Human-artificial intelligence interaction in gastrointestinal endoscopy.

The number and variety of applications of artificial intelligence (AI) in gastrointestinal (GI) endoscopy is growing rapidly. New technologies based on machine learning (ML) and convolutional neural networks (CNNs) are at various stages of development and deployment to assist patients and endoscopists in preparing for endoscopic procedures, in detection, diagnosis and classification of pathology during endoscopy and in confirmation of key performance indicators. Platforms based on ML and CNNs require regulatory approval as medical devices. Interactions between humans and the technologies we use are complex and are influenced by design, behavioural and psychological elements. Due to the substantial differences between AI and prior technologies, important differences may be expected in how we interact with advice from AI technologies. Human-AI interaction (HAII) may be optimised by developing AI algorithms to minimise false positives and designing platform interfaces to maximise usability. Human factors influencing HAII may include automation bias, alarm fatigue, algorithm aversion, learning effect and deskilling. Each of these areas merits further study in the specific setting of AI applications in GI endoscopy and professional societies should engage to ensure that sufficient emphasis is placed on human-centred design in development of new AI technologies.

Rectal neuroendocrine tumours and the role of emerging endoscopic techniques.

Rectal neuroendocrine tumours represent a rare colorectal tumour with a 10 fold increased prevalence due to incidental detection in the era of colorectal screening. Patient outcomes with early diagnosis are excellent. However endoscopic recognition of this lesion is variable and misdiagnosis can result in suboptimal endoscopic resection with subsequent uncertainty in relation to optimal long-term management. Endoscopic techniques have shown particular utility in managing this under-recognized neuroendocrine tumour.

Unlocking quality in endoscopic mucosal resection.

A review of the development of the key performance metrics of endoscopic mucosal resection (EMR), learning from the experience of the establishment of widespread colonoscopy quality measurements. Potential future performance markers for both colonoscopy and EMR are also evaluated to ensure continued high quality performance is maintained with a focus service framework and predictors of patient outcome.

Endoscopy training through the COVID-19 pandemic: maintaining procedural volumes and key performance standards.

Objective: Endoscopy departments have experienced considerable challenges in the provision of endoscopy services since the start of the COVID-19 pandemic. Several studies have reported a reduction of procedures performed by trainee endoscopists during the pandemic. The aim of this study was to assess the impact on colonoscopy training and quality in an academic centre throughout successive waves of the pandemic. Methods: This was a single-centre, retrospective, observational study comparing colonoscopies performed at a tertiary endoscopy centre in Ireland at different stages of the pandemic with those performed during a similar time frame prepandemic. Data were collected using electronic patient records. Primary outcomes were procedure volumes, adenoma detection rate and mean adenoma per procedure. Results: In the prepandemic period, 798 colonoscopies were performed. During the same period in 2020, 172 colonoscopies were performed. In 2021, during the third wave of the pandemic, 538 colonoscopies were performed. Percentages of colonoscopies performed by trainees were 46.0% (n=367) in 2019, 25.6% (n=44) in 2020 and 45.2% (n=243) in 2021. Adenoma detection rate was 21.3% in 2019, 38.6% in 2020 and 23.9% in 2021. Mean adenoma per procedure was 0.45 in 2019, 0.86 in 2020 and 0.49 in 2021. Caecal intubation rate was 90.74% in 2019, 90.9% in 2020 and 95.88% in 2021. Conclusion: The COVID-19 pandemic initially had a negative impact on overall colonoscopy volumes and training. Despite a reduction in procedural volume, key performance standards were maintained by trainees. Maintenance of hands-on training is essential to allow trainees achieve and retain competency in endoscopy.

Lymphocyte Activation Gene (LAG)-3 Is Associated With Mucosal Inflammation and Disease Activity in Ulcerative Colitis.

