patients. 44 (58.6%) patients were discharge home post procedure. Biliary sphincterotomy provided by expert endoscopist is a very effective and safe intervention for patients with refractory RUQ abdominal pain post cholecystectomy. Constant RUQ pain and failure to respond to medical management should raise the suspicion for papillary stenosis. Based on the response to sphincterotomy, we extrapolate that elevated CBD pressure is the cause of discomfort in these patients with visceral hypersensitivity. Therefore, low dose Amitriptyline should be used for patients with partial response to biliary sphincterotomy.Graph 1: One-sample z-test of symptom improvement in included patientsView Large Image Figure ViewerDownload Hi-res image Download (PPT)

Mo1239 COMPARATIVE OUTCOMES OF CAP ASSISTED ENDOSCOPIC RESECTION AND ENDOSCOPIC SUBMUCOSAL DISSECTION IN PATIENTS WITH DYSPLASTIC BARRETT'S ESOPHAGUS AND INTRAMUCOSAL ADENOCARCINOMA

Endoscopic eradication therapy (EET) is the standard of care for the treatment of Barrett’s Esophagus (BE) with dysplasia and intramucosal esophageal adenocarcinoma (EAC). Cap assisted endoscopic resection (cEMR) leads to piecemeal resection in larger lesions. Endoscopic submucosal dissection (ESD) allows for the en bloc removal of larger lesions. Both approaches are followed by mucosal ablation to achieve complete eradication of dysplasia and intestinal metaplasia. Limited data is available for the efficacy of these two approaches followed by mucosal ablation. We studied the outcomes of patients treated with these modalities at our institution, focusing on rates of complete remission of intestinal metaplasia (CRIM) and dysplasia (CRD). We queried a prospectively maintained database of all patients at our institution undergoing EMR and ESD for dysplastic/neoplastic BE and abstracted relevant demographic and clinical data. Only patients with at least one endoscopic follow up were included. Our primary outcome was the achievement of CRIM, defined as the absence of intestinal metaplasia on at least one follow up EGD at 2 years. Secondary outcomes included rates of complete remission of dysplasia (CRD), defined as the absence of dysplastic BE on at least one follow up EGD and rates of complications. All pathology was read by expert gastrointestinal pathologists. Time dependent analyses were conducted to compare outcomes in the two groups. A total of 633 patients, of whom 447 that underwent cEMR and 89 that underwent ESD were included. 447 (100%) and 59 (66.3%) in the cEMR group and ESD groups underwent ablation (either radiofrequency ablation or cryotherapy) post endoscopic resection. Demographic and clinical variables are displayed in Table 1. The ESD group had a higher proportion of EAC patients (48.3% vs 20.1%, p<0.05) but proportions of HGD+EAC were similar in both groups. More complications were associated with the EMR group, driven mainly by post-procedural bleeding rates. Rates of CRD were lower in the EMR group at 6 months but were comparable at 12 and 24 months (see table 1 and Figure 1). Rates of CRIM were comparable between the two groups at 12 and 24 months. A multivariate model (adjusting for age, gender, BMI, smoking status, length of BE, presence of hiatal hernia, and initial histology) showed that EMR was associated with a lower likelihood of achieving CRD (HR 0.487, 95% CI: 0.337-0.703, p <0.0001), but not CRIM at 2 years (HR 0.895, 95% CI: 0.537-1.491, p=.67). EMR and ESD both appear to be effective in the treatment of BE related neoplasia. ESD is safe when performed by expert endoscopists. While there is a higher likelihood of achieving CRD earlier in patients with ESD, both resection techniques when combined with endoscopic ablation appear equally efficacious in achieving CRIM at 24 months.Comparison of baseline characteristics and outcomes (CRIM, CRD and complications) in the two groups.View Large Image Figure ViewerDownload Hi-res image Download (PPT)

Sa2057 COMPARISON OF ARTIFICIAL INTELLIGENCE AND EXPERT ENDOSCOPIST TOWARD REAL-TIME ASSISTED DIAGNOSIS OF ESOPHAGEAL SQUAMOUS CELL CARCINOMA.

