A deep learning system for detection of early Barrett's neoplasia: a model development and validation study.

BACKGROUND: Computer-aided detection (CADe) systems could assist endoscopists in detecting early neoplasia in Barrett's oesophagus, which could be difficult to detect in endoscopic images. The aim of this study was to develop, test, and benchmark a CADe system for early neoplasia in Barrett's oesophagus. METHODS: The CADe system was first pretrained with ImageNet followed by domain-specific pretraining with GastroNet. We trained the CADe system on a dataset of 14 046 images (2506 patients) of confirmed Barrett's oesophagus neoplasia and non-dysplastic Barrett's oesophagus from 15 centres. Neoplasia was delineated by 14 Barrett's oesophagus experts for all datasets. We tested the performance of the CADe system on two independent test sets. The all-comers test set comprised 327 (73 patients) non-dysplastic Barrett's oesophagus images, 82 (46 patients) neoplastic images, 180 (66 of the same patients) non-dysplastic Barrett's oesophagus videos, and 71 (45 of the same patients) neoplastic videos. The benchmarking test set comprised 100 (50 patients) neoplastic images, 300 (125 patients) non-dysplastic images, 47 (47 of the same patients) neoplastic videos, and 141 (82 of the same patients) non-dysplastic videos, and was enriched with subtle neoplasia cases. The benchmarking test set was evaluated by 112 endoscopists from six countries (first without CADe and, after 6 weeks, with CADe) and by 28 external international Barrett's oesophagus experts. The primary outcome was the sensitivity of Barrett's neoplasia detection by general endoscopists without CADe assistance versus with CADe assistance on the benchmarking test set. We compared sensitivity using a mixed-effects logistic regression model with conditional odds ratios (ORs; likelihood profile 95% CIs). FINDINGS: Sensitivity for neoplasia detection among endoscopists increased from 74% to 88% with CADe assistance (OR 2·04; 95% CI 1·73-2·42; p<0·0001 for images and from 67% to 79% [2·35; 1·90-2·94; p<0·0001] for video) without compromising specificity (from 89% to 90% [1·07; 0·96-1·19; p=0·20] for images and from 96% to 94% [0·94; 0·79-1·11; ] for video; p=0·46). In the all-comers test set, CADe detected neoplastic lesions in 95% (88-98) of images and 97% (90-99) of videos. In the benchmarking test set, the CADe system was superior to endoscopists in detecting neoplasia (90% vs 74% [OR 3·75; 95% CI 1·93-8·05; p=0·0002] for images and 91% vs 67% [11·68; 3·85-47·53; p<0·0001] for video) and non-inferior to Barrett's oesophagus experts (90% vs 87% [OR 1·74; 95% CI 0·83-3·65] for images and 91% vs 86% [2·94; 0·99-11·40] for video). INTERPRETATION: CADe outperformed endoscopists in detecting Barrett's oesophagus neoplasia and, when used as an assistive tool, it improved their detection rate. CADe detected virtually all neoplasia in a test set of consecutive cases. FUNDING: Olympus.

Optimal timing of simethicone administration prior to upper endoscopy: A multicenter, single-blind, randomized controlled trial.

Background and study aims Simethicone is useful as premedication for upper endoscopy because of its antifoaming effects. We aimed to evaluate the effect of timing of simethicone administration on mucosal visibility. Patients and methods In this multicenter, randomized, endoscopist-blinded study, patients scheduled for upper endoscopy were randomized to receive 40 mg simethicone at the following time points prior to the procedure: 20 to 30 minutes (early group), 0 to 10 minutes (late group) or 20 mg simethicone at both time points (split-dose group). Images were taken from nine predefined locations in the esophagus, stomach, and duodenum before endoscopic flushing. Each image was scored on mucosal visibility by three independent endoscopists on a 4-point scale (lower scores indicating better visibility), with adequate mucosal visibility defined as a score ≤ 2. Primary outcome was the percentage of patients with adequate total mucosal visibility (TMV), reached if all median subscores for each location were ≤ 2. Results A total of 386 patients were included (early group: 132; late group: 128; split-dose group: 126). Percentages of adequate TMV were 55%, 42%, and 61% in the early, late, and split-dose group, respectively ( P < 0.01). Adequate TMV was significantly higher in the split-dose group compared to the late group ( P < 0.01), but not compared to the early group ( P = 0.29). Differences between groups were largest in the stomach, where percentages of adequate mucosal visibility were higher in the early (68% vs 53%, P = 0.03) and split-dose group (69% vs 53%, P = 0.02) compared to the late group. Conclusions Mucosal visibility can be optimized with early simethicone administration, either as a single administration or in a split-dose regimen.

