Introduction

Micronutrient deficiency is an underestimated problem in the westernized population with unclear effects on the manifestation of acute or chronic disease. Westernized lifestyle instigates disrupted gut-microbial crosstalk and chronic low-grade inflammation¹. Tryptophan (Trp) is an essential amino acid used for de novo synthesis of cellular nicotinamide adenine dinucleotide (NAD). We have previously shown increased Trp degradation in several chronic inflammatory disorders, including IBD^2,3. Whether Trp deficiency is a widespread phenomenon in the western population is unknown. Here, we tested the hypothesis that Trp deficiency is frequent in the western population and associates with increased overall disease severity in internal-medicine and especially gastrointestinal (GI) diseases.

Aims & Methods

To screen the micronutrient status of the healthy population, we assessed quantitative amino acid levels in n=15.188 healthy controls (HC) with a commercially available home-care dry blood spot test for targeted metabolomics (cohort 1). Self-reported symptoms were assessed with a digital questionnaire. To validate Trp deficiency (defined as the lowest Trp quartile) in acute internal-medicine diseases (cohort 2), we measured serum Trp by HPLC in HC (n=291) and patients presenting in the emergency ward of the University Hospital Schleswig-Holstein, Kiel (n=4349, categorized by ICD10; thereof GI diseases n=516 (K00-K93: diseases of the digestive system); 10/2023-09/2024). Next to laboratory and clinical assessment at presentation, disease courses were followed up digitally for at least 6 months. Statistical analyses were performed with linear mixed models (LMM) and logistic regression correcting for age and sex at birth; predictive performance was evaluated via area under the receiver-operating-characteristic (AUROC; R version 4.4.1).

Results

In cohort 1, 39.3% of HC presented with Trp deficiency. Trp deficiency significantly inversely correlated with the amount of self-reported symptoms, such as e.g. GI symptoms. Validating these findings in internal-medicine and a subgroup of GI diseases in cohort 2, we found significantly reduced serum Trp compared to HC (mean Trp 56.58±10.84µM vs. 45.06±16.65 (all disease groups), 45.21±20.69 (GI diseases)), especially when the serum inflammation marker C-reactive protein was not elevated (<5 mg/l). Here, Trp deficiency was observed in 1.4% of HC, while it was strongly increased in all disease groups (26.6%) and GI diseases (30.2%). Using LMM, we found that patients that were hospitalized, admitted to the ICU or died within the follow-up period had a significantly lower Trp compared to patients with favourable outcomes (outpatient/no ICU/alive). Patients with Trp at the lowest compared to the highest quartile underwent longer hospitalisation and ICU treatment (all diseases: 3.17/0.37 days (d) longer; GI diseases: 3.91/0.48d longer) as well as a higher number of in-hospital procedures (all diseases: 1.47 more; GI diseases: 3.3 more). The predictive value of Trp for complicated outcomes (hospitalisation, hospitalisation >7d, ICU treatment, death) as assessed by AUROC was low-moderate (all diseases: AUC 0.69, 0.67, 0.61, 0.72; GI diseases: AUC 0.61, 0.6, 0.54 and 0.62, respectively).

Conclusion

We report that Trp deficiency is frequent in the westernized population and associates with more severe disease courses. Our data suggests Trp degradation along the de novo NAD synthesis pathway as an overarching metabolic hallmark of chronic low-grade inflammation and a risk factor for severe disease courses of internal-medicine and GI diseases.

References

1. Christ, A. & Latz, E. The Western lifestyle has lasting effects on metaflammation. Nature Reviews Immunology vol. 19 267–268 Preprint at https://doi.org/10.1038/s41577-019-0156-1 (2019).
2. Harris, D. M. M. et al. Tryptophan degradation as a systems phenomenon in inflammation – an analysis across 13 chronic inflammatory diseases. EBioMedicine 102, 105056 (2024).
3. Nikolaus, S. et al. Increased Tryptophan Metabolism Is Associated With Activity of Inflammatory Bowel Diseases. Gastroenterology 153, (2017).

Introduction

Artificial Intelligence (AI)- enabled endoscopy and histology offer accurate, objective, and rapid assessment of disease activity in Ulcerative Colitis (UC). Emerging multi-source AI models integrating diverse datasets may enhance standardised disease evaluation and outcome prediction.