Narrow-band imaging (NBI) is currently regarded as the standard modality for diagnosing esophageal squamous cell carcinoma (SCC). We developed a computerized image-analysis system for diagnosing esophageal SCC by NBI and estimated its performance with video images. Altogether, 7181 non-magnified (non-ME) and 7530 magnified (ME) endoscopic images of 1571 pathologically confirmed superficial esophageal SCCs and 564 non-ME and 2744 ME images from non-cancerous tissue or normal esophagus were used as training data. Five- to nine- second video clips from 144 patients were selected as validation data. These video images were diagnosed by the AI system and 13 board-certified specialists (experts). The diagnostic process was divided into two parts—the first was detection (identify suspicious lesions), and the other was characterization (differentiate cancer from non-cancer). The sensitivities, specificities, and accuracies for the detection of SCC were, respectively, 91%, 51%, and 63% for the AI system and 79%, 72%, and 75% for the experts. The sensitivity of the AI system was significantly higher than that of the experts, but its specificity was significantly lower. The sensitivities, specificities, and accuracy for the characterization of SCC were, respectively, 86%, 89%, and 88% for the AI system and 74%, 76%, and 75% for the experts. The receiver operating characteristics curve showed that the AI system had significantly better diagnostic performance than the experts. Our AI system showed significantly higher sensitivity for detecting SCC and higher accuracy for characterizing SCC from non-cancerous tissue than endoscopic experts.

Su1441 PRIMARY NEEDLE KNIFE SPHINCTEROTOMY FOR BILIARY ACCESS IN SELECTED PATIENTS WITH HIGH RISK FOR POST-ERCP PANCREATITIS: A PROSPECTIVE PRELIMINARY FEASIBILITY STUDY

Needle knife fistulotomy may be useful as the first line of procedure for biliary access to prevent post-endoscopic retrograde cholangiopancreatography (ERCP) pancreatitis (PEP) based on recent clinical data. We aimed to evaluate the safety, effectiveness, and feasibility of primary needle knife sphincterotomy (NKS) for biliary cannulation in patients with a high risk of PEP. A total of 40 patients who need ERCP using primary NKS for biliary disease in our institution between June 2018 and November 2019 were prospectively enrolled in this study. The procedures were conducted by two expert biliary endoscopists (S.J. and J.S.P.) who had experience of ERCPs. They had one or more of the following risk factors of PEP; young age (<50 years), female gender, normal common bile duct (CBD) diameter (≤9 mm), normal serum total bilirubin level (0.2-1.2 mg/dL), obesity (body mass index: >30), or history of pancreatitis. The success rate of biliary cannulation, time to biliary cannulation, and the occurrence rate of PEP, as well as other post-ERCP complications, were assessed. Of 40 patients enrolled, 34 patients underwent primary NKS. Nine patients had 1 risk factor of PEP, seven had 2, eight had 3, seven had 4, and three had 5. The success rate of CBD cannulation using NKS was 94.1% (32/34). The median time for biliary cannulation was 243 seconds (range, 32-1,514 seconds). None of the patients experienced post-ERCP pancreatitis, cholangitis, and perforation. One case of minor bleeding (2.9%, 1/34) and 1 asymptomatic hyperamylasemia (2.9%, 1/34) occurred. Primary NKS might be safe, effective and feasible for biliary access in patients with high risk for PEP if the procedure is performed by expert hands.

Complicated sinusitis in developing countries: an African perspective.

