Introduction

While cold snare polypectomy (CSP) has been proven effective and safe in the resection of small polyps, it is not actively recommended for the pedunculated (Ip type) polyps in the European Society and Gastrointestinal Endoscopy and US Multi-Society Task Force on Colorectal Cancer guidelines. Although Japanese guidelines recommend that CSP should be limited to lesions less than 10 mm, preoperatively diagnosed as adenomas, the guideline does not address how to resect pedunculated (Ip type) adenomas. This study aims to examine the pros and cons of CSP for the resection of Ip type adenoma.

Aims & Methods

Among patients with colorectal adenomas resected endoscopically at Nippon Medical School Hospital in Japan, from January to July 2020, 85 lesions in 67 consecutive cases of type Ip, and 762 lesions in 346 cases of non-Ip type (types Is, Isp, and II) resected by CSP, were observed. Regarding clinicopathologic features, R0 resection rate, and adverse events, we evaluated the difference in the resection method (Ip-CSP group vs Ip-hot polypectomy (HP)/EMR group) as Study 1; and the macroscopic classifications (Ip-CSP group vs Non-Ip-CSP group) as Study 2.

Results

In Study 1, 23 lesions were included in the Ip-CSP group and 62 lesions were included in the Ip-HP/EMR group. A higher percentage of proximal colon lesions (74% vs. 27%, p<0.001) and smaller mean tumor size (5.1±1.6mm vs. 11±5.0mm, p<0.001) were observed in the Ip-CSP group, compared to Ip-hot polypectomy (HP)/EMR group. The R0 resection rate was also significantly lower in Ip-CSP group (61% vs. 90%, p=0.03). While the R0 resection rate tended to be low for lesions 6 mm or greater (67%/93%), no difference was observed among lesions less than 5 mm (59%/50%). No perforation or postoperative bleeding was observed in both groups. In Study 2 using 23 lesions for the Ip-CSP group and 762 lesions for the Non-Ip-CSP group, no significant difference was observed in tumor location or mean tumor diameter, as well as R0 resection rates (61% vs. 45%, p=0.13). No perforation was observed in either group. Postoperative bleeding was observed only in the Non-Ip group for 5 lesions (0.7%).

Conclusion

To achieve pathological complete resection for Ip adenomas, HP/EMR should be selected for adenoma with a diameter of 6 mm or greater. Conversely, CSP was considered acceptable for lesions of 5 mm or less. Further studies with a larger number of patients are warranted.

Introduction

Artificial intelligence (AI) characterization (CADx) systems for colonic polyps exhibit high diagnostic accuracy. However, they may only support, not replace, clinical decision-making and little is known of this human-AI interaction – even less so with trainee endoscopists. This issue is difficult to explore, but data from a study investigating CADx-supported trainee endoscopists compared to AI-blinded experts could give insight.

Aims & Methods

This analysis draws on data from a prospective observational study comparing the diagnostic performance of CADx-supported trainee endoscopists (using GI Genius®, Medtronic) to expert endoscopists blinded to AI support. Two key scenarios were explored: (a) instances in which the CADx system produced a ‘no prediction’ output, and (b) cases where trainees actively overruled AI-generated diagnoses. These were assessed to examine the impact of AI absence and human-AI discordance on diagnostic accuracy.

Results

The baseline cohort included 11 trainees performing 225 colonoscopies, mean age 63.8 years (SD 12.7), 48.9% male patients. Baseline ADR was 57.8%. Of 630 resected lesions (median size 3 mm, IQR 2;4), 40.0% were in the rectosigmoid, and 53.6% were non-adenomatous. CADx failed to provide a prediction in 19.2% (n=121) of cases. These ‘no prediction’ polyps were smaller (median 3 mm, p=0.048) and more often located in the right colon (72% vs. 58%, p=0.005. In the absence of AI support, trainee diagnostic accuracy declined significantly by 12.8% (63.6% vs. 76.4%, p=0.007). Interestingly, expert endoscopists—despite being blinded to AI—also performed worse on these polyps (64.7% vs. 73.9%, p=0.045), suggesting an inherent difficulty in classifying these lesions.
Trainees overruled AI suggestion leading to discordant diagnoses in 9.0% (57/630) of polyps, comprised of 20 (35.1%) non-adenomatous, and 37 (64.9%) adenomatous lesions. Trainees incorrectly overruled AI in 59.6% (34/57), which mostly consisted of downgrading lesions to NICE 1. Discordant polyps did not differ in size (3 mm vs. 3 mm, p=0.413), but were more often located proximally (77.2% vs. 58.1%, p=0.005). The difference in false negative rate (FNR) between trainees and AI was 62.16% (81.07% vs. 18.92%, p<0.0001), demonstrating a risk of missed adenomas when trainees override AI. Blinded expert endoscopists achieved identical overall diagnostic accuracy in those polyps as the AI system (59.6%). However, their error profiles differed – In this subset, blinded experts and AI performed equally in overall accuracy (59.6%), although with differing error profiles: AI showed a numerically lower FNR (18.9% vs. 35.1%, Δ16.2%, p=0.17), offset by a higher false positive rate (80.0% vs. 50.0%).