Efficacy and safety of a novel submucosal lifting gel used for endoscopic submucosal dissection: a study in a porcine model.

BACKGROUND AND STUDY AIMS: Endoscopic submucosal dissection (ESD) is technically demanding. A viscous gel for submucosal lifting might induce mechanical submucosal dissection facilitating easier and safer ESD. METHODS: In 12 female pigs (median 64 kg), ESDs of simulated lesions were performed at the posterior wall and greater curvature in the gastric body (one ESD per location) with randomly assigned injection fluids: gel or control fluid (0.9% saline with hydroxypropyl methylcellulose 3 mg/ml [7:1] and indigo carmine droplets). Additionally, 10 cc gel was injected into the submucosa at the anterior wall without ESD to assess effects of inappropriate injection. Pigs were euthanized at day 0, 3 or 28. In four additional pigs (euthanized day 3 or 28) 10 cc gel was injected into the muscularis propria (MP) after four endoscopic mucosal resections in the gastric body. RESULTS: Both fluid groups showed equal ESD-procedure times (28 [gel] vs. 26 min [control]) and complications. Gel-ESDs required less accessory interchanges (3.5 vs. 5.5; p = 0.01). Mechanical dissection after circumferential incision was achieved in 25% of gel-ESDs; none in control-ESDs. The severity of inflammation and fibrosis was equal in both fluid groups. Normal architecture and vital mucosa were found after inappropriate submucosal injection. MP-injections resulted in one transmural hematoma (day 3), and intramuscular encapsulation in 25% of the sites (day 28). LIMITATIONS: A pig's stomach differs from the human stomach. CONCLUSIONS: The mechanical dissection properties of the gel may reduce the need for submucosal dissection during ESD. The gel is safe when advertently injected in the submucosa and MP. The porcine model appeared suboptimal to evaluate the true mechanical dissection properties of the gel.

Effects of Lugol staining on stenosis formation induced by radiofrequency ablation of esophageal squamous epithelium: a study in a porcine model.

Preliminary data show higher stricture rates after radiofrequency ablation (RFA) for early esophageal squamous neoplasia compared with Barrett's esophagus. We studied the effects of Lugol stain (LS) directly prior to RFA on stricture formation in squamous epithelium. Of 16 pigs, the distal half of the esophagus was LS, followed by circumferential RFA (single application 12 J/cm(2) ) in the unstained and stained esophagus. Pigs were euthanized at day 0 (n = 4), 3 (n = 4), or 28 (n = 8). Histology was evaluated in four areas: blank-control (no RFA, no LS), blank-RFA (no LS), LS+RFA, and LS-control (no RFA). Stenosis severity in LS+RFA and blank-RFA at 28 days was assessed by the ratio of the mucosal diameter at the RFA area to the diameter 2 cm proximal of this zone. Histology showed submucosal edema in 50% of LS+RFA versus 0% in blank-RFA. Severity and depth of inflammation (day 3) was equal in LS+RFA and blank-RFA. Severity and depth of fibrosis (day 28) appeared more severe in LS+RFA. Consequently, stenosis was present in 100% (LS+RFA) versus 12.5% (blank-RFA). The stenosis-severity ratio was 0.40 (interquartile range 0.29-0.45) in LS+RFA versus 0.73 (interquartile range 0.64-0.78) in blank-RFA (P = 0.012). Limitations of this study were the difference in uptake of LS between pigs and humans, the difference in esophageal anatomy between pigs and humans, and between the proximal and distal esophagus within pigs. In conclusion, in the porcine squamous esophagus, stenosis rate and severity after RFA increased when preceded by LS. LS may be contributing in the altered response of squamous epithelium to RFA as compared with Barrett's esophagus.