Complicated rhinosinusitis continues to be a diagnosis not to be missed because of the potential life-threatening and life-changing outcomes. This is true for clinicians in both developed and developing countries. Advances in diagnostic imaging, appropriate antibiotic use and timely surgical intervention have resulted in better outcomes particularly in the developed world. This review focuses on the current management of complicated sinusitis in resource-limited settings. Recent publications indicate leaning toward endoscopic sinus surgery as a standalone procedure or in combination with traditional external approaches in the management of complicated rhinosinusitis with less morbidity and shorter recovery time. It is becoming increasingly evident that multidisciplinary specialties should be involved in the effective management of patients with complicated sinusitis. Early recognition, prompt surgical intervention and antibiotic therapy remain keys in the management of complicated rhinosinusitis. In many African countries and other developing countries, the absence of specialized imaging and endoscopic diagnostic equipment and endoscopic surgical facilities and expertise means that reliance must be placed on clinical acumen both for diagnosis and monitoring of treatment response of complicated sinus infections.

Tu1040 VIDEO GAME EXPERIENCE IS POSITIVELY CORRELATED WITH FASTER IMPROVEMENT ON ENDOSCOPY HANDLING SKILLS

The development of strong hand-eye coordination and psychomotor skills are crucial in endoscopy. Given the limited hours and procedural opportunities for endoscopy trainees, it is important to recognize factors that may aid or enhance in the acquisition of psychomotor skills. A recent study showed that video game training improved trainee performance in advanced virtual reality endoscopic simulators. In addition, video gaming experience has been linked to higher baseline laparoscopic skills. We sought to explore the effects of video gaming experience on endoscopic skills. We hypothesize that video game experience is positively correlated with baseline endoscope handling skills. We recruited first year gastroenterology trainees to participate in a six-week simulator-based training program from October to November 2019. We provided all trainees a previously validated simulator2 to practice tip control and scope handling by hitting 28 target points throughout the simulator. We instructed trainees to complete their practice runs as quickly as possible (expert endoscopist completion rate from previous study=mean 75 seconds (95%CI, 60-100s) following specified scope handling criteria. We abstracted participant demographics (age, sex, hand length, and hand span) and whether they had significant experience in video gaming, sports, and musical instruments. Our primary outcome was the number of runs needed before being able to consistently complete a trial run under 100 seconds. Our secondary outcome was the minimum completion time or the fastest time it took a trainee to complete one trial run. We performed descriptive statistics. We received exempt status from the local IRB. We enrolled 28 first-year gastroenterology trainees. We excluded one trainee who did not meet competency or finish the program in the analysis. A detailed list of the participant demographics is shown in Table. The gaming group reached the outer limit of 95%CI of expert-level competency (100-second completion rate) at a significantly faster rate than the non-gaming group (19.6±19.6 trial runs vs 45.5±23.2 trial runs, p=0.0044) (Figure). On average, those in the gaming group had a faster minimum completion time for one trial run, although the difference was not statistically significant (48±15 seconds vs 52±12 seconds, p=0.21). There were no significant differences for the age, hand length, and hand span between the trainees with significant gaming experience (n=14) and trainees without significant gaming experience (n=13). Both groups had similar athletic and musical experience. Video game experience is positively correlated with the faster acquisition of scope handling skills. We show that a cohort of fellows with previous video-gaming experience are able to acquire competency at twice the rate of their peers without video-gaming experience.FigureOn average, the gaming group took 19.6+19.6 trial runs to achieve competency and the non-gaming group took 45.5+23.24 runs. The difference between the two is statistically significant (p=0.0044).View Large Image Figure ViewerDownload Hi-res image Download (PPT)

Mo1727 HOT AVULSION OF RESIDUAL VISIBLE NEOPLASIA IN LARGE LATERALLY SPREADING TUMORS RESULTS IN LOW RECURRENCE RATES WITHOUT THE NEED OF ABLATION OF THE MARGINS