Conclusion

CADx systems generate ‘no prediction’ outputs in a substantial proportion of polyps, significantly impairing trainee diagnostic performance. These cases may represent inherently more challenging lesions. Despite CADx’s high accuracy, trainees overrule AI in almost every tenth polyp, most often incorrectly downstaging polyps. Structured training on appropriate use of CADx, emphasizing both its capabilities and limitations, may enhance clinical effectiveness. Further research on human-AI interaction and trust is warranted, as these factors will be critical for successful integration of CADx into routine endoscopic practice.

Disclosure

The original study (AC-CADx) received technical support by Medtronic. Medtronic was not involved in the design of the study or analysis of the data, neither in the original study nor this exploratory analysis.

Introduction

Barrett’s esophagus (BE) is a precursor to esophageal adenocarcinoma, with its prevalence reported to range from 1.9% to 6.8% in Western countries. We aimed to identify cost-effective pathways for BE screening by combining BE detection using multiple tools with subsequent risk-stratified surveillance.

Aims & Methods

A decision-analytic model was developed using a hypothetical cohort of the general population aged 50 years with BE prevalence rates of 1.0-7.0%. In addition to a no screening strategy, 24 potential screening pathways were evaluated combining six detection strategies (combinations of: guideline-based risk stratification, algorithmic risk stratification, swallowable devices like Cytosponge-trefoil factor 3, and endoscopy) and four surveillance strategies varying endoscopy intervals based on the BE length. Cost-effectiveness was assessed from an Australian health system perspective, using incremental cost-effectiveness ratio (ICER) thresholds for the USA (US$100,000/QALY), UK (US$43,228/QALY), and Australia (US$34,554/QALY).

Results

BE screening was cost-effective compared to no screening at BE prevalences of ≥1.0%, ≥2.4%, and ≥4.7% for the USA, UK, and Australian ICER thresholds, respectively. The optimal screening strategy was different between USA/non-USA thresholds; a 3-step detection (algorithmic risk stratification followed by swallowable devices and then endoscopy) with subsequent surveillance (no surveillance for short segment BE [SSBE] <3 cm, 2 yearly for long segment BE [LSBE] ≥3 cm; eradication for low-grade dysplasia [LGD]) for the UK and Australia. In contrast, for the USA, it was a 2-step detection (swallowable devices followed by endoscopy) with surveillance (no surveillance, SSBE ≤2 cm; 2 yearly, LSBE >2 cm; eradication, LGD) especially at ≥2.4% BE prevalence.

Conclusion

BE screening combining BE detection using multiple tools with subsequent risk-stratified surveillance based on segment length might be cost-effective, depending on BE prevalence in the community and ICER thresholds. Further work is needed to determine the actual BE prevalence in target communities.

Introduction

Among immune-related adverse events (irAEs) caused by immune checkpoint inhibitors (ICIs), gastrointestinal irAEs (GI-irAEs) are relatively rare, and their risk factors and prognostic implications remain unclear.

Aims & Methods

We retrospectively analyzed 1,104 patients who initiated ICI therapy at our institution between January 2016 and December 2022 and for whom survival data were available. GI-irAEs were defined as colitis with an increase of ≥3 bowel movements/day from baseline in the absence of other causes, or symptomatic gastritis confirmed histologically. Risk factors and prognostic impact of GI-irAEs were evaluated.

Results

The median duration of ICI therapy was 107 days. The cohort included 754 males (68%) with a median age of 69 years. The primary tumor types were lung cancer (32%), gastrointestinal cancers (19%), urothelial carcinoma (13%), head and neck cancer (13%), and melanoma (11%). The overall response rate was 27.8%, and the disease control rate was 49.0%. irAEs of any grade occurred in 310 patients (28%), and GI-irAEs were observed in 30 patients (2.7%). In univariate logistic regression analysis, use of anti-CTLA-4 antibody, combination ICI therapy, a prognostic nutritional index (PNI) ≥ 40, and platelet-to-lymphocyte ratio (PLR) ≥ 1.8 were associated with an increased risk of GI-irAEs. In multivariate analysis, only anti-CTLA-4 antibody use remained a significant risk factor (odds ratio: 9.2; 95% CI: 1.9–45). Kaplan–Meier analysis revealed that patients with GI-irAEs had significantly longer overall survival (OS) compared to those without irAEs in the overall cohort (median OS: 22.9 vs. 12.6 months, P < 0.0001). However, this difference was not observed in landmark analyses at 4 and 8 weeks post-treatment. Subgroup analysis based on PNI showed that in patients with PNI ≥ 40, those with GI-irAEs had significantly longer OS than those without irAEs (median not reached vs. 19.3 months, P = 0.0003). In contrast, in patients with PNI < 40, GI-irAE cases tended to have shorter OS than non-irAE cases (8.2 vs. 9.3 months). This trend was consistent in landmark analyses at 4 and 8 weeks.

