Introduction

Nodal metastases (N+) are one of the most powerful predictors of disease recurrence after surgery for non-functioning pancreatic neuroendocrine tumors (NF-PanNETs). However, the prognostic role of nodal metastases < 5 mm, defined as micrometastases, has been poorly investigated so far.

Aims & Methods

The aim of this study was to compare clinico-pathological features and survival outcomes between patients with NF-PanNETs without nodal involvement (N0), with nodal micrometastases (microN+) and with nodal metastases (macroN+). Consecutive patients who underwent a formal pancreatic resection for NF-PanNETs at San Raffaele Hospital (Milan, Italy) between 2018 and 2021 and were enrolled in the DETECTYON trial (NCT03918759) were considered for the purpose of the present study.

Results

Overall, 100 patients were included. Of these, 58 had N0 PanNETs, 15 had microN+ PanNETs and 27 had macroN+ neoplasms. Patients with macroN+ had significantly larger tumors [median 35 mm (IQR 28-47)] as compared to patients with microN+ [25 mm (IQR 24-35), p=0.040] and N0 neoplasms [26 mm (IQR 18-34), p=0.003]. The rate of G2-G3 neoplasms was comparable between patients with N0 and microN+ PanNETs (n=26/58 - 45% versus n=4/15 - 27%, p=0.203), whereas it was significantly higher among subjects with macroN+ tumors (n=21/27 - 78%). Patients with microN+ had a significantly lower Ki67 proliferative index as compared to those with macroN+ [2% (IQR 1-3) versus 6% (IQR 3-9), p=0.003]. Patients with N0 and microN+ neoplasms had microvascular invasion in 47% of cases (both), as compared to 82% of cases in patients with macroN+ (p=0.008). After a median follow-up of 37 months, 16 patients (16%) experienced disease relapse. Patients with N0 PanNETs had a 4-year DFS rate of 97% as compared with 88% and 43% in patients with microN+ (p=0.152) and macroN+ (p<0.001), respectively. At multivariable analysis, distant metastases (HR 5.826, p=0.026) and macroN+ (HR 6.281, p=0.034) were identified as independent determinants of disease relapse.

Conclusion

NF-PanNETs with microN+ seem to be associated with a risk of recurrence similar to N0 neoplasms upon their removal. MicroN+ may be regarded as a clinicopathological entity separate from macroN+, with a possible impact on postoperative surveillance protocols.

Disclosure

No conflicts of interest.

Introduction

It is unclear whether the insertion of an axis-orienting double-pigtail plastic stent (DPS) through biliary lumen-apposing metal stents (LAMS) in EUS-guided choledochoduodenostomy (CDS) improves the stent patency.

Aims & Methods

The aim is to determine whether this technical variant offers a clinical benefit in EUS-guided biliary drainage (CDS-type) for the management of distal malignant biliary obstruction. This is a multicentre, open-label, randomised controlled trial with two parallel groups. The primary endpoint was the rate of recurrent biliary obstruction (RBO), as a stent dysfunction parameter, detected during follow-up. Secondary endpoints: technical and clinical success (reduction in bilirubin >50% within 14days), adverse events, and others.

Results

Between November-2020 and October-2022, 123 patients (pancreatic cancer 85%, palliative status 80%) were screened and 84 with malignant biliary obstruction, and underwent EUS-CDS (rescue after failed ERCP, 62%), were randomised: 44 patients in LAMS group and 40 in LAMS-DPS group. The rate of RBO (29.5 vs 10%; OR 0.26 [IC95% 0.07–0.83],p=0.04) was lower, but with a longer procedure time (30 vs 20-min, p=0.02), in LAMS-DPS-group. Furthermore, less biliary reinterventions (12.5 vs 20.5%, p=0.38), longer time-to-RBO (155.5 vs 51.5 days, p=0.10), and shorter hospital stay (6 vs. 8.5,p=0.06) were encountered in LAMS-DPS group. No differences were found between technical (97 vs. 88%), clinical success (92 vs. 88.8%) and AEs (22.5 vs.11.4%). NCT04595058 ClinicalTrials.gov

Table1: Outcomes (RBO, recurrent biliary obstruction; BRI, biliary reinterventions).

Conclusion

This trial confirms that CDS with LAMS is an effective technique. The technical strategy of placing a pigtail within the LAMS is associated with a lower risk of RBO, and a non-significant trend of shorter hospital stay.