EMR of large laterally spreading tumors (LSTs) results in recurrent or residual adenomas (RRAs) ranging from 15 to 30%.Attempts to reduce this rate include careful optical attention to residual visible neoplasia with additional resection with cold biopsy forceps, wide margin snare excision, or ablation with argon plasma coagulation (APC).Invisible residual neoplasia has been addressed with snare tip soft coagulation of the margins,extended marginal excision,and APC which results in decreased recurrences.We introduced hot avulsion as a means to excise parts of the polyp not amenable to snare resection due to adherence to the submucosa, difficulty in snare orientation, and inability to entrap neoplastic tissue. Hot avulsion steps includes:1)Using a hot biopsy forceps to grasp visible residual neoplastic tissue;2)Application of continuous traction effectively tenting the tissue away from the resection base (in order to minimize possible thermal injury);3)Delivering short bursts of a cutting current resulting in tissue shearing.This method is efficient and employs readily available inexpensive accessories, and addresses recurrences related to central areas of the polyp as well as marginal residual neoplasia. A retrospective review of our endoscopy database was performed(5/16-9/19).Inclusion criteria included: EMR of polyps >2cm,use of hot avulsion, and at least one follow up colonoscopy.All index procedures for polypectomy were performed by single expert endoscopist (GBH). LSTs considered endoscopically unresectable (with features of invasive carcinoma) were excluded. 254 LSTs were resected using EMR and avulsion, of which 118 LSTs in108 patients met follow up criteria to be included in the study.The average age was 64 years with 65 females(60%).Average LST size was36mm. Most polyps were resected in the ascending colon(33%).Histopathologic examination revealed 29% tubular adenoma (TA),29% sessile serrated adenoma (SSA), and20% tubulovillous adenoma (TVA). High-grade dysplasia (HGD) was found in17.7% of polyps and intramucosal adenocarcinoma in 2%. 9 patients had more than one qualifying polyp for resection with EMR and avulsion .Average time to first follow up colonoscopy was eight months (3-36 months). Optically visible RRAs were found in 5 patients(4%)with repeat resection histology revealing TA(3), TVA(1)and SSA(1). Initial histopathology of patients with RRA included 2 TVA with HGD,1 TVA,1 TA, and 1 TA with HGD.Post-polypectomy biopsy of scar was performed in 100 out of the 118 cases. Hot avulsion as an adjunct to endoscopic mucosal resection results in very low recurrence rates equivalent to the lowest rates achieved with other more laborious adjunctive therapies. Moreover the strength of this study is the high rate of histologic interrogation of the scar(85%) not performed in most studies of recurrence.Adverse Events, Follow-up Colonoscopies and Recurrence Lesion CharacteristicsView Large Image Figure ViewerDownload Hi-res image Download (PPT)

Tu1041 RANDOMIZED CONTROLLED TRIAL OF NEEDLE-KNIFE FISTULOTOMY VERSUS TRANSPAPILLARY CANNULATION FOR PRIMARY BILIARY ACCESS DURING ENDOSCOPIC RETROGRADE CHOLANGIO-PANCREATOGRAPHY PERFORMED BY ADVANCED ENDOSCOPY TRAINEES.

