Introduction

Endoscopic submucosal dissection (ESD) is widely recognized as the preferred technique for removing large benign colorectal lesions and those with a risk of submucosal invasive cancer (SMIC), with the goal of achieving en-bloc resection. R0 resection is crucial for enabling organ-preservation strategies, particularly in endoscopically resected T1-colorectal cancer (T1-CRC).

Aims & Methods

This study aims to evaluate R0 resection rates in a large cohort of colorectal ESDs by assessing vertical and lateral margins and identifying predictive factors for vertical margin involvement. Furthermore, the presence of other prognostic risk factors for lymph node metastasis (LNM) will be analyzed in this large cohort of T1-CRC. This is a multicenter prospective cohort study in 13 centers, including all colorectal ESDs between 09/2019 and 09/2024. Procedural and histological outcomes were analyzed using univariate and multivariate analysis.

Results

7495 large colorectal polyps were included, of which 781 were T1-CRCs with a median size of 45 [35-65] mm. En bloc resection occurred in 96.7% and R0 resection in 74.9%. Lateral resection margins were free in 92.5%, with positive margins for adenocarcinoma in only 0.8% (low grade dysplasia in 2.6%, high grade in 1.2%, missing data in 2.6%). Vertical margins were free in 80.3%, and decreased significantly when deep SMIC was present to 75.0%, compared to 95.7% for superficial SMIC. However deep SMIC with up to 2000 µm invasion depth, demonstrated free vertical resection margins in still 80.7%, after which it gradually decreased to 73.1% and 67.7% at an infiltration depth of 2000-3000 µm and >3000 µm respectively. Predictive factors for free vertical resection margins include lesion size >50mm (OR=1.66, [1.01;2.71], p=0.045) and rectal location (OR=1.86, [1.04;3.3], p=0.035). Significant difference was observed in the presence of prognostic risk factors for LNM between superficial and deep SMIC, with lympho-vascular invasion (LVI) being present in 9.6% and high-grade tumor budding (Bd) in 17.1% in superficial SMIC, compared to 25.0% and 30.0% respectively, in deep SMIC. However, absence of LVI, Bd or poor differentiation was noted in 75.9% of superficial and 61.7% of deep T1-CRC. Concerning deep SMIC, no relation was observed in the presence of these risk factors with respect to the depth of infiltration.

Conclusion

Following ESD for T1-CRC, the likelihood of achieving free vertical resection margins decreases significantly when invasion depth exceeds 2000 µm. Careful lesion selection through optical diagnosis is therefore critical. To push organ preservation, intermuscular dissection should be considered for T1-CRC with deeper invasion in the rectum. Achieving R0 resection remains paramount, as additional prognostic factors are present in approximately 38% of cases, irrespective of invasion depth.

Introduction

Colitis-associated carcinoma (CAC) is a serious complication of long-standing ulcerative colitis (UC) and Crohn’s disease (CD). While the mutational landscape of CAC has been extensively characterised, its transcriptomic landscape remains incompletely defined. To investigate the tumour ecology and molecular mechanisms underpinning CAC development, we utilised FFPE tissues from human CAC patients and colon organoids. Through the integration of immunology-focused bulk gene expression profiling and spatial transcriptomics, we derived a spatially resolved gene signature of CAC at single-cell resolution.

Aims & Methods

To comprehensively profile the CAC tumour immune microenvironment, we generated a targeted bulk RNA expression dataset (Nanostring Immunology v2) from FFPE samples of 40 patients with IBD or CAC, and inflammation-free healthy controls (10 per group). We observed robust upregulation of SPP1 (encoding osteopontin, OPN) and FN1 (Fibronectin-1) in CAC. To expand these findings, we applied spatial transcriptomics (Xenium) on 5μm FFPE sections from CAC (10 patients), UC (3 patients), and healthy controls (3 patients), enabling spatial and single-cell resolution of SPP1 and its receptor expression. We further mapped ligand–receptor interactions to specific cell types, revealing distinct SPP1-mediated signalling in CAC. These results were validated at the protein level using whole-section FFPE confocal microscopy.