References

1-Garcia-Sumalla, A., Loras, C., Sanchiz, V. et al. Multicenter study of lumen-apposing metal stents with or without pigtail in endoscopic ultrasound-guided biliary drainage for malignant obstruction—BAMPI TRIAL: an open-label, randomized controlled trial protocol. Trials 2022; 23:181
2-Gornals JB, Consiglieri CF, Bergamino MA. Double pigtail for preventing ascending cholangitis after endoscopic ultrasonography-guided choledochoduodenostomy with lumen-apposing metal stent. Dig Endosc. 2016;28:100.
3-Garcia-Sumalla A, Loras C, Guarner-Argente C, et al. Is a coaxial plastic stent within a lumen-apposing metal stent useful for the management of distal malignant biliary obstruction? Surg Endosc. 2021;35:4873 –81

Disclosure

- A Garcia-Sumalla is a PhD student at the Faculty of Medicine and Health Sciences, University of Barcelona, and this paper is part of his doctoral project.
- Joan B Gornals acts as a PhD Director; is a consultant for Boston Scientific and has received a research grant from this company.
- Manuel Pérez-Miranda is a consultant and speaker for Medtronic, Olympus, Taewoong, and M.I.Tech.
- Enrique Vazquez-Sequeiros is a consultant for Boston Scientific.
Jose Ramon Aparicio is a consultant for Boston Scientific

Introduction

The success rate of endoscopic retrograde cholangiopancreatography (ERCP) is suboptimal in patients with complex surgically altered anatomy (e.g. Roux-en-Y anastomosis, pancreaticoduodenectomy) due to difficulty in advancement of scope to the target site and selective biliopancreatic duct cannulation. Although balloon-assisted enteroscopy can achieve a higher rate of intubation, its clinical application is limited by its small working channel and few compatible accessories.

Aims & Methods

We aimed to explore the feasibility and effectiveness of a novel endoscopic approach in patients with complex surgically altered anatomy.
This is a retrospective study of patients who underwent ERCP with complex surgically altered anatomy at our center with an annual ERCP volume of approximately 3000, between January 2018 and November 2022. The endoscopic procedures were performed by three senior endoscopists who had ERCP experience for at least 5 years. Long colonoscopy (measuring 168cm)-assisted ERCP was adopted as a first-line intervention, and selective biliopancreatic duct cannulation was performed after the target site (native papilla or the surgical anastomosis) was reached. For patients with failed cannulation, a plastic guide tube measuring 250 cm was inserted via the working channel of the colonoscope and its distal end was placed near the target site. The colonoscope was then withdrawn while the guide tube was left in-situ. By following the guide tube, a conventional side-viewing duodenoscope was cautiously advanced to the target site and biliopancreatic duct cannulation re-attempted.

Results

A total of 58 patients were recruited (mean age:56.7 years [standard deviation [SD]:14.4; male:36 [62.1%]]. The reconstruction methods were as follows – Roux-en-Y hepaticojejunostomy:32 (55.2%), Billroth-II gastrectomy with Braun anastomosis:11 (19.0%), pancreaticoduodenectomy:10 (17.2%) and Roux-en Y gastrectomy:5 (8.6%). Common bile duct stone was the most common indication for ERCP (n=30; 51.7%), followed by bilioenteric anastomotic stenosis (n=15; 25.9%), pancreaticojejunal anastomotic stenosis (n=7; 12.1%) and chronic pancreatitis (n=6; 10.3%). 55 (94.8%) of 58 had successful intubation by long colonoscope with a mean procedure time of 57.7minutes (SD:23.4). The mean time to reach target site was 19.1minutes (SD:9.0). The target site could not be identified in two patients and intestinal perforation occurred during endoscopic intubation in one patient. Among those with successful endoscopic intubation, biliopancreatic duct cannulation was achieved in 35 (63.6%) patients. For the remaining 20 patients with initial failure of biliopancreatic duct cannulation, intubation rate by duodenoscope was 100% by tracing the guide tube to target site. The success rate of biliopancreatic duct cannulation by duodenoscope was 85% (17/20) with a mean procedure time of 83.5 minutes (SD:21.1). The mean time to reach target site was 17.0minutes (SD:6.1). Notably, 90% (18/20) had either native papilla or pancreaticojejunostomy, in which selective biliopancreatic duct cannulation is known to be technically difficult with any type of forward-viewing scope due to unfavorable cannulation trajectory. For those with failed cannulation, two had anatomical stricture and one had impacted pancreatic stone.

Conclusion

ERCP by conventional duodenoscope with colonoscopy-assisted pre-placement of guide tube is a feasible, effective and safe alternative for patients with complex surgically altered anatomy in experienced endoscopists.

Introduction

Hiatal hernia (HH) is a common finding in endoscopic examinations that can cause significant symptoms, depending on its severity. The Hill classification assesses the degree of HH (type II to IV) by considering the width of the cardia and the hiatus on visual inspection during gastroscopy.

Aims & Methods

Our goal was to train a deep learning (DL) model that determines the Hill classification using a novel active learning (AL) pipeline. A total of 21,970 gastroscopic images representing 14,885 examinations from two endoscopy centers were retrospectively identified as the training dataset. A novel AL pipeline was developed to train the DL model. For comparison reasons, a second model was trained with randomly selected images from the same training dataset. An external, publicly available collection of HH images was annotated using the Hill classification and used as test data. Outcomes are the ability of the models to diagnose and classify HH and the time spent by the expert for annotation. The proposed AL pipeline is provided as Open-Source.