Needle-knife fistulotomy (NKF) is a technique traditionally used to achieve biliary access when transpapillary cannulation techniques are unsuccessful. The literature supports that NKF can be safely performed as a primary biliary access approach but due to technical aspects, this technique should be reserved for expert endoscopists. The aim of this study was to evaluate the efficacy and safety of NKF compared to standard transpapillary access when performed by trainee endoscopists. We conducted a randomized controlled prospective trial of endoscopic retrograde cholangio-pancreatography (ERCP) performed by advanced endoscopy trainees at a tertiary hospital (Thomas Jefferson University Hospital, Philadelphia, PA). Endoscopists included two advanced fellows. Consecutive patients with a native papilla requiring biliary access and agreeing to participate were enrolled in the study and randomized to one of two groups. One group underwent standard cannulation through the papillary orifice followed by standard sphincterotomy. The other group underwent a supra-papillary needle knife fistulotomy. For patient safety, patients found to have a papilla within or on the rim of a diverticulum were excluded. The assigned access technique was considered unsuccessful if access attempts exceeded 20 minutes. End points included: success of biliary access, time to biliary access, and complications with all patients called 48-72 hours after their procedures. A total of 85 patients were enrolled. There were 21 screen failures and 7 patients excluded on the basis of a periampullary diverticulum, leaving 57 study patients (30 transpapillary techniques and 27 NKF) performed by trainees. Mean age was 59.2 in the transpapillary group and 61.1 in the NKF group (p=0.704) with 40% and 41% males in transpapillary and NKF groups, respectively (p=0.955). Indications for transpapillary and NKF access were choledocholithiasis (56.7% vs 51.9%), malignancy (40% vs 40.7%), and other (3.3% vs 7.4%). There was no difference in successful biliary access under 20 minutes between transpapillary and NKF groups (73.3% vs 88.9%, p=0.137). Mean time to biliary access for transpapillary and NKF groups was 4.09 vs 7.48 minutes (p=0.02), respectively. There was also no difference in intraprocedural and postprocedural complications, including bleeding (0% vs 0%), perforation (4.6% vs 4.2%, p=0.95), and pancreatitis (4.6% vs 8.3%, p=0.603). Our study suggests that advanced endoscopy trainees can safely perform NKF for primary biliary access. Successful biliary access, time to access, and complication rates did not significantly differ between the two groups. NKF can be safely performed by trainee endoscopists and should be considered a viable approach for initial biliary access.

Expert endoscopists with high adenoma detection rates frequently detect diminutive adenomas in proximal colon.

Background and study aims Adenoma detection rate (ADR) is an important quality indicator in colonoscopy, and improved ADR decreases the incidence of colorectal cancer. We investigated differences in polyp detection according to the endoscopist’s ADR. Patients and methods We performed a propensity-score matching study using baseline patient characteristics of age, sex, body mass index, family history of colorectal cancer, smoking, drinking, indication for colonoscopy, bowel preparation, and colonoscope type. We compared polyp detection and colonoscopy procedures between patients who underwent colonoscopy by high-ADR endoscopists (high ADR group) and by low-ADR endoscopists (low ADR group). Results We matched 334 patients in the high ADR group with 334 in the low ADR group. The ADR was 44.0 % and 26.9 % for the high-ADR and low-ADR endoscopists, respectively. Proximal, nonprotruding, and diminutive adenomas were more frequently detected by high-ADR endoscopists than by low-ADR endoscopists (all P < 0.001); similarly, more high-risk adenomas were detected by high-ADR endoscopists ( P = 0.028). Furthermore, more sessile serrated polyps detected by high-ADR endoscopists ( P = 0.041). High-ADR endoscopists more frequently performed pancolonic chromoendoscopy ( P < 0.001). Conclusions Expert detectors often found nonprotruding and diminutive adenomas in the proximal colon along with increased detection rate of high-risk adenomas. Low-ADR endoscopists need to recognize the features of missed adenomas to improve their ADRs.

Triaging advanced GI endoscopy procedures during the COVID-19 pandemic: consensus recommendations using the Delphi method

Background and AimsThere is a lack of consensus on which GI endoscopic procedures should be performed during the COVID-19 pandemic, and which procedures could be safely deferred without having a significant impact on outcomes.MethodsWe selected a panel of 14 expert endoscopists. We identified 41 common indications for advanced endoscopic procedures from the ASGE Appropriate Use of GI Endoscopy guidelines. Using a modified Delphi method, we first achieved consensus on the patient-important outcome for each procedural indication. Panelists prioritized consensus patient-important outcome when categorizing each indication into one of the following 3 procedural time periods: (1) time-sensitive emergent (schedule within 1 week), (2) time-sensitive urgent (schedule within 1 to 8 weeks), and (3) non-time sensitive (defer for >8 weeks and then reassess the timing). Three anonymous rounds of voting were allowed before attempts at consensus were abandoned.ResultsAll 14 invited experts agreed to participate in the study. The prespecified consensus threshold of 51% was achieved for assigning patient-important outcome(s) to each advanced endoscopy indication. The prespecified consensus threshold of 66.7% was achieved for 40 of 41 advanced endoscopy indications in stratifying them into 1 of 3 procedural time periods. For 12 of 41 indications, 100% consensus was achieved; for 20 of 41 indications, 75% to 99% consensus was achieved.ConclusionsBy using a Modified Delphi method that prioritized patient-important outcomes, we developed consensus recommendations on procedural timing for common indications for advanced endoscopy. These recommendations and the structured decision framework provided by our study can inform decision making as endoscopy services are reopened.

