Introduction

In Western countries, endoscopic submucosal dissection (ESD) for early-stage Barrett’s esophageal adenocarcinoma (BEA) is usually combined with radiofrequency ablation (RFA) to eradicate the residual BE mucosa. In contrast, as RFA is not commercially performed in Japan, lesion-based ESD (L-ESD) has become the standard treatment for superficial BEA, followed by post-ESD surveillance. Although L-ESD is effective, metachronous recurrence may occur from preserved BE, and complete BE resection may be required depending on recurrence risk. While BE length has been associated with recurrence risk, it may not fully predict the recurrence. In Japan, image-enhanced magnification endoscopy with peri-lesional biopsies before ESD is used to assess lesion margins. Thus, we focused on the association between peri-lesional biopsy and the risk of metachronous recurrence.

Aims & Methods

We aimed to investigate clinicopathological features of BEA with metachronous recurrence after curative L-ESD, with a particular focus on peri-lesional biopsy. This was a retrospective, multicenter cohort study conducted at three high-volume Japanese institutions between April 2006 and August 2024. Patients with lesions assessed by image-enhanced magnification and peri-lesional biopsy before L-ESD were included. We excluded cases where lesions for which the definitive pathological diagnosis after L-ESD was not BEA. Patients were classified into two groups: no recurrence (non-R) and metachronous recurrence (MR). In peri-lesional biopsies, “negative” was defined when all biopsy specimens in each case showed no atypia or malignancy, and “non-negative” was defined when any specimen exhibited atypia or malignancy. We evaluated clinicopathological features and explored potential risk factors. The cutoffs for metachronous recurrence risk in the Prague C and M classification were also calculated using receiver operating characteristic curves.

Results

A total of 207 patients with 215 lesions were included. Of these, 200 patients (200 lesions) were in the non-R group, and 7 patients (15 lesions) in the MR group. In the non-R and MR groups, median ages were 67 and 77 years (p=0.19); long-segment BE rates were 26.5% and 85.7% (p=0.002); median Prague C lengths were 1 and 5 cm (p=0.007); and M lengths were 2 and 6 cm (p=0.002), respectively. Median tumor sizes were 15 and 13 mm (p=0.22); depths were pT1a: 69% and 86.7%, and pT1b: 31% and 13.3% (p=0.24); unclear or positive horizontal margins were 9.0% and 20.0% (p=0.17), respectively. The cutoff value for the risk of metachronous recurrence in the Prague C length was 5.0 cm (area under the curve [AUC]=0.78, 95% confidence interval [CI]: 0.56-0.99) and M length was 5.0 cm (AUC=0.83, 95%CI: 0.64-1.0).
The non-negative rate of peri-lesional biopsies was significantly higher in the MR group than in the non-R group (35.7% [5/15] vs. 12.5% [25/200]; p=0.04). In five lesions of the MR group, peri-lesional biopsies were all negative for the initial BEA (0% [0/22] of peri-lesional biopsies were non-negative), however, those for metachronous BEA were included non-negative species in each five lesion (50.0% [11/22] of peri-lesional biopsies were non-negative).

Conclusion

BE length is an indicator for risk stratification of recurrence. In addition, given the high incidence of non-negative results in peri-lesional biopsies at the time of metachronous recurrence, the temporal progression of atypia in the surrounding BE may have advanced significantly until metachronous recurrence. This finding is crucial for determining the potential candidates for BE eradication.

Introduction

Acute upper gastrointestinal bleeding and timing to endoscopy are a recognised as risk factors for mortality. European Guidelines * suggest that the hemodynamical resuscitation could be performed before the endoscopy, however, indications are lacking about the timing of blood transfusions.

Aims & Methods

The aim of the study is to evaluate the transfusion timing to reduce death risk in patients with upper gastrointestinal bleeding. This is a prospective multicentre cohort study, including all consecutive hospitalized patients with upper gastrointestinal bleeding (UGIB). Clinical, endoscopic source of bleeding, medical and endoscopic treatment, time to endoscopy, and timing to transfusion were collected. We adopted the ESGE* guidelines to define the need for transfusion, timing to endoscopy and mortality. We considered timing to transfusion before or after the endoscopic procedure. Statistics: multivariate backward logistic regression was used to estimate the adjusted odds ratio death risk for variables with main clinical relevance or highly statistical significance at univariate analyses. Variables identified by the logistic regression were used into the propensity-score matching to compare the death risk between the two groups.

