Introduction

Endoscopic ultrasound-guided gastroenterostomy (EUS-GE) offers advantages over duodenal stenting and surgical bypass. EUS-GE technique, however, remains non-standardized. A parallel enteric tube (PET) facilitates luminal distention, a critical step during EUS-GE.

Aims & Methods

We primarily aimed at standardizing PET-assisted EUS-GE, secondarily assessing reproducibility and outcomes.
Prospective multicenter study on consecutive adult patients with unresectable malignant gastric outlet obstruction undergoing primary EUS-GE between August 2019-April 2021. Hierarchical task analysis predefined procedural steps into tasks and subtasks. Subtasks were further categorized into essential (performed in all centers and in over 85% of the procedures) or optional. Subtask methodology was considered established if performed similarly in all centers or variable if not. Procedure times, injected fluid volume, luminal diameters obtained, accessories used, and clinical outcomes were recorded.

Results

Seven endoscopists performed EUS-GE in 65 patients (50.8% male, median [IQR] age 77.5 [65.7-86.5] years). EUS-GE was categorized into 4 tasks and 10 subtasks (Table 1). Five essential subtasks involved an established methodology (guidewire and PET placement, endoscope exchange, delivery system insertion and LAMS deployment).
Technical and clinical success rates were 98.5% and 83.3%, respectively. AEs occurred in 10 (15.4%) patients (2 mild, 7 moderate, 1 severe). Success and AE rates were not different between expert and non-experts. Median (IQR) procedure time was 26.2 (19.7-37.1) min. The short and long axis of the targeted small bowel loop was 27 (23-30) x 40 (33-60) mm. Procedure time was longer (35 [30.6-43.7] vs 21.8 [16.4-29.5] min, p<0.001) and injected fluid volume higher (510 [439-870] vs 415 [255-480] ml, p=0.01) in non-experts.

Table 1. Proposed taxonomy for the PET-assisted EUS-GE

Conclusion

PET-assisted EUS-GE was standardized, identifying its key steps and technique variants. PET-assisted EUS-GE appears to be a reproducible procedure among advanced endoscopists with different levels of experience. (Clinical Trials Identification Number: NCT04660695).

Disclosure

Carlos Chavarríareceived research and training support by a grant from The Foundation of the Institute of Health Sciences Studies of Castilla y León (IECSCYL).
Vanessa Martín-Álvarezhas nothing to declare.
José Ramón Aparicio is a consultant for Boston Scientific.
Jose Carlos Subtil has nothing to declare.
Francisco Javier Garcia-Alonso has nothing to declare.
Juan J. Vila is a consultant for Boston Scientific and speaker for Cook Endoscopy and Olympus.
Belén Martínez-Moreno has nothing to declare.
Victoria Busto-Bea has nothing to declare.
Ramon Sanchez-Ocaña has nothing to declare.
Carlos de la Serna-Higuera has nothing to declare.
Manuel Perez-Miranda Miranda is a consultant and speaker for Boston Scientific, Olympus, and M.I.Tech.

Introduction

The use of artificial intelligence (AI) systems for decision support has recently gained great popularity in the medicine.¹ However, recent studies have shown possible limitations emerging from the interaction between humans and AI. Among these, the term “automation bias (AB)” refers to human errors deriving from over-relying on AI’s false alarms such that the human changes a correct initial interpretation.² AB is poorly investigated in the medical field and was never tested in the gastroenterological setting.

Aims & Methods

The aim of the study was to investigate the impact of AI on human diagnostic performance in gastrointestinal endoscopy, including AB.
Ninety high-quality endoscopic videoclips (16 seconds) extracted from small bowel (SB) capsule endoscopy (N=60) and proctosigmoidoscopy (N=30) were edited with the addition of a mock AI system designed with diagnostic accuracy, sensitivity and specificity of 80%, 86% and 75%. The ground truth was settled on the agreement of 9 established experts about the presence of SB lesions with significant hemorrhagic potential according to the Saurin classification (30 videoclips), inflammatory lesions relevant for the diagnosis of SB Crohn's disease according to the Lewis score (30 videoclips) and assessing the Ulcerative Colitis (UC) Endoscopic Index of Severity (30 videos). The videoclips were randomly divided into three sessions and submitted around Europe to fully trained endoscopist, trainees and endoscopic technicians. Participants were instructed to the accuracy measures of AI annotations and scored each videoclip first without, then with AI support. Participants’ performances with and without AI were calculated. Subgroup analyses were made after stratification by participants’ experience level in each endoscopic setting and test performances.

