Introduction

Hereditary diffuse gastric cancer (HDGC) associated with CDH1 germline pathogenic variants (PV), carries a high lifetime risk of signet ring cell carcinoma (SRCC). Clinical and molecular events driving progression from early SRCC to advanced cancer are elusive, therefore there is uncertainty to whom and when to recommend risk reducing total gastrectomy (RRTG). We hypothesise a high number of early SRCC lesions (SRCC burden) could be a risk factor for progression to clinical gastric cancer. This study aimed to identify predictors of SRCC burden and provide evidence to modify surveillance intervals.

Aims & Methods

We analyzed data from 53 CDH1 PV carriers who underwent RRTG prospectively recruited between Jan 2006 and Dec 2024 (RRTG dataset) and a separate cohort of 95 CDH1 carriers with ≥2 surveillance endoscopies and 0-2 positive biopsies at baseline (endoscopy dataset). Targeted biopsies (TB) and systematic random biopsies (RB) were obtained using the Cambridge protocol. Multivariable negative binomial regression was used to assess the association of clinical (age, sex, family history, ethnicity, CDH1 variant type) and endoscopic factors (mean number of positive TB and RB per endoscopy) with SRCC burden in RRTG specimens. A logistic regression model with significant predictors was trained to classify patients into low (≤20 foci) and high (>30 foci) cancer burden, validated over 1,000 iterations. Temporal trends in biopsy findings were analyzed using linear mixed-effects models in endoscopy dataset. Surveillance effectiveness at 6- and 12-month intervals was compared in a subset of 38 patients with at least one positive endoscopy followed by subsequent surveillance.

Results

Among 53 RRTG patients (median: 3 surveillance endoscopies over 19 months), 47 (90%) had pT1aN0M0 cancer, and 6 (11%) had pT0N0M0. The mean number of SRCC foci on RRTG specimens was 13, with a median of 33 (IQR: 2-37) and a range of 0 to 273. Among clinical factors, whereas the average number of positive TB (p = 0.001) and RB (p < 0.001) per endoscopy were independent predictors. The model based on endoscopy factors achieved an AUC of 0.88 (0.87, 0.89) for predicting cancer burden under RRTG. No significant trend in biopsy positivity over time was observed (Estimate = 0.0046, p = 0.176). Extending surveillance intervals from 6 to 12 months in patients with low-level positivity did not significantly reduce detection of increased positive biopsies (22% vs. 19%, p = 0.78). Furthermore, using ≥3 positive biopsies as a trigger for recommending RRTG, only 6 patients (16%) met this threshold. The median time after the first positive endoscopy to meeting this criterion was 9 months (95% CI: 0-29) under the 6-month surveillance strategy and 12 (95% CI: 3-33) months under the 12-month surveillance strategies.

Conclusion

Endoscopic surveillance using targeted and random biopsies by experienced endoscopists provides a reliable estimate of SRCC burden in HDGC. Our findings suggest that extending surveillance intervals and delaying surgery in patients with low early SRCC burden is clinically safe.

Introduction

Aims & Methods

Results

Conclusion

Introduction

Mitochondrial dynamics refers to coordinated cycles of mitochondrial fission and fusion that allow mitochondrial networks to re-model their shape, distribution and size. These transient processes are involved in multiple cellular functions and the perturbation of mitochondrial dynamics is reported in models of Inflammatory Bowel Diseases (IBD). The role of Dynamin-related protein 1 (Drp1), a critical GTPase for mitochondrial fission, remains understudied in an intestinal disease context.

Aims & Methods

This research aims to examine how mitochondrial networks are altered in mouse and cell culture-based models of intestinal injury and IBD. Cell culture-based assays were performed with the MODE-K intestinal epithelial cell line to explore the role of mitochondrial dynamics in vitro. A Drp1-specific knockout (Drp1 KO MODE-K) was generated using CRISPR-based approaches . Mitochondrial fission was also targeted pharmacologically using P110 and Mdivi-1 fission inhibitors. Intestinal tissue from IBD-related mouse models including TNF^ΔARE (ileitis), Il10^-/- (colitis), Hsp60 ^Δ/ΔIEC (metabolic injury) and Hsp60 ^Δ/ΔIEC;Il10^-/- (metabolic injury/colitis) mice were used for histological analyses and ex vivo organoid cultures. TOM20 was used to stain mitochondrial networks. A novel IBD mouse model with a Drp1 intestinal epithelial cell specific knockout Drp1 ^Δ/ΔIEC;Il10^-/- was generated to assess the role of Drp1-mediated mitochondrial fission in vivo.