BACKGROUND AND AIMS: Lymphocyte activation gene [LAG]-3 is an immune checkpoint and its expression identifies recently activated lymphocytes that may contribute to inflammation. We investigated the role of LAG-3 by analysing its expression and function in immune cells from blood and tissue of patients with ulcerative colitis [UC]. METHODS: The phenotypic properties of LAG-3+ T cells were determined by flow cytometry, qRT-PCR and single-cell RNA-sequencing. LAG-3+ cells were quantified and correlated with disease activity. The functional effects of LAG-3+ cells were tested using a depleting anti-LAG-3 monoclonal antibody [mAb] in a mixed lymphocyte reaction [MLR]. RESULTS: LAG-3+ cells in the blood were negligible. LAG-3+ lymphocytes were markedly increased in inflamed mucosal tissue and both frequencies of LAG-3+ T cells and transcript levels of LAG3 correlated with endoscopic severity. LAG-3 expression was predominantly on effector memory T cells, and single-cell RNA-sequencing revealed LAG3 expression in activated and cytokine-producing T cell subsets. Foxp3+CD25hi Tregs also expressed LAG-3, although most mucosal Tregs were LAG-3-. Mucosal LAG-3+ cells produced mainly interferon γ [IFNγ] and interleukin-17A. LAG-3+ cell numbers decreased in patients who responded to biologics, and remained elevated in non-responders. Treatment with a depleting anti-LAG-3 mAb led to a reduction in proliferation and IFNγ production in an MLR. CONCLUSIONS: LAG-3+ cells are increased in the inflamed mucosa, predominantly on effector memory T cells with an activated phenotype and their cell numbers positively correlate with disease activity. Depleting LAG-3 eliminates activated proliferating T cells, and hence LAG-3 could be a therapeutic target in UC.

Benefits of exercise in patients with rheumatoid arthritis: a randomized controlled trial of a patient-specific exercise programme.

BACKGROUND: Patients with rheumatoid arthritis (RA) tend to be more overweight, take less physical exercise, exhibit decreased cardiorespiratory fitness and demonstrate reduced muscle strength compared with age- and sex-matched controls. Impaired cognitive function in RA is an important associated factor, although it has been less well-recognized. The aim of this study was to investigate the effects of a specifically designed exercise programme on body composition, aerobic capacity, muscle strength and cognition in RA. METHODS: Sixty-six patients with RA were randomized to a specifically designed, personalized exercise programme or standard care. Assessments included body composition, fitness, grip strength and cognitive testing, in addition to disease related measures. RESULTS: Significant improvements in C-reactive protein (p = 0.025), fatigue scores (p = 0.047) and truncal fat (p = 0.004) were observed in the exercise group compared with controls. Median waist circumference was significantly reduced (94.0 to 91.4 cm, p < 0.0001). Improvements were also seen in aerobic capacity (23.2 to 27.6 ml/kg/min, p = 0.002) and in median right (12.0 to 13.0 kg, p = 0.025) and left grip strength (8.0 to 10 kg, p = 0.005). Cognitive function improved in the exercise group, with median Montreal Cognitive Assessment score 25.5 at 0 months compared to 28.0 at 3 months (p = 0.001). CONCLUSION: This study demonstrates that exercise has a significant and positive impact on cognitive function in RA. Furthermore, physical activity is safe and effective in chronic inflammatory joint disease and is recommended as a vital component in the holistic management of these patients.Key Points• A dedicated physical exercise programme is feasible and safe in patients with rheumatoid arthritis (RA).• Physical exercise helps reduce fatigue scores and improves cardiovascular fitness in stable RA patients.• Physical exercise has a positive impact on cognition in patients with RA.• A structured exercise programme should be an integral part of chronic disease management protocols for patients with RA.

Diagnostic Accuracy of Endoscopic Trimodal Imaging and Chromoendoscopy for Lesion Characterization in Ulcerative Colitis.

BACKGROUND: During surveillance colonoscopy of patients with long-standing ulcerative colitis [UC], a variety of dysplastic and non-dysplastic lesions are detected. The aim of this study was to address the diagnostic accuracy of endoscopic characterization of endoscopic trimodal imaging [ETMI] and chromoendoscopy [CE]. ETMI includes the combination of autofluorescence imaging [AFI], narrow band imaging [NBI] and white light endoscopy [WLE]. METHODS: This is a pre-specified additional analysis of a multi-centre, randomized controlled trial that compared AFI with CE for dysplasia detection in 210 patients with long-standing UC [FIND-UC trial]. In the AFI arm, endoscopists used the ETMI system to record AFI colour, Kudo pit pattern using NBI and WLE for lesion characterization. For AFI, purple colour and ambiguous colour combined with pit pattern type III-V on NBI was considered dysplastic. Kudo pit pattern was described in the CE arm. For pit pattern description using NBI and CE, type III-V was considered dysplastic. Histology was the reference standard. RESULTS: In total, 52 dysplastic and 255 non-dysplastic lesions were detected. Overall sensitivity for real-time prediction of dysplasia was 76.9% (95% confidence interval [CI] 46.2-95.0) for ETMI, and 81.6% [95% CI 65.7-92.3] for CE. Overall negative predictive value [NPV] for ETMI was 96.9% [95% CI 92.0-98.8] and 94.7% [90.2-97.2] for CE. CONCLUSIONS: Sensitivity for endoscopic differentiation of dysplastic lesions detected during surveillance of patients with long-standing UC seems limited using ETMI and CE. Future research is warranted as the high NPV indicates that these techniques are valuable for the exclusion of dysplastic lesions [NTR4062].