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

The performance of magnifying endoscopy with narrow-band imaging (ME-NBI) using a computer-aided diagnosis (CAD) system in diagnosing early gastric cancer (EGC) is unclear. Here, we aimed to clarify the differences in the diagnostic performance between expert endoscopists and the CAD system using ME-NBI. The CAD system was pretrained using 1492 cancerous and 1078 noncancerous images obtained using ME-NBI. One hundred seventy-four videos (87 cancerous and 87 noncancerous videos) were used to evaluate the diagnostic performance of the CAD system using the area under the curve (AUC), accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). For each item, comparisons were made between the CAD system and 11 experts who were skilled in diagnosing EGC using ME-NBI with clinical experience of more than 1 year at our hospital. The CAD system demonstrated an AUC of 0.8684. The accuracy, sensitivity, specificity, PPV, and NPV were 85.1% (95% confidence interval [95% CI], 79.0-89.6), 87.4% (95% CI, 78.8-92.8), 82.8% (95% CI, 73.5-89.3), 83.5% (95% CI, 74.6-89.7), and 86.7% (95% CI, 77.8-92.4), respectively. The CAD system was significantly more accurate than 2 experts, significantly less accurate than 1 expert, and not significantly different from the remaining 8 experts. The overall performance of the CAD system using ME-NBI videos in diagnosing EGC was considered good and was equivalent to or better than that of several experts. The CAD system may prove useful in the diagnosis of EGC in clinical practice.

Current status and trend in training for endoscopic submucosal dissection: A nationwide survey in Korea.

BackgroundAlthough endoscopic submucosal dissection (ESD) is widely used, the current status and trend in its training have yet to be fully evaluated. We aimed to investigate how ESD endoscopists have been trained in actual clinical practice.MethodsEndoscopists aged <45 years who have completed a gastroenterology fellowship or were currently in a fellowship for ≥2 years were included. We conducted a nationwide survey on the ESD training experiences of these endoscopists.ResultsAmong 79 young Korean endoscopists invited to participate in the survey, 68 (86.1%) trained in 24 major hospitals responded to the questionnaire. Twenty, 25, and 23 participants belonged to the second-year fellow, <5 years after training, and ≥5 years after training groups, respectively. Sixty-nine percent of the participants observed ≥50 ESD cases before starting ESD under supervision by an expert endoscopist. Additionally, 22% experienced ≥20 supervised ESDs during the training period. The proportion of the participants who underwent a hands-on course differed among the groups (≥5 years after training, 13.0%; <5 years after training, 40.0%; and second-year fellow, 50.0%; P = 0.027). ESD under supervision, observation, and hands-on course were the preferred methods for learning ESD (91.1%, 80.9%, and 35.3%, respectively). Overall, 42.6% of the participants were satisfied with their training program. More experience in supervised ESD (≥20 cases) was associated with an increased satisfaction (odds ratio, 6.65; 95% confidence interval, 1.62–36.31).ConclusionsObservation and performance of ESD under the supervision of an expert endoscopist are the primary methods for learning ESD. Hands-on course program has been used more frequently in recent years.

How To Maximize Trainee Education During the Coronavirus Disease-2019 Pandemic: Perspectives From Around the World

How To Maximize Trainee Education During the Coronavirus Disease-2019 Pandemic: Perspectives From Around the World

Inter-Observer Agreement of a New Endoscopic Score for Ulcerative Colitis Activity: Preliminary Experience.