Results

2,017 out of 3324 patients received a transfusion and were included in the analysis; the mean age was 69.7 [±15.1], male was 1,365 (67.7%); the source of bleeding was non-variceal in 1634 patients (81 %), 178 patients died (8.8%); 69 out of 700 (9.9 %) died in the group transfusion before endoscopy; 109 out 1317 (8.3 %) in the group transfusion after endoscopy with an OR 1.2 (C.I. 95% .86 to 1.68) p<= 0.23. Mortality between the two groups did not significantly differ when haemoglobin levels at admission were considered, nor after adjustment for potential confounding variables, including ASA score, in-hospital bleeding, haematemesis at presentation, shock at admission, altered mental status and timing to endoscopy. Propensity score matching analysis indicated that among patients who received a transfusion, the risk of death decreased when the transfusion was administered before the endoscopy and the haemoglobin level was below 7 g/dL (coefficient = -0.06, p = 0.06; see Table).

Table. Propensity-score matching Haemoglobin level before VS after endoscopy
Admission Heamoglobin value gr/L	Coefficient	SE.	z	P>|z|	[95%conf. nterval]
< 7 gr/l	-0.06	0.03	-1.86	0.063	-0.11 0.00
7 - 8 gr/l	-0.03	0.04	-0.95	0.342	-0.10 0.04
8 - 10 gr/l	0.02	0.03	0.52	0.605	-0.05 0.08

Conclusion

Red pack cell transfusions before or after endoscopy did not modify the death risk in patients with acute upper gastrointestinal bleeding; the death risk was less frequent in patients transfused before endoscopy and with a haemoglobin value lower than 7 gr/l. However, this risk increased in those patients transfused before endoscopy and with a haemoglobin value > 8 gr/l.

References

* Ian M. Gralnek et al. Endoscopy 2021; 53: 300–332

Introduction

Individuals with Lynch syndrome (LS) are predisposed to different tumors, including gastric (GC) and duodenal (DC) cancer, especially those with specific mutations or familial aggregation. However, the role of esophagogastroduodenoscopy (EGD) in these patients is still uncertain, with heterogeneous recommendations among different guidelines.

Aims & Methods

To estimate the overall incidence of GC and DC in individuals with LS, to determine the incidence of GC and DC in higher-risk subgroups within these patients, as well as to clarify the usefulness of screening or surveillance by EGD.

Systematic search of Pubmed and Embase for cohort studies and clinical trials from inception to April 2024. Incidences of GC and DC were calculated, as well as the relative risk (RR) and their respective 95% confidence intervals in those studies comparing patients under surveillance with EGD vs without. A meta-analysis was performed using a random-effects model. Heterogeneity was assessed with the I² statistic.

Results

A total of 584 studies were obtained, of which 30 were finally included, with a total of 31264 patients with LS. The overall incidence of GC was 2.04% (95%CI 1.63-2.45). The incidence of GC was 2.60% (95%CI 1.61-3.60) in MLH1 carriers and 2.91% (95%CI 1.63-2.45) in MSH2 carriers. In those with LS and family history, the incidence of GC was 25.94% (95%CI 15.01-36.87). The incidence of DC was 1.61% (95%CI 1.21-2.00).
Only 3 studies compared patients with LS in follow-up with EGD vs those without follow-up, evidencing a protective effect of EGD for GC [RR 0.31 (95%CI 0.14-0.69)], but not for DC [RR 0.65 (95%CI 0.22-1.32)].

Conclusion

The incidence of GC and DC in patients with LS is higher than that of the general population, especially in those patients with MLH1, MSH2 mutations or family aggregation. Surveillance with EGD could help to reduce this incidence through the diagnosis of premalignant lesions.

Introduction

In recent years, artificial intelligence (AI) has been widely explored to enhance capsule endoscopy (CE), with the goal of improving the efficiency of the reading process. While most AI models have been developed for small bowel and colon analysis, the development of esophagogastric (E-G) models has been limited due to the scarcity of frames captured during the procedure, making it challenging to develop a robust model. This study aims to develop an interoperable ubiquitous model capable detecting pleomorphic lesions in the E-G tract.