Chromoendoscopy versus autofluorescence imaging for neoplasia detection in patients with longstanding ulcerative colitis (FIND-UC): an international, multicentre, randomised controlled trial.

BACKGROUND: Patients with longstanding ulcerative colitis undergo regular dysplasia surveillance because they have an increased colorectal cancer risk. Autofluorescence imaging and chromoendoscopy improve dysplasia detection. The aim of this study was to determine whether autofluorescence imaging should be further studied as an alternative method for dysplasia surveillance in patients with longstanding ulcerative colitis. METHODS: This prospective, international, randomised controlled trial included patients from an ulcerative colitis-dysplasia surveillance cohort from five centres in the Netherlands and the UK. Eligible patients were aged 18 years or older who were undergoing dysplasia surveillance after being diagnosed with extensive colitis (Montreal E3) at least 8 years before study start or with left-sided colitis (Montreal E2) at least 15 years before study start. Randomisation (1:1) was minimised for a previous personal history of histologically proven dysplasia and concomitant primary sclerosing cholangitis. The coprimary outcomes were the proportion of patients in whom at least one dysplastic lesion was detected and the mean number of dysplastic lesions per patient. The relative dysplasia detection rate, calculated as the ratio of the detection rates by autofluorescence imaging and chromoendoscopy, needed to be more than 0·67 (using an 80% CI) for both primary outcomes to support a subsequent large non-inferiority trial. Outcomes were analysed on a per-protocol basis. The trial is registered at the Netherlands Trial Register, number NTR4062. FINDINGS: Between Aug 1, 2013, and March 10, 2017, 210 patients undergoing colonoscopy surveillance for longstanding ulcerative colitis were randomised for inspection with either autofluorescence imaging (n=105) or chromoendoscopy (n=105). Dysplasia was detected in 13 (12%) patients by autofluorescence imaging and in 20 patients (19%) by chromoendoscopy. The relative dysplasia detection rate of autofluorescence imaging versus chromoendoscopy for the proportion of patients with ulcerative colitis with at least one dysplastic lesion was 0·65 (80% CI 0·43-0·99). The mean number of detected dysplastic lesions per patient was 0·13 (SD 0·37) for autofluorescence imaging and 0·37 (1·02) for chromoendoscopy (relative dysplasia detection rate 0·36, 80% CI 0·21-0·61). Adverse events were reported for two patients in the autofluorescence imaging group (one patient had intraprocedural mild bleeding, and one patient had abdominal pain) and for three patients in the chromoendoscopy group (two patients had intraprocedural mild bleeding, and one patient had perforation). INTERPRETATION: Autofluorescence imaging did not meet criteria for proceeding to a large non-inferiority trial. Therefore, existing autofluorescence imaging technology should not be further investigated as an alternative dysplasia surveillance method. FUNDING: Olympus Europe and Olympus Keymed.

Diagnostic Yield of Dysplasia in Polyp-adjacent Biopsies for Patients with Inflammatory Bowel Disease: A Cross-sectional Study.

INTRODUCTION: Patients with inflammatory bowel disease (IBD) undergoing polypectomy are recommended by current guidelines to have biopsies taken from adjacent mucosa to determine whether there is dysplasia present. With improvements in endoscopic imaging, it is now possible to characterize colonic lesions with higher levels of confidence than previously. We have reviewed the diagnostic yield of polyp-adjacent biopsies in IBD. MATERIALS AND METHODS: A systematic search of our histopathology database revealed cases in which polyps had been endoscopically resected or biopsied in patients with IBD. Endoscopy reports and medical records were reviewed, and patient demographic and disease-specific details were recorded, along with details of polyp characteristics and histopathology outcomes. RESULTS: Three hundred and two polyps were biopsied or resected in 131 patients undergoing 178 colonoscopies. The median polyp size was 4 mm (range 1-45), and the predominant morphology was Paris 0-Is (n = 98, 32%). The histology was tubular adenoma in 76 (25%), tubulovillous adenoma in 14 (5%), hyperplastic in 112 (37%), post-inflammatory in 32 (11%), sessile serrated polyp in 31 (10%), traditional serrated adenoma in 2 (0.7%), flat high-grade dysplasia or cancer in 2 (0.7%) and other in 33 (11%). Dysplasia in adjacent biopsies was detected in 2 patients (0.7%), and was endoscopically visible in both cases. The proportion of endoscopically unsuspected dysplasia was 0/300 (0%, 95% CI 0-1.6%). CONCLUSION: The diagnostic yield for polyp-adjacent biopsies in patients with IBD is negligible. With high-definition technology and chromoendoscopy, it may no longer be necessary to biopsy endoscopically normal adjacent tissue to detect invisible dysplasia.