Predictive factors for initial treatment response after circumferential radiofrequency ablation for Barrett's esophagus with early neoplasia: a prospective multicenter study.

BACKGROUND AND STUDY AIMS: Radiofrequency ablation (RFA) is safe and effective for the eradication of neoplastic Barrett's esophagus; however, occasionally there is minimal regression after initial circumferential balloon-based RFA (c-RFA). This study aimed to identify predictive factors for a poor response 3 months after c-RFA, and to relate the percentage regression at 3 months to the final treatment outcome. METHODS: We included consecutive patients from 14 centers who underwent c-RFA for high grade dysplasia at worst. Patient and treatment characteristics were registered prospectively. "Poor initial response" was defined as < 50 % regression of the Barrett's esophagus 3 months after c-RFA, graded by two expert endoscopists using endoscopic images. Predictors of initial response were identified through logistic regression analysis. RESULTS: There were 278 patients included (median Barrett's segment C4M6). In poor initial responders (n = 36; 13 %), complete response for neoplasia (CR-neoplasia) was ultimately achieved in 86 % (vs. 98 % in good responders; P < 0.01) and complete response for intestinal metaplasia (CR-IM) in 66 % (vs. 95 %; P < 0.01). Poor responders required 13 months treatment (vs. 7 months; P < 0.01) for a median of four RFA sessions (vs. three; P < 0.01). We identified four independent baseline predictors of poor response: active reflux esophagitis (odds ratio [OR] 37.4; 95 % confidence interval [CI] 3.2 - 433.2); endoscopic resection scar regeneration with Barrett's epithelium (OR 4.7; 95 %CI 1.1 - 20.0); esophageal narrowing pre-RFA (OR 3.9; 95 %CI 1.0 - 15.1); and years of neoplasia pre-RFA (OR 1.2; 95 %CI 1.0 - 1.4). CONCLUSIONS: Patients with a poor initial response to c-RFA have a lower ultimate success rate for CR-neoplasia/CR-IM, require more treatment sessions, and a longer treatment period. A poor initial response to c-RFA occurs more frequently in patients who regenerate their endoscopic resection scar with Barrett's epithelium, and those with ongoing reflux esophagitis, neoplasia in Barrett's esophagus for a longer time, or a narrow esophagus.

Radiofrequency ablation and endoscopic resection in a single session for Barrett's esophagus containing early neoplasia: a feasibility study.

BACKGROUND AND STUDY AIM: Endoscopic resection with radiofrequency ablation (RFA) 6 weeks later safely and effectively eradicates Barrett's esophagus with high grade dysplasia (HGD) and early cancer. After widespread endoscopic resection, related scarring may hamper balloon-based circumferential RFA (c-RFA). However c-RFA immediately followed by endoscopic resection in the same session might avoid the impact of scarring and reduce laceration and stenosis risk. We aimed to assess the feasibility of such an approach. PATIENTS AND METHODS: Patients with Barrett's esophagus ≥ 3 cm and ≥ 1 visible lesion (HGD/early cancer) were included. Visible lesions were marked with cautery, and c-RFA (12 J/cm2) was delivered using two applications and a cleaning step, followed by resection of the delineated area. Outcome measures were surface regression of Barrett's esophagus at 3 months, need for subsequent c-RFA, complications, and quality of resection specimens. RESULTS: 24 patients (20 men, 4 women; mean age 68 years, standard deviation [SD] 12; Barrett's esophagus median length C6M8) underwent single-session c-RFA + endoscopic resection, providing a median of 4 (interquartile range [IQR] 2 - 6) resection specimens (early cancer 18 patients; HGD 6). Complications included 1 perforation, 4 bleedings, and 5 stenoses; all were managed endoscopically. Specimens allowed assessment of neoplasia depth, differentiation, and lymphatic/vascular invasion. Median Barrett's esophagus surface regression at 3 months was 95 %. No patient required a second c-RFA procedure and 40 % required repeat endoscopic resection for visible lesions. Complete response for neoplasia was achieved in 100 % and complete response for intestinal metaplasia (CR-IM) in 95 %. CONCLUSIONS: c-RFA followed by endoscopic resection in the same session is feasible, but technically demanding and associated with a substantial rate of complications and repeat endoscopic resection. This approach should be reserved for selected cases in expert centers, with endoscopic resection and RFA 6 - 8 weeks later remaining the standard combined approach.