Conclusion

The impact of GI-irAEs on prognosis appears to differ depending on the patient’s nutritional status. Particular caution is warranted for patients in poor general condition receiving ICIs, especially those involving anti-CTLA-4 antibodies, as GI-irAEs may serve as a poor prognostic factor in this population.

Introduction

Women have been historically underrepresented in gastrointestinal endoscopy research. Understanding the current patterns of female authorship is essential to assess progress and identify persistent gaps within this field.

Aims & Methods

This study aimed to evaluate the representation of women as first and senior (last) authors the three highest-impact journals in gastrointestinal endoscopy (Endoscopy, Gastrointestinal Endoscopy and Digestive Endoscopy) from 2015 to 2024. For each publication, we extracted bibliometric data of original articles, review articles and editorials, including the names of the first and senior authors, their primary country of affiliation, and the year of publication. Countries and territories were categorized by geographic region. Author gender was predicted using Genderize.io software (https://genderize.io), and only authors with ≥80% gender prediction accuracy were included.

Results

We analyzed 4,414 publications. Among first authors, 885 (20.15%) were women from 36 countries and 3,507 (79.85%) were men from 49 countries. Most female first authors were affiliated with institutions in North America (36.61%), while most male first authors were based in Asia (40.82%). The greatest gender disparity among first authors was observed in Asia (male-to-female ratio = 6.8), and the lowest in Europe (male-to-female ratio = 2.5). Among senior authors, 433 (11.51%) were women from 32 countries and 3,328 (88.49%) were men from 47 countries. Female senior authors were most frequently from Europe (38.57%), whereas male senior authors were again predominantly from Asia (41.12%). The largest disparity in senior authorship was in Oceania (male-to-female ratio = 44), and the lowest in Latin America (male-to-female ratio = 4.1). Notably, the odds of a woman being the first author nearly doubled when the senior author was also a woman. Overall, over the 10-year period, the male-to-female ratio among first authors improved from 4.46 (2015–2016) to 3.25 (2023–2024), and among senior authors from 9.10 to 6.43.

Conclusion

Women remain underrepresented in first and senior authorship positions in gastrointestinal endoscopy research. These disparities may be influenced by limited female representation and mentorship in the field. Although gender gaps have narrowed over the past decade, sustained and coordinated efforts are needed to promote equity in academic publishing.

Clinical Case Summary

Immune checkpoint inhibitors (ICI) revolutionized both treatment and outcome of multiple malignancies. However, they are frequently associated with immune-related adverse events (irAE), commonly affecting the gastrointestinal tract with an increased incidence and severity when combining agents.
A 41-year-old man presented with abdominal pain and bloody diarrhea after treatment with dual ICI therapy for lymphogenic progression after partial nephrectomy in the past. The initial work-up revealed positive stool samples for C.difficile – antibiotic treatment with negativization of stool samples was completed with no significant improvement. A colonoscopy was then performed, revealing severe ulcerative pancolitis with no ileal involvement. Based on the use of ICI 40 days prior to manifestation, the diagnosis of immune-related pancolitis was made. Local and systemic steroids at 1 mg/kg were initiated, achieving clinical remission. ICI therapy was permanently discontinued.
During steroid tapering a clinical relapse occurred – dosage was increased to 2 mg/kg with only partial improvement. A follow-up colonoscopy showed moderate pancolitis with a narrow fistula 30cm above the dentate line with an intraluminal free-moving foreign body (clips) from the nephrectomy, which was not detected on MRI. A second line therapy with three intravenous infusions of Vedolizumab was applied. A complete clinical response (resolvement of diarrhea and normalization of calprotectin levels) was achieved for three months, followed by a new relapse of symptoms with endoscopically confirmed moderate pancolitis. Considering the chronically recurrent course, the patient was put on an ongoing Vedolizumab treatment with no further relapses up to present date.
Immune-related colitis is a condition with an unpredictable natural course. Not all patients benefit from induction therapy only. Nevertheless, the duration of immunosuppressive treatment for long-term remission remains unclear at times.

References

1. Colonoscopy revealing severe ulcerative pancolitis
2. Follow up colonoscopy with a free moving foreign body
3. MRI and CT findings