Diagnostic and therapeutic treatment modalities for acute lower gastrointestinal bleeding: a systematic review.

BACKGROUND AND STUDY AIMS: Investigations for lower gastrointestinal bleeding (LGIB) include flexible sigmoidoscopy, colonoscopy, computed tomographic angiography (CTA), and angiography. All may be used to direct endoscopic, radiological or surgical treatment, although their optimal use is unknown. The aims of this study were to determine the diagnostic and therapeutic yields of endoscopy, CTA, and angiography for managing LGIB, and their influence on rebleeding, transfusion, and hospital stay. PATIENTS AND METHODS: A systematic search of MEDLINE, PubMed, EMBASE, and CENTRAL was undertaken to identify randomized controlled trials (RCTs) and nonrandomized studies of intervention (NRSIs) published between 2000 and 12 November 2015 in patients hospitalized with LGIB. Separate meta-analyses were conducted, presented as pooled odds (ORs) or risk ratios (RR) with 95 % confidence intervals (CIs). RESULTS: Two RCTs and 13 NRSIs were included, none of which examined flexible sigmoidoscopy, or compared endotherapy with embolization, or investigated the timing of CTA or angiography. Two NRSIs (57 - 223 participants) comparing colonoscopy and CTA were of insufficient quality for synthesis but showed no difference in diagnostic yields between the two interventions. One RCT and 4 NRSIs (779 participants) compared early colonoscopy (< 24 hours) with colonoscopy performed later; meta-analysis of the NRSIs demonstrated higher diagnostic and therapeutic yields with early colonoscopy (OR 1.86, 95 %CI 1.12 to 2.86, P  = 0.004 and OR 3.08, 95 %CI 1.93 to 4.90, P  < 0.001, respectively) and reduced length of stay (mean difference 2.64 days, 95 %CI 1.54 to 3.73), but no difference in transfusion or rebleeding. CONCLUSIONS: In LGIB there is a paucity of high-quality evidence, although the limited studies on the timing of colonoscopy suggest increased rates of diagnosis and therapy with early colonoscopy.

Endoscopic approach to polyp recognition.

Identification and complete resection of colorectal polyps provide a significant mortality benefit from colorectal cancer. With improvements in colonoscopic techniques and advanced endoscopic imaging techniques, polyp detection has taken on greater complexity since the establishment of bowel cancer screening programmes internationally. All endoscopists operating within symptomatic and screening populations should be aware of endoscopic features associated with advanced neoplasia. Chromoendoscopy and advanced imaging techniques, such as narrow spectrum technologies (narrow band imaging, flexible spectral imaging colour enhancement (FICE) and i-Scan digital contrast (iSCAN)), have specific classification systems to support accurate lesion characterisation. This review summarises the evidence in relation to polyp detection, recognition and characterisation as well as the identification of features of invasion. Future areas of interest include optimal management of large polyps, incorporation of a 'detect, resect and discard' strategy for small and diminutive polyps, expected wider use of computer decision support tools (artificial intelligence and deep learning) and the use of fluorescently labelled molecular probes to improve detection and assessment of neoplasia.

Osteoclastic-Type Giant Cell Tumours of the Pancreas: A Homogenous Series of Rare Tumours Diagnosed by Endoscopic Ultrasound.

BACKGROUND: Giant cell tumors (GCT) of the pancreas are a rare form of pancreatic cancer. Although data are limited, clinical outcomes appear to depend largely on histological subtype with osteoclastic tumors carrying a better prognosis. We report on a homogenous series of patients with osteoclastic-type GCTs of the pancreas presenting to a national pancreatico-biliary gastrointestinal oncology center. METHODS: Patients underwent endoscopic, radiological and histopathological assessments. Data were collected in relation to consecutive patients presenting with osteoclastic-type tumors of the pancreas and analyzed with survival as a primary end point. RESULTS: Four patients were treated over a 4-year period. Median age was 77 years with equal gender distribution. Median tumor size was 42 mm. Histology was osteoclast-type giant cells in all 4 patients. Two patients underwent surgery with curative intent. Median overall survival was 13.1 months. CONCLUSION: This is the largest reported series of osteoclast-type histology in GCTs of the pancreas.