Ulcerative colitis (UC) endoscopic scores translate mucosal damage into values standardizing image analysis. Due to potential limits of current endoscopic activity indexes, we have elaborated on a new score, the “Extended Mayo Endoscopic Score (EMES),” and evaluated its inter-observer agreement in a multicenter endoscopy team, comparing concordance with the Mayo subscore. Sixteen UC consecutive patients underwent follow-up colonoscopy. Recorded videos were anonymously loaded on a web platform. Thirteen expert endoscopists evaluated UC activity using both Mayo and EMES. EMES was described in every colon segment: erythema (0: absent, 1: mild, 2: moderate, 3: severe), vascular pattern (0: normal, 1: reduction, 2: disappearance), erosions and ulcers (0: absent, 1: from 1 to 5, 2: 6 to 10, 3: >10). Weighted Fleiss’ kappa with 95% confidence interval (CI) and p-value defined inter-rater agreement. Global inter-observer agreement of EMES was moderate (kappa = 0.56, 95% CI = 0.46–0.67, p < 0.001). The evaluation of each colonic segment showed moderate agreement for all segments: ascending (kappa = 0.46, 95% CI = 0.32–0.60, p < 0.001), transverse (kappa = 0.48, 95% CI = 0.29–0.67, p < 0.001); descending (kappa = 0.49, 95% CI = 0.35–0.64, p < 0.001), sigmoid (kappa = 0.52, 95% CI = 0.39–0.65, p < 0.001) and rectum (kappa = 0.55, 95% CI = 0.42–0.69, p < 0.001). Mayo subscore agreement was similar to global EMES (kappa = 0.53, 95% CI = 0.39–0.66, p = 0.001). Therefore, our report emphasizes the importance of assessing inter-observer agreement for EMES, but also for other known scoring systems, including the Mayo subscore.

A167 MANAGEMENT OF LARGE COLORECTAL POLYPS IN NOVA SCOTIA

Abstract Background Colorectal cancer (CRC) has the third highest cancer incidence for males and females in Canada1. CRC rates are decreasing, likely due to implementation of CRC screening programs and removal of precancerous polyps. In Nova Scotia, management of large (&amp;gt; 2 cm) polyps is varied. Some patients are referred for endoscopic mucosal resection (EMR), while others referred for surgical resection. EMR is effective2 and associated with less morbidity and mortality compared to surgery3. Furthermore, EMR is less expensive than surgery and requires less resources when performed as a day procedure. Management of these polyps are unknown in Canada and are potentially changing with gathering evidence on the safety and effectiveness of EMR. North American and European guidelines recommend EMR for non-invasive colonic polyps as first line management. Aims This projects aim is to determine how large polyps are being treated in a group of FIT positive patients from Nova Scotia. Methods This is a retrospective observational study. The Nova Scotia Colon Cancer screening database was searched for FIT positive patients in 2017. Patients found to have &amp;gt; 2 cm colonic polyps were included. Malignant polyps were excluded. Electronic charts were reviewed to determine method of polyp removal. Endoscopic and histologic polyp information was obtained. Results 196 patients had at least one polyp &amp;gt; 2 cm. Within the central health zone where pathology results were available 78 patients had &amp;gt; 2cm polyps. 13 were excluded for malignancy, one lost to follow up and one patient was miscoded. In total 63 patients were included, 8 of which underwent a surgical resection. Of these surgically resected polyps the average size was 4 cm and 3 cm for the EMR group. The average time to removal was 3 months in the surgical group and one month in the EMR group. The average age was 68 years in the surgical group and 67 years in the EMR group. The endoscopists referring for surgery were 75% surgeon and 25% GI and the endoscopists performing EMR were 45% surgeon and 55% GI. 1 of 8 (12.5%) patients experienced complications (anastomotic leak) in the surgical group and 2 of 56 (3.6%) patients experienced post polypectomy bleeding in the EMR group. Conclusions In this study 12.5% of large nonmalignant polyps found in a group of FIT positive patients from Nova Scotia were removed surgically via resection. Polyps were on average larger by 1 cm in the surgical group and patients experienced a higher rate of complication. This study highlights the need for a structured system of referral and assessment for endoscopic removal of advanced polyps by expert endoscopists, to reduce the number of surgical resections. Additional investigation of additional years, and health zones outside the tertiary central health zone is ongoing. This will help provide more validity to our current data, and allow comparison of academic vs community management of large polyps. Funding Agencies None