Observer agreement in the assessment of narrowband imaging system surface patterns in Barrett's esophagus: a multicenter study.

BACKGROUND AND STUDY AIMS: The clinical utility of narrow-band imaging (NBI) for Barrett's esophagus is limited by the multiplicity of classification schemes. We evaluated the interobserver agreement and accuracy of a new consensus-driven simplified binary classification of NBI surface patterns.

Radiofrequency ablation for the endoscopic eradication of esophageal squamous high grade intraepithelial neoplasia and mucosal squamous cell carcinoma.

BACKGROUND AND STUDY AIMS: Radiofrequency ablation (RFA) with or without prior endoscopic resection safely and effectively removes early neoplasia in Barrett's esophagus. We speculated that this approach might also be suited for early squamous neoplasia of the esophagus. The aim of the study was to assess our initial experiences with RFA for high grade intraepithelial neoplasia (HGIN) and esophageal squamous cell cancer (ESCC) limited to the mucosa. PATIENTS AND METHODS: This was a prospective case series study in two tertiary centers. Patients with at least one unstained lesion (USL) of the esophagus using Lugol's chromoendoscopy and squamous HGIN/ESCC upon biopsy were included. In the case of nonflat USLs, endoscopic resection was performed for staging and to render the mucosa flat. After endoscopic resection and subsequent circumferential RFA, chromoendoscopy was repeated every 3 months with focal RFA of residual USLs. Follow-up chromoendoscopy was repeated at 6 months and annually thereafter. The main outcome measure was complete histological response for any squamous intraepithelial neoplasia or ESCC. RESULTS: A total of 13 patients (10 HGIN, three ESCC) were included. Following endoscopic resection in nine patients, the median extent of USLs was 4 cm and 50 % of circumference. All 13 patients achieved a complete response after a median of 2 RFA sessions (IQR 1 - 3 sessions). RFA-related complications included two mucosal lacerations (at the endoscopic resection scar) and one intramural hematoma, none requiring therapy. Endoscopic resection-/RFA-related complications were three stenoses. Dilation resulted in perforation in one patient (managed with a covered stent). There were no recurrences (median follow-up 17 months [IQR 11 - 22 months]). CONCLUSIONS: This study suggests that RFA with or without prior endoscopic resection for esophageal squamous HGIN and mucosal ESCC is feasible and effective.

Safety and efficacy of multiband mucosectomy in 1060 resections in Barrett's esophagus.