Results

TOM20 immunofluorescence staining of mouse tissue sections show that mitochondrial morphology, localisation and mass are diverse under different inflammation and injury-based conditions. The colitis susceptible Il10^-/- mice typically show more fragmented, rounder mitochondria with a decrease in mitochondrial mass compared to Il10^+/+ controls. Hsp60^Δ/ΔIEC mice have a conditional deletion of the Heat shock protein 60 (Hsp60) in the epithelium which results in metabolic injury, a transient tissue injury with irregular tissue architecture and altered mitochondrial metabolism. This Hsp60 loss further accelerates Il10^-/- pathology in Hsp60 ^Δ/ΔIEC;Il10^-/- mice. At day 4 (D4) post tamoxifen treatment, before the onset of crypt loss, Hsp60 ^Δ/ΔIEC;Il10^-/- mice show an increase in a phosphorylated form of Drp1 (pDrp1^Ser616) indicating increased mitochondrial fission. At day 8 (D8), during peak metabolic injury, both Hsp60 ^Δ/ΔIEC;Il10^+/+ and Hsp60 ^Δ/ΔIEC;Il10^-/- mice show a significant increase in mitochondrial mass compared to controls. Furthermore, co-stimulation of ex vivo wild-type organoids with the inflammatory cytokine Tnf and the Drp1-specific inhibitor P110 increases mitochondrial unfolded protein response (UPR) markers at the mRNA level, highlighting the interplay between mitochondrial dynamics and stress responses in injury-based inflammatory conditions. Drp1 KO MODE-K cells also show a noticeable increase in dense, peri-nuclear TOM20 staining when compared to control cells. In extension, the novel Drp1 ^Δ/ΔIEC;Il10^-/- model has proved viable and preliminary data indicates that Drp1 ^Δ/ΔIEC;Il10^+/+ mice have reduced body weights compared to Drp1 ^flox/flox;Il10^+/+ mice. Further characterisation will provide insight into the interplay between Drp1, mitochondrial dynamics and intestinal disease development.

Conclusion

Mitochondrial networks are altered in response to different types of inflammation or injury in models of IBD. Regulated mitochondrial fission and fusion is important for intestinal epithelial cell homeostasis and Drp1 loss or inhibition affects intestinal epithelial cell responses.

Introduction

Several Randomized Controlled Trials (RCTs) have evaluated the role of artificial intelligence (AI) Computer-Aided Detection (CADe) systems to enhance polyp detection. However, these studies have focused exclusively on locally deployed hardware-based AI systems, whereas cloud-based systems may improve the accessibility and upgrading of CADe technology. ‘CADDIE’ (Odin Vision, an Olympus company) is a cloud-based CADe system that receives the live endoscopy image stream via the hospital’s internet network, where the AI algorithm processes the images, and overlays real-time bounding boxes on the endoscopy screen to highlight suspected polyps.

Aims & Methods

This prospective RCT aimed to evaluate the efficacy and safety of the novel cloud-based CADe-system in improving polyp detection.

This international, multi-centre, parallel-group RCT was conducted across eight centres in four European countries. Participating endoscopists had a baseline of >1000 colonoscopy procedures and an adenoma detection rate (ADR) of ≥25%. Patients aged ≥40 years and at average risk for colorectal cancer undergoing screening or surveillance indications for colonoscopy were recruited. Patients were randomised in a 1:1 ratio to standard colonoscopy (SoC) or CADDIE-assisted (CADe-arm) colonoscopy. The co-primary endpoints were superiority in adenomas per colonoscopy (APC) as the efficacy endpoint, and non-inferiority for the Positive Percent Agreement (PPA), defined as the percentage of resections that are histologically confirmed adenomas, SSLs and large (>10mm) hyperplastic polyps of the proximal colon, as the safety endpoint.