Results

A total of 12.660 videoclips (both with and without AI) were visualized in 422 sessions (182, 118 and 122 participants completed the first, second and third session, respectively). A strong and significant difference between the respondents' performance (N=281) when they were unaided by AI (average accuracy (M)=0.78, [0.77, 0.79] and when they were supported by the AI (M=0.81, [0.80, 0.82], paired t-test, t(276)=11.2, p<0.001) was found. Size effect and the evidence in favour of a significant positive effect of AI were large, 0.67 (Cohen's D) and very large (Bayes Factor>100), respectively. Interestingly, no significant difference between trainees and specialists both in regard to their performance, either unaided, M=0.77 [0.74, 0.79] vs M=0.78, [0.77, 0.79] or aided M=0.80 [0.78, 0.82] vs M=0.81 [0.80, 0.82] and their accuracy improvement (M=0.03, [0.02, 0.05] vs M=0.03, [0.02, 0.04] (unaided t-value=-0.89, p=0.37; aided t-value=-0.69, p=0.49, difference t-value=0.54, p=0.59) were found. A strong leveller effect emerged, with low-performers improving their accuracy much more than top-performers (average improvement 5%, [0.04, 0.06] vs 1%, [0.00, 0.02], paired T test, t=6.53, p<0.001). Nine percent of the final decisions involved disregarding correct AI advice in favour of incorrect human judgment, while AB occurred in 2% of cases.

Conclusion

In this study, the application of AI to assess SB bleeding, SB inflammatory lesions, and UC activity brings significant improvement in diagnostic accuracy, with the benefit appearing to be greater in the low-performers. However, in rare cases, AI may induce diagnostic errors due to AB. Future studies directly evaluating the impact of AI in real-life clinical settings are needed to confirm these data.

References

1. Cabitza F, Campagner A, Angius R, Natali C, Reverberi C. 2023. AI Shall Have No Dominion: on How to Measure Technology Dominance in AI-supported Human decision-making. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (CHI '23). Association for Computing Machinery, New York, NY, USA, Article 354, 1–20. https://doi.org/10.1145/3544548.3581095
2. Reverberi, C., Rigon, T., Solari, A. et al. Experimental evidence of effective human–AI collaboration in medical decision-making. Sci Rep 12, 14952 (2022). https://doi.org/10.1038/s41598-022-18751-2

Introduction

This study aimed to analyze the correlation between different attributes of doctors who prescribe colonoscopies and the cleanliness of the intestine to guide the development of colonoscopy application protocols.

Aims & Methods

Data on colonoscopy cases conducted in the gastroenterology department of Hangzhou First People's Hospital between April 2018 and March 2021 were collected. The gender, age, professional attributes of the prescribing doctors, and Boston Bowel Preparation Scale (BBPS) score were recorded. In addition, the correlation between the prescribing doctors' characteristics and the intestine's cleanliness was analyzed.

Results

The study included 22,522 patients with a mean BBPS score of 6.83±1.94. There were 16,459 male and 6,063 female doctors with similar BBPS scores (P = 0.212). The study found no significant difference in BBPS scores between 19,338 physicians and 3,184 non-physicians (P = 0.154). However, BBPS scores differed significantly between 18,168 gastroenterologists and 4,354 non-gastroenterologists (P = 0.016) and between 19,990 intestinal-related specialties and 2,532 non-intestinal-related specialties (P = 0.000). In addition, BBPS scores were significantly different between 18,126 prescribing doctors with colonoscopy ability and 4,352 without (P = 0.014). However, there was no significant difference in BBPS scores among doctors of different ages (P = 0.190). The study found significant differences in BBPS scores between male and female patients and those under or over 40 years (P = 0.000).

Conclusion

To improve colonoscopy preparation quality, priority should be given to doctors in gastroenterology, intestinal-related specialties, and with colonoscopy operation ability. Their expertise may result in better education and improved bowel cleanliness.

Clinical Case Summary

A 74-year-old, female patient was referred to the study center for pancreatic cysts. Magnetic resonance imaging revealed multiple pancreatic cysts up to 15 mm in size distributed from the pancreatic head to the pancreatic tail and enlargement of the main pancreatic duct diameter to 8 mm. T1- and T2-weighted imaging demonstrated showed a 10 mm mass with iso-intensityin the main pancreatic duct in the pancreatic body with diffusion restriction on diffusion-weighted imaging. Endoscopic ultrasonography also demonstrated a mass located in the main pancreatic duct in the pancreatic body, but determining whether the tumor was located in the pancreatic duct or was an extension of the intraparenchymal lesion into the main pancreatic duct was difficult. On contrast-enhanced abdominal computed tomography, the mass was difficult to identify as a lesion because it showed the same contrast as the surrounding pancreatic parenchyma. Endoscopic retrograde choangiopancreatography revealed enlargement of the duodenal papillary entrance, mucus discharge, and a mass-like, translucent image with an irregular surface in the main pancreatic duct in the pancreatic body. Endoscopic ultrasound-guided fine needle aspiration led to the diagnosis of adenocarcinoma, and a distal pancreatectomy was performed on the assumption that the lesion derived from IPMN. Pathological analysis led to the diagnosis of IOPN consisting of papillary cells with an eosinophilic cytoplasm, positive immunostaining for MUC5AC and MUC6, and partially positive staining for MUC1 and HepPar-1. The mass was determined to be within the pancreatic parenchyma by comparing a resection specimen with preoperative images. Fusion gene analysis revealed the presence of ATP1B1-PRKACA, confirming the diagnosis of IOPN.
It is important always to bear in mind the possibility of IOPN if pancreatic cysts are observed. IOPN is relatively rare, and more cases are needed for a better understanding of this disease.