INTRODUCTION: Multiband mucosectomy (MBM) is a relatively new technique for endoscopic resection in Barrett's esophagus. This ligate-cut technique uses a modified variceal band ligator allowing for six consecutive resections without prior submucosal lifting. The aim was to evaluate the safety of MBM and its efficacy for complete endoscopic removal of delineated target areas in Barrett's esophagus. METHODS: Prospective registration of all MBM procedures in Barrett's esophagus was carried out between November 2004 and October 2009 in two hospitals. Prior to MBM, the target area was delineated with coagulation markings, followed by endoscopic resection until the delineated area was completely resected. Primary end points were acute (during procedure) plus early complications (<30 days) and the rate of complete endoscopic resection of the delineated target area. RESULTS: A total of 243 MBM procedures, with 1060 resections, were performed in 170 patients. MBM was performed for focal lesions (n=113), for Barrett's esophagus removal as part of a (stepwise) radical endoscopic resection protocol (n=117), and as escape treatment after radiofrequency ablation (n=13). The only acute complication was bleeding (in 3%, endoscopically managed); no perforations occurred despite absence of submucosal lifting. Early complications consisted of delayed bleeding (in 2%, endoscopically managed) and stenosis, which occurred in 48% of patients treated in a (stepwise) radical resection protocol; patients treated for focal lesions or in escape treatment showed no stenosis. Complete endoscopic resection was achieved in 91% of the focal lesions, in 86% of cases treated under the (stepwise) radical endoscopic resection protocol, and 100% for escape treatment after radiofrequency ablation. CONCLUSION: MBM is a safe and effective technique for the removal of delineated target areas in Barrett's esophagus.

Risk of lymph node metastasis associated with deeper invasion by early adenocarcinoma of the esophagus and cardia: study based on endoscopic resection specimens.

BACKGROUND: Most risk estimations for lymph node metastasis in adenocarcinoma of the esophagus and cardia (AEC) with invasion into the muscularis mucosae (m3) or submucosa are based on surgical series. This study aimed to correlate the lymph node metastasis rate with m3 and submucosal infiltration depth of AEC in endoscopic resection specimens. METHODS: Patients undergoing endoscopic resection for AEC between January 2000 and March 2008 at two centers were included if the endoscopic resection specimen showed m3 or submucosal cancer. Infiltration into the muscularis mucosae was defined as m3. Submucosal invasion was classified as sm1 (≤ 500 µm) or sm2/3 (> 500 µm). Exclusion criteria were chemotherapy or radiotherapy and nonradical endoscopic resection. RESULTS: 82 patients included 57 with m3, 12 with sm1, and 13 with sm2/3 cancers. Of the tumors, 13 were poorly differentiated and five showed lymphovascular invasion. After initial endoscopic resection, seven patients underwent surgery and 75 endoscopic therapy. No lymph node metastases were found in 158 lymph nodes of the esophagectomy specimens and none of the endoscopically treated patients were diagnosed with lymph node metastasis during a median follow-up of 26 months (interquartile range [IQR] 14 - 41). CONCLUSION: This study suggests that lymph node metastasis risk for m3 and submucosal AEC may be lower than has been assumed on the basis of surgical series, and that current guidelines are valid regarding suitability of m3 AECs for endoscopic therapy. It may also suggest that selected patients with submucosal cancers are also eligible for endoscopic management. Confirmation of these results is needed in larger series with longer follow-up.

[Iron-deficiency anaemia: a sign that requires an adequate explanation]. / IJzergebreksanemie: zoeken naar een overtuigende verklaring.

In three patients, aged 79, 58 and 59 years, respectively, iron-deficiency anaemia (IDA) was diagnosed. All three had a right-sided colonic adenocarcinoma. In the first patient, a cardiologic cause was looked for instead of a simple laboratory investigation of the anaemia. The second patient received iron supplementation without sufficient diagnostic evaluation, and in the third patient the abdominal X-ray was inadequate for evaluation of the ascending colon; moreover, she responded well to iron supplementation therapy. IDA is a common problem in clinical practice that may have various causes. In the Netherlands, gastrointestinal bleeding is the major cause of IDA in men and women over 50 years of age. The three patients described illustrate that IDA should be considered a clinical sign. An accurate evaluation of the IDA to detect a convincing explanation is therefore necessary. Various causes can be found on duodenoscopy and colonoscopy, e.g. neoplasms, ulcers, angiodysplasia or polyps. Carcinoma of the ascending colon is a frequent cause of IDA, especially in those over 50 years of age and without upper gastrointestinal symptoms. Evaluation to exclude a right-sided carcinoma of the colon has a high priority in these cases.