Results

The AL-based model was generated in eight consecutive steps, covering about 10% of the unlabeled training data. The model achieved mean accuracy, sensitivity, specificity, and F1-score of 0.878, 0.692, 0.896, and 0.67 respectively, when classifying the four different types of Hill. In diagnosing the existence of a HH (type I vs II-IV), the model achieved 0.828, 0.749, 0.902, 0.808 for the same metrics. The median annotation time per image was 3.62s (Q1-Q3; 3.14-4.69). Obtaining a model via the random image selection that performs similarly, required at least 30% more annotations and thus expert time commitment.

Conclusion

In this work we present a novel AL pipeline, enabling examiners to develop new models for gastrointestinal endoscopy without requiring programming skills. Using the proposed pipeline resulted in a well performing model in less time spent by the expert annotator, compared to traditional model training.

Disclosure

The authors have no conflicts of interest to disclose.

Introduction

Atrophic gastritis(AG) is mainly caused by Helicobacter pylori(Hp)infection or autoimmunity. Hp-related AG affects both corpus and antral mucosa resulting in multifocal AG(MF-AG), whereas AG driven by autoimmunity is corpus-restricted(CR-AG).
Corpus mucosa gastric atrophy carries an increased risk of gastric dysplasia(GD) and cancer(GC);the neoplastic risk in CR-AG is considered lower, but still debated. GD-GC lesions in AG may arise in the antrum, incisura angularis or, less frequently,the corpus.Data on characteristics and potential differences of GD-GC lesions between MF-AG and CR-AG are lacking.

Aims & Methods

This study aimed to assess clinical,endoscopic,and histological characteristics of GD-GC in MF-AG and CR-AG pts.
A multicenter,cross-sectional study was conducted across 14 Italian GI centers. Data on GD-GC in patients with MF-AG or CR-AG(2001-2023) were retrospectively collected based on clinical,endoscopic, and histological charts: site,dimension,endoscopic,and histologic features of GD-GC, perilesional gastric antral and corpus mucosa(updated Sydney system),and patients’ clinical data (age,sex,outcome after GD-GC diagnosis,Hp infection, parietal cell antibodies(PCA), family history for GC). Inclusion criteria were:patients with MF-AG or CR-AG and GD or GC diagnosis,aged>18 years, availability of endoscopic and histological description of GD-GC and perilesional gastric mucosa, completeness of clinical data.Data were stratified for MF-AG/CR-AG and the GD-GC, perilesional mucosa, and clinical characteristics were compared.

Results

A total of 84 pts[(F46[54.8%];age 70(33-90yrs)] were included. MF-AG and CR-AG were observed in 45(53.6%) and 39(46.4%) pts,respectively. High-grade(HG)-GD,low-grade(LG)-GD,and non-cardia GC were diagnosed in 6(7.1%),31(36.9%), and 47(56.0%)pts.
Age, GC family history,smoking habits, and PCA positivity were similar between MF-AG and CR-AG.Women(69.2% vs 42.2%,p=0.01) and pernicious anemia(29% vs 9.1%,p=0.04) were more frequent in CR-AG than in MF-AG.
The proportion of GD and GC in MF-AG and CR-AG pts was similar:HG-GD was present in 4(8.9%) vs 2(5.1%),LG-GD in 17(37.8%) vs 14(35.9%),GC in 24(53.5%) vs 23(59.0%)pts(p>0.05).
Compared to MF-AG,in CR-AG pts GD-GC were more frequently polypoid lesions (51.6% vs 27.3%,p=0.048) and more frequent in the corpus (55.3% vs 28.6%,p=0.02), but also occurred in the antrum(34.2%) and incisura(10.5%).The intestinal-type GC was the most prevalent and was more frequent in MF-AG than in CR-AG (87%vs60.9%,p>0.046), while diffuse GC was 2 times more frequent in CR-AG than in MF-AG, albeit not reaching statistical significance (26.1% vs 13%, p=0.27).
Concerning GC differentiation grade and staging, no significant differences were observed between MF-AG and CR-AG.Endoscopic treatment of GD-GC was similar in both groups. CR-AG patients underwent surgery more frequently than MF-AG(48.6% vs 23.1%,p=0.0207), but outcome was similar with 90.7% MF-AG and 92.1% CR-AG pts being alive.
Histopathological features of corpus mucosa (acute-chronic inflammation, severity of atrophy,pseudopyloric and intestinal metaplasia) were similar between MF-AG and CR-AG(p>0.05). The histological evidence of Hp in the antrum and corpus was low in both groups (2.3% vs 2.9%,p=0.87).

Conclusion

Non-cardia GC and GD may occur in both,MF-AG and CR-AG, displaying differences in topography and endoscopic presentation, but similarities in perilesional gastric corpus mucosa, differentiation, staging and outcome.
Endo-histological surveillance should be considered in pts with gastric corpus atrophy, irrespective of its etiology and extension.

References

Lahner E, et al. Dig Liver Dis. 2019 Dec;51(12):1621-1632.
Lenti MV, et al. Nat Rev Dis Primers. 2020 Jul 9;6(1):56.
Rugge M, et al. Gut. 2023 Jan;72(1):30-38.