Aims & Methods

This study aimed to develop a novel AI model fusing endoscopic and histological findings to improve the assessment of disease remission and predict early response to therapy in UC clinical trials. A novel multimodal AI fusion algorithm was developed by integrating paired endoscopic videos and histological whole-slide images (WSIs) from the phase 2 clinical trial of Mirikizumab in UC (NCT02589665). The endoscopy branch of the model was trained with 291 white-light videos, using a convolutional neural network to select informative frames and the BioMedCLIP foundational model to extract features. The histology branch utilised the CONCH foundational model and was fine-tuned on 291 WSIs to obtain patch-level features. Features from both modalities were aggregated and fused using multi-head self-attention. The model’s ability to assess histological remission and response to therapy at weeks 12 and 52 was evaluated. Histological remission was defined as Geboes ≤2B.0, while a response to therapy at different time points was based on histological remission or improvement (Geboes <3.1).

Results

The fusion model outperformed single-modality assessments for histological remission, achieving a sensitivity of 89.72% (95% CI: 82.35–94.76), specificity of 89.67% (95% CI: 84.34–93.67), and accuracy of 89.69% (95% CI: 85.61–92.94). It demonstrated remarkable performance in assessing response to therapy at 12 and 52 weeks, with sensitivity of 97.96% (95% CI: 89.15–99.95), specificity of 86.84% (95% CI: 71.91–95.59) and accuracy of 93.10% (95% CI 85.59 – 97.43) for histological remission at week 52. Substantial agreement was observed between the AI fusion model and central readout.

	Histologic remission			Response to therapy
	Endoscopy-only	Histology-only	Endo-Histo fusion	Histologic improvement week 12	Histologic improvement week 52	Histologic remission week 12	Histologic remission week 52
Sensitivity (95% CI)	75.80% (66.46 – 83.47)	85.98% (77.93 – 91.94)	89.72% (82.35 – 94.76)	86.54% (74.21 – 94.41)	93.88% (88.13 – 98.72)	79.59% (65.66 – 89.76)	97.96% (89.15 – 99.95)
Specificity (95% CI)	73.91% (66.94 - 80.10)	88.59% (83.08 – 92.79)	89.67% (84.34 – 93.67)	76.00% (61.83 – 86.94)	81.58% (65.67 – 92.26)	73.58% (59.67 – 84.74)	86.84% (71.91 – 95.59)
PPV (95% CI)	62.79% (53.84 - 71.14)	81.42% (73.01 – 88.11)	83.48% (75.41 – 89.75)	78.95% (66.11 – 88.62)	86.79% (74.66 – 94.52)	73.58% (59.67 – 84.74)	90.57% (79.34 – 96.87)
NPV (95% CI)	83.95% (77.37 – 89.24)	91.57% (86.48 – 95.21)	93.75% (89.09 – 96.84)	84.44% (70.54 – 93.51)	91.18% (76.32 – 98.14)	79.59% (65.66 – 89.76)	97.06% (84.67 – 99.93)
Accuracy (95% CI)	74.57% (69.16 – 79.47)	87.63% (83.29 – 91.18)	89.69% (85.61 – 92.94)	81.37% (72.45 – 88.40)	88.51% (79.88 – 94.35)	76.47% (67.04 – 84.31)	93.10% (85.59 – 97.43)
F1 Score (95% CI)	68.64% (62.31 – 74.51)	83.64% (78.07 – 88.27)	86.49% (81.27 – 90.69)	82.57% (74.13 – 89.17)	90.20% (82.71 – 95.20)	76.47% (67.04 – 84.31)	94.12% (87.64 – 97.81)

Abbreviations: CI, confidence interval; NPV, negative predictive value; PPV, positive predictive value.

Conclusion

This innovative multimodal fusion AI model enhances the assessment of histological remission and accurately predicts response to therapy. By potentially standardising central readouts and enabling automated disease assessment, this novel tool marks a significant advancement towards precision medicine in clinical trials.

Disclosure

Marietta Iacucci received research grants and consultant fees from Olympus, Pentax, Eli Lilly, and Janssen. All other authors declare no competing interests.

Introduction

Helicobacter pylori (H. pylori) infection is a major cause of gastric diseases, with the virulence factor CagA carried by exosomes playing a crucial role in regulating inflammation. Ferroptosis and M1 polarization are key mechanisms in promoting inflammation, yet the role of CagA in inducing macrophage ferroptosis and M1 polarization to exacerbate gastric mucosal inflammation remains unclear.