Aims & Methods

We included 59,482 E-G frames, from 774 CE procedures of 5 centers, to develop a Convolutional Neural Network (CNN). The dataset was divided following an exam-based split, with 90% allocated for training – including a 5-fold cross validation – while the remaining was used for testing. The primary outcomes were: sensitivity, specificity, accuracy and area under the curve (AUC).

Results

During training, the CNN achieved mean sensitivity of 85.0% (IC95% 76.5–93.5), specificity of 96.3% (IC95% 93.9–98.7), accuracy of 93.6% (IC95% 91.7–95.4), and AUC-ROC of 0.98 (IC95% 0.97–0.98). During testing, the sensitivity, specificity and accuracy of CNN were 92.2%, 95.1% and 94.6%, respectively.

Conclusion

CE-AI models capable of assessing E-G are crucial step toward developing a truly robust panendoscopic model. This ubiquitous and interoperable model capable of lesion detection in both esophagus and stomach, achieved good overall accuracy. However, prospective real-world studies comparing its performance with standard upper endoscopy are still needed to validate its clinical applicability.

Introduction

Acute GI bleeding (GIB) can present significant mortality. Topical Endoscopic Haemostatic Agents (TEHAs) have been added to endoscopic treatment modalities for GIB’s. There is no dedicated international consensus on the role of TEHAs in GIB control. We aimed to develop the first expert led consensus statements to provide international guidance on the use of TEHAs in acute GIB’s.

Aims & Methods

A team of 15 world renowned experts in the field of acute GIB from 8 countries were recruited. All panel members have published extensively and possess significant expertise in the cognitive and endoscopic aspects of GIB management. A first meeting was held to define statements and divide them in 5 subgroups: Indications of therapy, mechanism of action, usage, safety and published evidence. Using the RAND/UCLA appropriateness method, they voted on statements by combining expert collective judgement and best available evidence in a two-round voting process. Extensive literature searches were carried out on each statement to provide the latest evidence base to support decisions. A second meeting was held to discuss the results of the first meeting and refine statements. A second round of voting was conducted to rate refined statements. Statements were rated on a 9-point interval scale (1 to 9). A higher score implied a “more appropriate” statement i.e. an important statement in the management of GIB patients. Four statistical methods were used to delineate statements that satisfied all criteria of appropriateness: 1. mean absolute deviation from the mean, 2. BIOMED concerted action on appropriateness definition, 3. RAND method and 4. inter-percentile range adjusted for symmetry methods. For a statement to be considered appropriate it had to meet all statistical definitions of appropriateness showing consensus agreement.

Results

24 statements were presented in round one. 15 statements met all four statistical definitions of appropriateness and proceeded to round two. Following round two, a list of 11 final consensus statements were scored as appropriate and met all four statistical definitions of appropriateness showing overall consensus being reached (Table 1). There was 100% agreement that TEHAs are effective in achieving haemostasis in malignancy related GIB’s (statement 2), can be used as ‘salvage’ therapy for non-variceal GIB’s (statement 4), and can be used as a bridge to definitive non-endoscopic therapy for persistent GIB’s (statement 5). These represent the most important indications for TEHAs that endoscopists should recognise.

Table 1: Final eleven statements voted as appropriate reaching expert consensus after round two. (Score 1-3 = inappropriate statement, score 4 – 6 = uncertain about statement, score 7 – 9 = appropriate statement), TEHAs = Topical Endoscopic Haemostatic Agents