Results

Immunology gene expression profiling (Nanostring) revealed strong upregulation of SPP1 (osteopontin, OPN) in colitis-associated cancer (CAC) compared to ulcerative colitis (UC) and Crohn’s disease (CD). Spatial transcriptomics (Xenium) demonstrated spatial heterogeneity of SPP1, primarily arising from CD68+ macrophages in CAC, as validated by OPN/CD68 co-immunostaining. Ligand-receptor analysis identified SPP1 as a key outgoing signal from macrophages, engaging CD44 and integrin receptors on T cells. In immune subsets, CD4+ T cells showed upregulation of FOXP3 and CTLA4, indicating a regulatory T cell phenotype enriched in CAC. CD8A+ T cells displayed upregulation of GZMK, alongside downregulated JAK-STAT signalling and upregulated LDHA, hypoxia and glycolysis hallmark pathways, suggesting metabolic reprogramming and impaired effector function in the hypoxic CAC microenvironment.

Conclusion

We identified a CAC-specific upregulation of SPP1 at both RNA and protein levels, expressed by CD68 macrophages. Spatially, SPP1/OPN+ macrophages and CD8+ T cells were mutually exclusive, suggesting a role for OPN in driving T cell exclusion and immunosuppression. Ligand-receptor analysis highlighted SPP1-CD44/integrin interactions in immune modulation. CD4+ T cells showed a regulatory phenotype, while CD8+ T cells showed metabolic reprogramming and functional impairment. These features define a CAC-specific immune landscape and a gene signature with potential to detect pre-malignant changes.

Introduction

Crohn ‘s disease (CD) is characterized by chronic inflammation in the gut, where severe complications such as fibrotic strictures require surgical resection. Stricture development involves excessive extracellular matrix (ECM) deposition due to continuous activation of intestinal myofibroblasts, that further contribute to intestinal fibrogenesis. Emerging data suggests that stricture-related intestinal myofibroblasts overexpress a serine protease called fibroblast activation protein (FAP). Furthermore, type III collagen deposition is increased during fibrosis in all layers of the intestinal tract, and studies have shown elevated collagen degradation in serum from patients with CD. The present study aimed at evaluating the potential of FAP-cleaved type III collagen [C3F] as a serum marker for intestinal fibrosis in CD.

Aims & Methods

The cohort consisted of 49 patients with luminal CD and 62 patients with stenotic CD. Clinical assessment was conducted for all patients at baseline and 12 months post study initiation. For the luminal phenotype, endoscopy was performed at both time points. For the stenotic group, radiographical assessment was done prior to surgery and endoscopy was performed at 12 months. C3F was measured in serum at all timepoints for baseline, post-operative day (POD), 1, 3, 6 and 12 months. To compare C3F levels at baseline for luminal vs stenotic phenotype, Mann-Whitney U-test was conducted. For the stenotic group, mixed-effect analysis, with a Dunnett’s multiple comparison applying Bonferroni Correction, was conducted to compare C3F levels for baseline vs POD and Spearman’s rank correlation was applied.

Results

C3F baseline levels were significantly higher in patients with stenotic CD compared to luminal CD (Table 1). In the stenotic group, C3F was significantly increased in patients at POD compared to baseline (Table 1). In the luminal group, C3F was significantly elevated in patients at 12 months compared to baseline. Furthermore, C3F showed a positive correlation with C-reactive protein (CRP) (ρ=0.350, p=0.006) and a significant negative correlation with disease duration (ρ= -0.257, p=0.048) for patients with the stenotic CD phenotype (Table 1).


Table 1. C3F levels are elevated in the stenotic CD vs the luminal CD group, with increased levels at POD compared to baseline for the stenotic phenotype. Spearman’s rank correlation, depicting the relationship between C3F and routine blood demographics, as well as demographics. Significant findings (p<0.05) are highlighted in bold.

Conclusion

C3F was elevated in patients with stenotic CD, with even higher levels after the surgical resection. The data suggests that C3F reflects fibroblast activity during fibrostenosis and tissue remodeling following surgery. The positive correlation with CRP supports the association with inflammation leading to collagen deposition and fibrosis. Interestingly, C3F shows negative correlation with disease duration suggesting that C3F could reflect fibrosis at early stages of stenotic CD. These findings highlight the potential use of C3F as a novel biomarker for intestinal fibrosis in CD.

Disclosure

TTK, MSA, MP are employed at Nordic Bioscience. ACBJ, JHM and MAK are employees and stock owners at Nordic Bioscience.