Results

Overall, 22 endoscopists and 841 participants were included in the analysis. The CADe-arm significantly increased APC by 33% (0.62 vs 0.82, Ratio 1.33 [95%CI 1.06-1.67], p=0.01) and ADR (35.9% vs 43.2%, difference of 7.3% [95%CI 0.7%-13.9%], p=0.03) compared to the SoC-arm. The mean detection of sessile serrated lesions (SSLs) (0.03 vs 0.08, Ratio 3.30 [95%CI 1.41-7.57], p<0.01), large adenomas (0.04 vs 0.09; Ratio 1.93 [95%CI 1.03-3.62], p=0.04), and large polyps (0.05 vs 0.12; Ratio 2.36 [95%CI 1.33-4.17], p<0.01) also significantly improved. A higher proportion of large polyps (27.3% vs. 37.0%) and large adenomas (15.8% vs 24.2%) were non-polypoid ‘flat’ morphology in the CADe-arm compared to the SoC-arm. Other detection metrics, including neoplastic polyps per colonoscopy (0.66 vs 0.91; Ratio 1.39 [95%CI 1.12-1.73], p<0.01) and polyps per colonoscopy (1.06 vs 1.41 (Ratio 1.35 [95%CI 1.13-1.61], p<0.01), also significantly increased.

Non-inferiority in PPA was achieved (53.4% vs 53.9%; 0.5% (-5.0%,∞)), with no increase in unnecessary resections.

Conclusion

This trial demonstrates the safety and efficacy of a cloud-based CADe-system to increase endoscopists' polyp detection across multiple metrics, notably for large adenomas, large polyps, and SSLs.

Disclosure

AR, CS, LBL, RK have received medical consultancy fees from Odin Medical Ltd

Introduction

The Mayo Endoscopic Subscore (MES) is the most widely used endoscopic index for assessing disease activity in ulcerative colitis (UC). Although artificial-intelligence　(AI)–assisted colonoscopy is expected to reduce inter-observer variability, the MES 4-point scale (0–3) lacks sufficient sensitivity to detect subtle changes in mucosal activity, limiting its ability to capture incremental improvement or deterioration. Moreover, because MES focuses on qualitative features—such as bleeding and erosions—without considering the extent or surface area of inflammation, it provides an incomplete estimate of a patient’s overall inflammatory burden.

Aims & Methods

We sought to develop and validate a quantitative, AI–driven colonoscopic score—Quantitative Ulcerative Colitis Assessment using Deep-learning (QUAD)—and to determine whether it predicts clinical relapse in patients with UC. The QUAD system was trained on 84 743 colonoscopic images from 998 UC patients at three centers and comprises three main functions: i) Image filtering: Automatic exclusion of low-quality images that are unsuitable for analysis owing to light saturation (halation), inadequate insufflation, excessive proximity to the mucosa, chromoendoscopy, or virtual chromoendoscopy. ii) Frame-level scoring: Each frame is divided into four quadrants; inflammation in each quadrant is graded from 0 to 3, and the quadrant scores are summed to yield a QUAD score of 0–12. iii) Video-level aggregation: Full-length withdrawal videos are processed to generate (i) an Entire QUAD score (0–60), representing pan-colonic inflammation, and (ii) a Distal QUAD score (0–12) for the rectosigmoid segment (final 20 % of the video).
The QUAD system contains mainly three function: i) automatically exclude low quality images that is ineligible for analysis in cases of halation, inadequate air de­livery, proximity to mucosa, chromoendoscopy, and virtual chromoendoscopy; ii) fully auto­mated twelve-class score output based on frame analysis: Each frame was divided into four quadrants; inflammation in each quadrant was graded 0–3 and summed to yield a QUAD score of 0–12; and iii) full-length withdrawal videos were processed to generate (I) an Entire QUAD (0–60) for considering whole-colon inflammation status and (II) a Distal QUAD (0–12) for the distal rectum–sigmoid segment (final 20 % of the video). An open-label cohort of adults with quiescent UC (partial Mayo score ≤ 2 for ≥6 months) was prospectively enrolled at a single center (August 2022–February 2023), and followed for 24 months after baseline colonoscopy; clinical relapse was defined as partial Mayo score > 2. The predictive value of QUAD scores for relapse was evaluated.

Results

Frame-level QUAD scores correlated strongly with expert-assigned MES values (Spearman r = 0.68, p < 0.01). Among 156 participants (median age, 51 years; 49 % male), 22 experienced relapse during follow-up. Relapse occurred in 19.4 % (19/98) of patients with a QUAD score > 3 versus 5.2 % (3/58) with a score ≤ 3 (log-rank p = 0.01).
In video analysis, an Entire QUAD > 6 did not significantly stratify relapse risk (16.8 % [17/101] vs 9.1 % [5/55]; log-rank p = 0.17). In contrast, a Distal QUAD > 3 was strongly associated with relapse (43.8 % [7/16] vs 10.7 % [15/140]; log-rank p < 0.01).