Final list of agreed consensus statements	Median score	No. of experts 1-3 score	No. of experts 4-6 score	No. of experts 7-9 score
Statement 1: When utilised, TEHAs can be used when there is an actively bleeding ulcer/lesion in the Gastrointestinal tract	9	0	0	15 (100%)
Statement 2: TEHAs are effective in achieving immediate haemostasis in patients with malignancy related upper Gastrointestinal bleeding	9	0	0	15 (100%)
Statement 3: TEHAs can be used in combination with other haemostatic modalities in patients with high-risk upper Gastrointestinal bleeding	9	0	1 (7%)	14 (93%)
Statement 4: TEHAs may be used as salvage/rescue treatment in patients with acute non variceal upper Gastrointestinal bleeding, when other endoscopic interventions have failed	9	0	0	15 (100%)
Statement 5: TEHAs can be used as a bridge to definitive non endoscopic therapy (for example interventional radiology or surgery) in persistent upper gastrointestinal bleeding	9	0	0	15 (100%)
Statement 6: Whereas TEHAs are highly effective, standard endoscopic therapy remains the gold standard and should be considered early in the treatment algorithm for Gastrointestinal bleeding	9	0	0	15 (100%)
Statement 7: TEHAs are an effective treatment option for lower Gastrointestinal bleeding	7	0	3 (20%)	12 (80%)
Statement 8: TEHAs are an effective agent for haemostasis particularly in units where tertiary expertise is not available Statement 9: The application of TEHAs should be targeted at a specific bleeding site or lesion	7 9	0 0	3 (20%) 0	12 (80%) 15 (100%)
Statement 10: The endoscopic technique for application of TEHAs has a short learning curve Statement 11: Activation failuire, including spray cathether occlusion and impaired visualisation are the most common technical failuires related to the TEHAs device use and should be recognised early with backup options available in the endoscopy unit if this occurs (ie another of the same or alternative TEHA)	8 9	0 0	2 (13%) 0	13 (87%) 15 (100%)

Conclusion

We present the first dedicated international consensus statements aimed at providing guidance to clinicians on best practice use of TEHAs in patients with acute GIB’s. There was consensus on the need for future trials to compare use of different haemostatic powders in GIB’s.

Introduction

Accurate real-time differentiation of colorectal polyps during colonoscopy is critical for optimal decision-making and cost-effective care. Artificial intelligence (AI)-based tools such as MagentiQ-Eye Colo have been developed to assist endoscopists in predicting polyp histology based on endoscopic features. However, real-world validation against histopathology remains limited.

Aims & Methods

We conducted a retrospective study using colonoscopy data collected at our endoscopy unit. Histologically confirmed colorectal polyps were included, with subtypes classified as adenomas (LGD and HGD), sessile serrated adenomas (SSA), or hyperplastic polyps (HP). The CADe AI tool MagentiQ-Eye Colo was applied retrospectively to stored colonoscopy video recordings to generate polyp histology predictions ("Neoplastic" or "Non-Neoplastic"). These were compared to the corresponding histopathological diagnoses. Diagnostic performance metrics including sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated.

Results

A total of 257 polyps were analyzed. Of these, 198 polyps received a conclusive AI-based prediction, while 59 polyps were labeled as "Inconclusive" and were excluded from performance analysis.
Overall, MagentiQ-Q achieved a sensitivity of 96.8% and specificity of 22.3% in detecting neoplastic polyps, with an overall accuracy of 58.1%.
A focused subgroup analysis was performed on 133 polyps classified as adenomas, SSA, or HP:
MagentiQ-Eye achieved:

• 96.2% sensitivity and 3.8% specificity for adenomas (adenoma villosum, adenoma tubulare),
• 70.8% sensitivity and 29.2% specificity for sessile serrated adenomas (SSA),
• 43.3% sensitivity and 56.7% specificity for hyperplastic polyps (HP).

Further analysis of adenoma subtypes showed:

• LGD (n=50): 96.0% predicted as neoplastic.
• HGD (n=29): 97.4% predicted as neoplastic.

Size-based analysis revealed that MagentiQ was more likely to classify polyps as neoplastic with increasing estimated size: in the "Advanced" size category, 98.7% of polyps were predicted as neoplastic, in the "Small" category, this rate was 84.0%, in the "Diminutive" category, the neoplasia prediction rate dropped to 61.5%.

Conclusion

The MagentiQ-Eye Colo AI system shows excellent sensitivity in identifying neoplastic polyps, especially conventional adenomas including both LGD and HGD. However, overprediction of neoplasia was frequent in non-neoplastic lesions, notably hyperplastic polyps and a subset of SSA cases, leading to reduced specificity. Moreover, the system yielded inconclusive predictions in 23% of cases, suggesting a need for algorithm refinement and prospective validation to support its clinical integration.