Aims & Methods

CagA-positive (CagA+) and CagA-negative (CagA-) H. pylori strains were co-cultured with gastric epithelial cells, and their supernatants were collected. Exosomes were isolated using ultracentrifugation to obtain CagA+Exo and CagA-Exo. Changes in macrophage M1/M2 polarization markers and ferroptosis markers were analyzed under CagA+Exo and CagA-Exo treatment. The activation of the JAK/STAT1 signaling pathway was assessed by Western blot, and STAT1 expression in the nucleus was evaluated using a nuclear protein extraction kit. Chromatin immunoprecipitation (ChIP) was used to identify the binding site of STAT1 on the SLC7A11 promoter. Biopsy samples from the gastric mucosa of gastritis patients infected with CagA+ or CagA- H. pylori were analyzed for macrophage polarization and ferroptosis markers.

Results

Exosomal CagA was found to induce macrophage M1 polarization and ferroptosis, with the ferroptosis inhibitor SLC7A11 playing a key role in CagA-induced ferroptosis. Exosomal CagA activated the JAK/STAT1 signaling pathway, leading to M1 polarization and ferroptosis. Additionally, CagA+Exo increased nuclear STAT1 expression, which subsequently bound to the SLC7A11 promoter to regulate its transcription. In gastric mucosa biopsies from CagA+ H. pylori-infected gastritis patients, increased levels of iNOS (a marker of M1 polarization) and decreased expression of SLC7A11 were observed.

Conclusion

Our findings indicate that exosomal CagA promotes macrophage M1 polarization and ferroptosis through the JAK/STAT1 signaling pathway. This study provides new insights and potential therapeutic targets for H. pylori-associated gastritis.

Introduction

Deep learning systems in gastrointestinal endoscopy are traditionally pretrained on large, diverse datasets comprising generic, non-medical images before being trained on scarce endoscopic data. Foundation models are models pretrained based on domain-specific datasets. For endoscopy, foundation models require large datasets of general endoscopic images. Yet datasets for developing such models remain limited. In this study, we present GastroNet-5M, a dataset comprising 4,820,653 endoscopic images of approximately 500,000 procedures.

Aims & Methods

GastroNet-5M is composed of anonymized endoscopic images collected from eight Dutch hospitals between 2012 and 2020. Using a self-supervised learning approach, this dataset enabled the development of a foundation model for various endoscopic AI applications. In this study, we compared GastroNet-5M-pretrained models against current state-of-the-art models with traditional pretraining approaches and all publicly available endoscopic foundation models.
We evaluated diagnostic performance on 9 classification and 9 segmentation tasks, including Barrett’s neoplasia detection, colorectal polyp characterization, and gastric cancer invasion depth prediction. All models were trained from scratch for each task. Data efficiency was assessed by repeating all classification tasks with progressively smaller training sets (i.e. 100%, 50%, 25%, 10% and 5%). Robustness was evaluated across 12 datasets reflecting data heterogeneity such as variation in endoscope manufacturer, imaging modality (e.g., virtual chromoendoscopy), and image quality.

Results

GastroNet-5M-pretrained models outperformed all benchmarks in 8 of 9 classification tasks, with an average AUC improvement of 6.8% (Table 1). For segmentation, they achieved higher Dice scores compared to all other models across all 9 tasks, with an average gain of 15.4%. In data efficiency experiments, GastroNet-5M models achieved the highest AUC in 5 of 9 tasks when training data was limited, while no other model led in more than one. In robustness testing, they outperformed other models in 8 of 12 tasks evaluating heterogeneous data conditions.

Table 1. Area-Under-the-Curve scores of the GastroNet-5M-pretrained model and benchmark models for all classification tasks.