Conclusion

AI-generated QUAD provides a sensitive, fully quantitative measure of mucosal inflammation and stratifies relapse risk in clinically quiescent UC. In particular, the automatically derived Distal QUAD—focused on the distal colon and independent of operator image selection—offers a practical, objective tool for longitudinal monitoring and treatment tailoring in UC.

Introduction

While ileocecal resection (ICR) is the most common surgical procedure in Crohn's disease (CD), non-ICR surgeries account for only 20% of cases and remain underrepresented in the literature. This study aimed to identify specific clinical risk factors for surgical recurrence in CD patients undergoing small bowel or colonic resections without involving the ileocecal valve.

Aims & Methods

This multicenter retrospective cohort study included 257 patients with CD (124 small bowel resections, 133 colonic resections) across 13 Italian centers from 1978 to 2023. Univariate and multivariate logistic regression and Cox proportional hazards analyses were performed to identify clinical risk factors associated with surgical recurrence.

Results

The cohort comprised two distinct groups: small bowel resections (n=124), characterized by male predominance (63.71%), a high prevalence of stricturing disease (60.48%), and a median follow-up of 75.5 months; and colonic resections (n=133), featuring a more balanced gender distribution (47.37% male), higher rates of penetrating disease (47.37%), and a median follow-up of 88.0 months. Penetrating disease behaviour emerged as the only significant predictor of recurrence in logistic regression analysis (OR 2.16, 95% CI 1.03-4.52, p=0.041), while factors typically predictive in ICR, such as smoking (OR 0.90, p=0.843) and prior resections (OR 0.96, p=0.913), were non-significant. Cox regression identified prior biologic exposure as a significant predictor of earlier recurrence (HR 2.52, 95% CI 1.22-5.18, p=0.012), with TNF experience specifically associated with increased risk (HR 2.28, 95% CI 1.12-4.67, p=0.024). The cumulative number of risk factors (HR 1.35, 95% CI 1.01-1.82, p=0.046) and postoperative intra-abdominal septic complications (HR 2.75, 95% CI 1.18-6.44, p=0.020) also contributed to recurrence risk. Disease phenotypes significantly differed between resection types (p<0.001), with stricturing disease most prevalent in small bowel resections (60.48% vs. 36.09% in colonic) and penetrating disease more common in colonic resections (47.37% vs. 36.29% in small bowel). Results are summarized in Table 1.

Table 1: Univariate logistic regression analysis of predictors for postoperative surgical recurrence in patients undergoing small bowel or colonic resections. Abbreviations: CI: Confidence Interval; CD: Crohn’s Disease; IBD: Inflammatory Bowel Disease; POR: Postoperative recurrence; TNF: Tumour Necrosis Factors.

Conclusion

This study reveals distinct risk factors for surgical recurrence after non-ICR surgeries in CD. Unlike ICR where smoking, perianal disease, and prior resections predominate, non-ICR recurrence is primarily associated with penetrating disease phenotype and prior biologic exposure. These findings demonstrate that risk stratification models developed for ICR are not applicable to non-ICR surgeries, highlighting the need for procedure-specific approaches to postoperative management. Tailored prophylactic strategies should address the unique recurrence dynamics of small bowel and colonic resection surgeries.

Introduction

Autoimmune atrophic gastritis (AAG) is an immune-mediated disorder characterized by atrophy of the gastric corpus mucosa and preservation of the antrum. This condition is marked by the replacement of specialized oxyntic glands with pseudopyloric metaplasia (PPM) and/or intestinal metaplasia (IM) (1,2). Two main subtypes of gastric IM have been described: complete (cIM) which resembles the epithelium of the small intestine, containing Paneth cells (PanC), absorptive cells with a brush border (BB), and goblet cells (gobC), and incomplete (icIM) which exhibits histological features similar to the colonic epithelium, typically lacking PanC and/or absorptive cells (3,4). Gastric IM may progress to dysplasia (GD) and ultimately to cancer (GC), with the icIM subtype being more strongly associated with an elevated risk of neoplastic transformation (2,5).Surveillance of atrophic gastritis is particularly recommended in icIM (6,7). AAG has been reported to be associated to cIM rather than icIM(8) and also for this reason GC risk in AAG has been reported to be negligible (9).