The Italian Society for Digestive Endoscopy (SIED) accreditation and quality improving project based on international standards.

Background and study aims Accreditation of endoscopy services, using valid quality indicators, may address failures to comply with quality standards between endoscopy services. The aim of this work was to present the Italian Society for Digestive Endoscopy (SIED) accreditation model and its effectiveness. Methods A team of eight endoscopists identified quality indicators derived from international guidelines and assessed them in each center voluntarily requesting accreditation. During a 1-day site visit, two expert endoscopists, the representative of the independent and international administrative certification body and a professional nurse evaluated the endoscopy center, by direct observation of the endoscopy team and examination of the medical records Results In all centers we noted shortcomings in instrument reprocessing. In 30 of 40 centers (75 %) the information in the nursing charts was incomplete. Sampling for Helicobacter pylori had not been done in 12 of 40 centers (30 %). In six of 40 centers (15 %) the adenoma detection rate for each endoscopist had not been evaluated. Post-polypectomy intervals were inappropriate in 12 of 40 centers (30 %). We noted a statistically significant difference ( P < 0.001) between the answers to the SIED checklist of indicators submitted to the inspection team for accreditation before the site visit and the situation found for colonoscopy on site. As of June 30, 2018, 18 endoscopy centers had been accredited and 10 centers had not yet being accredited because they had not completed the measures to correct points raised at the visits. Conclusions Numerous Italian endoscopy centers fail to meet important quality indicators. Our accreditation program can provide means for detecting these problems and correcting them by implementing SIED standards.

Digital Single-operator Cholangioscopy (DSOC) Improves Interobserver Agreement (IOA) and Accuracy for Evaluation of Indeterminate Biliary Strictures

Visual characteristics seen during digital single-operator cholangioscopy (DSOC) have not been validated. The aim of this 2-phase study was to define terminology by consensus for the visual diagnosis of biliary lesions to develop a model for optimization of the diagnostic performance of DSOC. In phase 1 (criteria identification), video-cholangioscopy clips were reviewed by 12 expert biliary endoscopists, who were blinded to the final diagnosis. Visual criteria were consolidated into the following categories: (1) stricture, (2) lesion, (3) mucosal features, (4) papillary projections, (5) ulceration, (6) abnormal vessels, (7) scarring, (8) pronounced pit pattern.During the second phase (validation), 14 expert endoscopists reviewed DSOC (SpyGlass DS, Boston Scientific) clips using the 8 criteria to assess interobserver agreement (IOA) rate. In phase 1, consensus for visual findings were categorized into 8 criteria titled the "Monaco Classification." The frequency of criteria were: (1) presence of stricture-75%, (2) presence of lesion type-55%, (3) mucosal features-55%, (4) papillary projections-45%, (5) ulceration-42.5%, (6) abnormal vessels-10%, (7) scarring-40%, and (8) pronounced pit pattern-10%. The accuracy on final diagnosis based on visual impression alone was 70%.In phase 2, the IOA rate using Monaco Classification criteria ranged from slight to fair. The presumptive diagnosis IOA was fair (κ=0.31, SE=0.02), and overall diagnostic accuracy was 70%. The Monaco classification identifies 8 visual criteria for biliary lesions on single-operator digital cholangioscopy. Using the criteria, the IOA and diagnostic accuracy rate of DSOC is improved compared with prior studies.