Dataset	Application	GastroNet-5M FM	Sanderson et al.1	Van der Putten et al.2	EndoFM3	ConvNeXt Base4	DINOv2 ViT-Base (ImageNet)5,6	YOLOv117
ARGOS8	Barrett's Neoplasia Detection	91.6 (87.1-96.1)	89.9 (86.4-93.4)	72.6 (69.4-75.9)	84.9 (83.7-86.1)	89.2 (87.3-91.0)	82.2 (77.6-86.8)	85.2 (83.9-86.5)
BONSAI CADe9	Barrett's Neoplasia Detection	96.7 (96.6-96.9)	90.5 (90.2-90.7)	86.1 (85.0-87.2)	90.4 (90.0-90.9)	96.7 (96.5-96.9)	96.5 (96.3-96.7)	91.8 (90.7-92.9)
BONSAI CADx10	Barrett's Neoplasia Characterization	97.2 (96.9-97.6)	92.1 (91.7-92.4)	93.1 (92.4-93.9)	94.4 (94.0-94.7)	96.4 (96.0-96.8)	96.8 (96.1-97.5)	96.2 (95.7-96.6)
GIANA11	Angiodysplasia Detection	99.5 (99.3-99.6)	91.7 (91.0-92.3)	94.7 (94.0-95.4)	98.7 (98.7-98.8)	99.6 (99.5-99.7)	99.6 (99.6-99.7)	99.4 (99.2-99.6)
HyperKvasir-MES12	IBD Disease Activity	88.9 (88.7-89.1)	65.2 (61.8-68.7)	79.0 (77.4-80.6)	85.0 (84.2-85.8)	84.5 (83.9-85.2)	70.8 (61.2-80.4)	83.5 (81.8-85.2)
LIMUC13	IBD Disease Activity	95.5 (95.2-95.7)	92.7 (92.6-92.8)	92.1 (91.6-92.7)	94.2 (94.1-94.4)	94.3 (94.2-94.5)	94.5 (94.3-94.8)	93.3 (92.7-93.8)
Maastricht14	Colorectal Polyp Diagnosis	92.8 (91.4-94.3)	73.6 (70.8-76.5)	76.9 (73.5-80.3)	90.5 (90.0-91.1)	89.9 (89.2-90.7)	81.0 (69.8-92.3)	89.3 (88.1-90.5)
NJDTH15	Gastric Cancer Invasion Depth	95.0 (94.1-95.8)	83.3 (82.3-84.3)	79.6 (78.3-80.9)	88.3 (87.8-88.8)	89.4 (87.8-90.9)	91.0 (88.4-93.6)	90.5 (89.2-91.8)
POLAR16	Colorectal Polyp Diagnosis	85.2 (84.3-86.1)	67.6 (67.2-68.1)	64.0 (62.2-65.9)	71.0 (70.2-71.8)	77.5 (75.9-79.1)	67.4 (59.3-75.5)	77.2 (75.2-79.3)

Conclusion

GastroNet-5M, a multicenter dataset of over 5 million unlabeled endoscopic images, offers a valuable resource for pretraining deep learning models in endoscopy. The use of GastroNet-5M enhances diagnostic accuracy, reduces the need for scarce application-specific endoscopic imagery, and improves robustness of endoscopic AI systems to the inevitable data heterogeneity in clinical practice. GastroNet-5M is publicly accessible for further research and development.

References

1. Sanderson E, Matuszewski BJ. arXiv preprint arXiv:2401.06278. 2024.
2. van der Putten J, de Groof J, Struyvenberg M, et al. Artif Intell Med. 2020;107.
3. Wang Z, Liu C, Zhang S, et al. arXiv preprint arXiv:2306.16741. 2023.
4. Liu Z, Mao H, Wu C-Y, et al. arXiv preprint arXiv:2201.03545. 2022.
5. Dosovitskiy A, Beyer L, Kolesnikov A, et al. arXiv preprint arXiv:2010.11929. 2020.
6. Oquab M, Darcet T, Moutakanni T, et al. arXiv preprint arXiv:2304.07193. 2023.
7. Jocher G, Qiu J. Ultralytics YOLOv11, v11.0.0. 2024.
8. de Groof AJ, Struyvenberg MR, van der Putten J, et al. Gastroenterology. 2020;158(4):915–29.e4.
9. Fockens KN, Jukema JB, Boers T, et al. United European Gastroenterol J. 2023;11(4):324–36.
10. Jukema J, Kusters C, Jong M, et al. Endoscopy (ESGE Days). 2023.
11. GIANA Challenge. Available from: https://giana.grand-challenge.org
12. Borgli H, Thambawita V, Smedsrud PH, et al. Sci Data. 2020;7(1):283.
13. Polat G, Kani HT, Ergenc I, et al. Inflamm Bowel Dis. 2023;29(9):1431–9.
14. van der Zander QEW, Schreuder RM, Fonollà R, et al. Endoscopy. 2021.
15. Tang D, Zhou J, Wang L, et al. Front Oncol. 2021;11:654093.
16. Houwen BB, Hazewinkel Y, Giotis I, et al. Endoscopy. 2022;54(Suppl 1):eP133.

Introduction

Gastric mucosal visibility is essential for high-quality upper gastrointestinal (GI) endoscopy. While simethicone and N-acetylcysteine (NAC) have been shown to improve mucosal visibility, the optimal time interval between preparation and endoscopy remains undefined.