Introduction

The Mayo Endoscopic Subscore (MES) is widely used in clinical trials and daily practice to assess disease activity and response to treatment in UC.

Aims & Methods

The objective of this study was to identify any nuances in patients with MES > 1 treated with non-TNF biologics and targeted oral therapies. The Ipsos UC Therapy Monitor, a multi-centre online medical chart review study of patients with UC, was conducted between July and September 2024 among gastroenterologists practicing across hospital and private practices in the UK, France, Germany, Italy, and Spain. Physicians were screened for practice duration and patient volume. Charts of patients prescribed advanced therapy were included in the analysis.

Results

205 sampled physicians collectively reported 820 UC patients. Of the reported patients with MES > 1, 62 were treated with non-TNF biologics and 62 were treated with novel oral therapies. Among the patients treated with non-TNF biologics, a higher proportion were female vs those receiving oral therapy (48% vs 31%); they were also older in age (mean years: 43 vs 38).
The patients treated with non-TNF biologics had a higher prevalence of comorbidities, including depression/anxiety (24% vs 19%), cardiovascular disease (8% vs 0%), and dyslipidemia/hyperlipidemia/hypercholesterolemia (11% vs 7%), compared to those treated with novel oral therapy.
Mean CRP levels were higher in the oral therapy group (10.2 mg/L vs 9.1 mg/L), while mean Mayo scores were slightly lower (4.7 vs 5.2). Fecal calprotectin levels were elevated in patients treated with non-TNF biologics vs oral therapy group (441 µg/g vs 503 µg/g).
When analysing reasons for prescribing selected advanced therapy between the two patient groups, a greater proportion of patients treated with non-TNF biologics had safety reasons specified more frequently by the sampled physicians.

Table 1: % of reported UC patients receiving advanced therapy for specific cited reasons

Conclusion

This study highlights key differences in patient characteristics and treatment considerations among UC patients with MES > 1 treated with non-TNF biologics compared to those treated with targeted oral therapies. Patients on non-TNF biologics tend to be older, more often female, and have a higher prevalence of comorbidities. Physicians prescribing non-TNF biologics prioritize safety, while those prescribing oral therapies focus on rapid symptom control. These findings highlight the importance of considering patient characteristics and treatment goals when selecting advanced therapies for UC patients with MES > 1. Further investigation using comparator cohort is warranted.

Introduction

The pathogenesis of ulcerative colitis (UC) involves complex immune responses, including Th1, Th2, and Th17 pathways, and is characterized by substantial heterogeneity. Cytokine profiling may contribute to understanding individual disease mechanisms and guiding personalized therapeutic strategies; however, its clinical application remains limited due to technical complexity. This study aimed to investigate whether endocytoscopic and histopathological findings correlate with mucosal cytokine expression levels and could serve as surrogate markers.

Aims & Methods

Thirty-five patients with UC receiving treatment at our institution were enrolled. Endocytoscopy (ultra-magnifying endoscopy) was performed in all patients, and the findings were classified according to the Endocytoscopic Classification system (EC classification: EC-A, EC-B, EC-C, EC-D). Mucosal biopsies were obtained from the most inflamed areas identified during endocytoscopy. Histological evaluation included the Matts histological grade and assessments of neutrophil, eosinophil, and lymphocyte infiltration; goblet cell depletion; crypt distortion; and epithelial damage, each graded on a four-point scale (absent, mild, moderate, severe). mRNA expression levels of cytokines (IFN-γ, TGF-β1, TNF-α, IL-1β, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12, IL-13, IL-17, and IL-23) were quantified using RT-PCR. Statistical analyses were performed to assess correlations between endocytoscopic/histopathological findings and cytokine expression levels.