References

Aatur D. Singhi, Laura D.Wood, Emma Parks, Michael S.Torbenson, Matthaus Felsenstein. Recurrent Rearrangements in PRKACA and PRKACB in Intraductal Oncocytic Papillary Neoplasms of the Pancreas and Bile Duct : Gastroenterology. 2020 Feb ; 158(3) : 573-582.
Eriko Ikeda. Histological Classification of Intraductal Papillary Mucinous Neoplasms of the Pancreas : Journal of Biliary Tract & Pancreas. 2020 ; 41(1) : 27-31

Introduction

Autoimmune pancreatitis (AIP) is a chronic, relapsing form of pancreatitis, often affecting multiple organs. Although the exact pathomechanism is incompletely understood, plasmacytoid dendritic cells (pDC) and their cytokines IL-33 and interferon alpha-1 (IFNA1) are known to play a central role in AIP pathophysiology. Through activation of IFN-regulatory factors 3 and 7 (IRF3/IRF7) in pDC, IFNA1 transcription is initiated, likely offsetting the inflammatory cascade. AIP often leads to exocrine atrophy and insufficiency, however it is currently unknown whether pDCs affect acinar cell regeneration or their cell death.

Aims & Methods

This study aims to understand the effect of the IFN-cascade offset by pDCs on acinar regeneration capacity and cell death in AIP. The MRL/Mp mouse model for AIP was used to first characterize the immune cell infiltrate in AIP compared to acute pancreatitis (AP) via immunohistochemistry (IHC) staining. To test the impact of pDCs and Ifna1 release on pancreatic acinar cells, pDCs were generated from bone marrow of wild type (WT) and Irf3/7-knock out (Irf3/7-KO) mice and murine acinar cells were incubated with supernatant of WT and Irf3/7-KO pDCs, or pDC-related cytokines IL-33 and IFN-a. Gene expression via qPCR, cell stress via LDH assay and cytokine array were performed to measure the effects on acinar cells.

Results

AIP in MRL/Mp mice exhibited a stronger immune infiltrate in comparison to mice with an AP. The immune infiltrate in AIP was characterized by a dense invasion of CD8+ T cells, presence of CD11c+ cells (pDCs) and complete absence of neutrophils. Furthermore, Irf3 and Irf7 expressions were increased not only in immune but also acinar cells in AIP, whereas decreased in acinar cells and not at all expressed in immune cells in AP. Proliferation markers were not expressed in acinar cells in AIP, whereas showed some expression in acinar cells in AP. Acinar cells treated with pDC supernatant of WT revealed an increase in apoptosis signaling genes, whereas acinar cells treated with pDC supernatant of Irf3/7-KO mice showed a downregulation of apoptotic genes.

Conclusion

This study reveals that pDCs dictate acinar cell death signaling by regulating apoptosis via the IFN-cascade. This might be a distinct feature of AIP, in which the clinical course is moderate and only apoptosis is present, compared to AP, in which the disease course is much more severe and necrosis is a prominent feature. Further studies are needed to examine the role of pDC in necrosis and acinar regeneration.

References

1. Hart PA, Zen Y, Chari ST. Recent Advances in Autoimmune Pancreatitis. Gastroenterology. 2015;149(1):39–51.
2. Watanabe T, Yamashita K, Arai Y, Minaga K, Kamata K, Nagai T, et al. Chronic Fibro-Inflammatory Responses in Autoimmune Pancreatitis Depend on IFN-α and IL-33 Produced by Plasmacytoid Dendritic Cells. J Immunol. 2017;198(10):3886–96.
3. Minaga K, Watanabe T, Arai Y, Shiokawa M, Hara A, Yoshikawa T, et al. Activation of interferon regulatory factor 7 in plasmacytoid dendritic cells promotes experimental autoimmune pancreatitis. J Gastroenterol. 2020;55(5):565–76.