Aims & Methods

This randomized controlled trial evaluated the effect of timing on mucosal visibility and explored the upper limit beyond which visibility begins to decline.
In this single-center, randomized controlled trial, adult patients undergoing elective upper GI endoscopy were randomized into four groups: placebo, simethicone + NAC with endoscopy performed after 10–20 minutes, 20–30 minutes, and >30 minutes. The primary outcome was the proportion of patients achieving adequate gastric mucosal visibility (GMV), defined as a score of <7 on a validated visibility scale. A pre-specified subgroup analysis of the >30 minute group was performed by stratifying patients into 20-minute intervals (31–50, 51–70, and 71–90 minutes) to assess the durability of mucosal clarity over time.

Results

Comparison	Adequate gastric mucosal visibility (%)	P
10–20 min vs 20–30 min	27% vs 62%	0.001
10–20 min vs >30 min	27% vs 67%	0.001
10–20 min vs Placebo	27% vs 2.6%	0.001
20–30 min vs >30 min	62% vs 67%	0.174
20–30 min vs Placebo	62% vs 2.6%	0.001

1,140 patients (median age 42years, 63.2% males) were included in the study. Adequate GMV was significantly higher in all preparation groups compared to placebo (p < 0.001). The 20–30 minute and >30 minute groups demonstrated the highest rates of adequate visibility (61.5% and 73.5%, respectively), both significantly superior to the 10–20 minute group (26.9%) and placebo (2.6%). In the >30 minute subgroup analysis, the 31–50 minute interval demonstrated the highest adequacy (73.5%). The 51–70 minute group showed moderate adequacy (52.1%), which was not significantly different from the 20–30 minute group (p = 0.207). However, the 71–90 minute group had reduced adequacy (43.8%) and did not differ significantly from the 10–20 minute group (p = 0.145). No adverse events were reported across study arms.

Conclusion

Endoscopy performed between 20 and 50 minutes after mucosal preparation yields optimal gastric mucosal visibility. A decline in visibility beyond 50 minutes suggests a practical upper limit for endoscopy timing following preparation. These findings support structured, time-sensitive scheduling following mucosal cleansing to enhance procedural quality.

Disclosure

None

Introduction

Quality patient information is the cornerstone of personalised cancer care, empowering treatment decisions and improving patient outcomes. As a team of colorectal nurse specialists in a large hospital trust in the Northeast of England we undertook a quality improvement project to enhance information provision for patients diagnosed with bowel cancer. This initiative was prompted by recognition of both the critical importance of accurate information for cancer patients and the need to adapt service delivery during the COVID-19 pandemic

Results & Findings

The quality improvement project began with a comprehensive survey designed to evaluate patients' perceptions of current information provision and gather suggestions for enhancement. Twenty patients, deliberately selected to represent diverse demographic profiles, participated in the survey. A striking finding was that 80% of respondents sought additional information about their condition from internet sources, particularly from trusted websites like NHS Choices, Macmillan, and Cancer Research UK. The same percentage expressed interest in more visual content such as photos of staff, and most indicated a preference for accessing digital information alongside traditional printed materials.
Following the survey, two focus group workshops were conducted in collaboration with the NHS organisation dedicated to personalised care approaches. The focus groups examined patient experiences across the entire care pathway, from initial referral through treatment and follow-up, guided by principles of realistic medicine and shared decision-making.
We presented several innovative proposals for information delivery: an interactive diagram of the colorectal cancer journey accessible via the Trust's website; patient contact cards containing team details, phone numbers, website links, and QR codes; a webpage featuring QR codes linking to specific information resources; a video depicting the patient journey from referral through investigations; and plans to refresh the existing service webpage. These proposals aligned with the survey finding that most patients used online resources for information.

Summary & Discussion

This quality improvement initiative has provided us with valuable insights and confidence to advance information provision strategy in ways that genuinely reflect patient needs and experiences. The findings support a multimedia approach combining digital and traditional resources, with particular emphasis on accessible online information that patients can engage with at their own pace. As a team we created a patient information pathway that mirrors the actual patient journey and authentically reflects diverse patient experiences.

Conclusion

The project demonstrates the value of collaborative approaches to service improvement that meaningfully involve patients as partners. By putting patients at the heart of t project, the colorectal nursing team was positioned to develop information resources that truly meet patient needs during a challenging time in their lives. This patient-centred approach to information provision exemplifies the team's commitment to not only meet but exceed patient expectations in their privileged role as cancer care providers.

Disclosure

NONE