Aims & Methods

To characterize the IM subtype in AAG patients who developed GC or GD in order to clarify the relationship between cIM or icIM and gastric neoplastic lesions in AAG.
We conducted a case-control study by performing histopathological assessment of gastric IM on formalin-fixed paraffin-embedded gastric corpus mucosa biopsies obtained from two groups:
1) cases (AAG GC+ or GD+): AAG pts (n=15, F 86%, median age 69) who developed GC (n=8) or GD (n=8) at AAG diagnosis or during long-term follow-up (FU) (median 7, range 0-13 yrs).
2) controls (AAG GC-GD-): AAG pts (n=38, F 60%, age 56 yrs) monitored for at least once 3 yrs after diagnosis who did not develop GC or GD over long-term FU (median 6 yrs, 3-17).
AAG was histologically diagnosed (1,2) and pts (cases or controls) with H.pylori (Hp) positivity at histology and/or serology and/or previous history of Hp were excluded.
Histopathological subtyping of gastric IM was performed on haematohylin-eosin-stained slides. cIM was defined by the presence intestinal metaplastic epithelium containing Paneth cells (PanC), goblet cells (gobC) and absorptive enterocytes with a brush border (BB)(9).

Results

AAG GC+GD+ (cases, n=15) and AAG GC-GD- (controls, n=37) were similar for sex, age, 1^st degree family history of GC, parietal cell autoantibodies positivity, presence of PPM and long-term FU.
The presence of PanC (13/15, 87% vs 31/37, 84%, p=0.80), gobC (15/15, 100% vs 37/37, 100%, p=1.00), and BB (14/15, 93% vs 35/37, 95%, p=0.86) was similar in AAG GC+GD+ and AAG GC-GD- pts. The presence of all 3 markers of cIM (PanC, gobC, BB) was found in 12/15 (80%) AAG GC+GD+ and 31/37 (81%) AAG GC-GD- pts (p=0.75). Also, the various combinations of these markers (PanC+gobC, PanC+BB, gobC+BB) were similar between groups: 87% vs 81%, 67% vs 81%, and 93% vs 95%, respectively (p>0.05). All pts had at least 2 of 3 markers.

Conclusion

Eighthy percent of AAG pts with GC or GD exhibited all defining histological features of cIM, including the presence of PanC, gobC, and absorptive cells with a BB, compared to 81% of AAG pts without neoplastic lesions.
In AAG with and without GC or GD, cIM is more common than icIM and gastric neoplastic lesions may occur in the absence of icIM. Therefore, icIM is not reliable as predictor for neoplastic risk and further studies are needed to identify new markers of neoplastic progression in AAG pts.

References

1) Lenti MV et al. Autoimmune gastritis. Nat Rev Dis Primers 2020
2) Rugge M et al. RE.GA.IN: the Real-world Gastritis Initiative-updating the updates. Gut 2024
3) El Zimaity HMT et al. Gastric intestinal metaplasia: subtypes and natural history. J Clin Pathol 2001
4) Sugano K et al. Gastric Intestinal Metaplasia: Real Culprit or Innocent Bystander as a Precancerous Condition for Gastric Cancer? Gastroenterology 2023
5) Du S et al. Gastric Cancer Risk of Intestinal Metaplasia Subtypes: A Systematic Review and Meta-Analysis of Cohort Studies. Clinical and Translational Gastroenterology 2021
6) Shah SC et al. AGA Clinical Practice Update on the Diagnosis and Management of Atrophic Gastritis: Expert Review. Gastroenterology. 2021
7) Dinis-Ribeiro M et al. Management of epithelial precancerous conditions and early neoplasia of the stomach (MAPS III): European Society of Gastrointestinal Endoscopy (ESGE), European Helicobacter and Microbiota Study Group (EHMSG) and European Society of Pathology (ESP) Guideline update 2025. Endoscopy 2025
8) Genta R et al. Incomplete Intestinal Metaplasia Is Rare in Autoimmune Gastritis. Dig Dis 2023
9) Rugge M et al. Autoimmune gastritis: long-term natural history in naïve Helicobacter pylori-negative patients. Gut 2023