Results

The median age of participants was 44 years (IQR: 35–57), with 24 males and 11 females. Disease types included pancolitis (n = 16), left-sided colitis (n = 14), and proctitis (n = 5). The median Mayo score was 1 (IQR: 0–5). EC classification scores showed significant positive correlations with the expression levels of IL-1β, IL-6, IL-8, IL-13, IL-17, and IL-23 (Spearman’s rank correlation, p < 0.05; r = 0.555, 0.526, 0.584, 0.441, 0.422, and 0.449, respectively). Similarly, Matts histological grades were significantly correlated with the expression of IFN-γ, IL-1β, IL-6, IL-8, IL-13, IL-17, and IL-23 (p < 0.05; r = 0.493, 0.756, 0.692, 0.730, 0.656, 0.623, and 0.448, respectively). Histopathological features—including neutrophil and lymphocyte infiltration, goblet cell depletion, epithelial damage, and crypt distortion—were significantly associated with the expression of IFN-γ, IL-1β, IL-6, IL-8, IL-13, IL-17, and IL-23. Additionally, eosinophil infiltration was positively correlated with the expression levels of IL-1β, IL-5, and IL-13.

Conclusion

This study demonstrates that endocytoscopic and histopathological indicators of inflammation are associated with mucosal expression levels of several cytokines—particularly inflammatory and Th17-related cytokines—in patients with ulcerative colitis (UC). These findings suggest that such evaluations may serve as surrogate markers, potentially improving disease assessment and informing therapeutic decision-making.

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

Crohn’s disease (CD) is a human immune-mediated inflammatory disease (IMID) of unknown etiology that affects the gastrointestinal (GI) tract. CD is now a global health issue due to its lifelong nature and significant burden on patients and public health [1]. Phase III clinical trials have shown that treatment with oral Janus Kinase (JAK) inhibitors can alleviate disease in patients with CD. However, the United States (US) Food and Drug Administration (FDA) has issued warnings about side effects for this class of oral drugs which target JAKs systemically in the human body. Therefore, new targeted delivery approaches are needed to block JAK activity selectively in the GI tract. The JAK1/2-STAT1 intracellular signaling pathway is activated by the Th1 cytokine IFN-γ alone or in synergy with another Th1 cytokine, TNF-α, and SOSC1 is an endogenous negative regulator of JAK1/2. This pathway activates the expression of inflammatory genes in multiple cell types typically associated with active or inflamed CD [2]. RNA interference (RNAi) mediated gene silencing using small interfering RNA (siRNA) targeting JAK1/2-STAT1 and SOCS1 overexpression using exogenous SOCS1 mRNA are promising methods to reduce the JAK-STAT pathway activation.

Aims & Methods

The GENEGUT project aims to develop oral RNA-based therapies for ileal CD. In this study, we have developed new approaches to deliver siRNAs targeting JAK1, JAK2, and STAT1 into relevant cell models, including human intestinal epithelial cell (IEC) lines. Epithelial cell models that replicate the inflammatory microenvironment in CD were used by exposing IECs to inflammatory cytokines IFN- γ/TNF- α, leading to increased cytokine production and cell death.

Results

Selected siRNAs effectively blocked JAK1, JAK2 and STAT1 expression and activity, preventing the downstream activation of inflammatory and cell death pathways. SOCS1 overexpression also suppressed JAK1/2-STAT1 pathway activation and enhanced cell survival.

Conclusion

Ongoing work is focused on optimizing the delivery of siRNA and mRNA into differentiated epithelial monolayers, co-cultures and patient-derived organoids to advance translational applications.

References

1. Dulai PS, Singh S, Vande Casteele N, Boland BS, Rivera-Nieves J, Ernst PB, Eckmann L, Barrett KE, Chang JT, Sandborn WJ. Should We Divide Crohn's Disease Into Ileum-Dominant and Isolated Colonic Diseases? Clin Gastroenterol Hepatol. 2019 Dec;17(13):2634-2643. doi: 10.1016/j.cgh.2019.04.040. Epub 2019 Apr 19. PMID: 31009791; PMCID: PMC6885453.
2. Cordes F, Foell D, Ding JN, Varga G, Bettenworth D. Differential regulation of JAK/STAT-signaling in patients with ulcerative colitis and Crohn's disease. World J Gastroenterol. 2020 Jul 28;26(28):4055-4075. doi: 10.3748/wjg.v26.i28.4055. PMID: 32821070; PMCID: PMC7403801.

Disclosure

Funded by the European Union (GA 101057491) and supported the Swiss State Secretariat for Education, Research and Innovation (SERI) under contract number 22.00119. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Health and Digital Executive Agency (HaDEA). Neither the European Union nor the granting authority can be held responsible for them. UK participants in Horizon Europe Project GENEGUT are supported by UKRI grant number 10042451 (Bangor University).