Introduction

In ulcerative colitis (UC) clinical trials, endoscopic disease activity is commonly assessed via central reading of the overall Mayo Endoscopic Score (MES). Segmental MES can provide a more detailed evaluation of mucosal response, but it is limited by high inter-rater variability in both anatomical landmark localization and disease activity grading. Automatic MES grading can assist reducing reader discrepancies and detect local mucosal changes that may not be estimated by global scoring. In this study, we developed and evaluated an artificial intelligence (AI)-based system for automated segmental MES grading enabling disease extent quantification and granular endoscopic remission estimation.

Aims & Methods

In contrast to previous approaches, the proposed AI system introduces a novel hierarchical framework for automated MES grading at the frame-, segment-, and video-level. The system comprises three modules, trained on 5,897 retrospectively acquired, aggregated, and anonymized endoscopic videos. A foundation model (FM), trained on 22 million frames, was developed to extract clinically relevant frames from full-length colonoscopies. A colon segmentation module was built upon the FM to localize three anatomical segments: descending colon (DC), sigmoid colon (SI), and rectum (RM). An automated MES classification module was then trained to perform frame-level MES grading. Hierarchical aggregation of local scores, combined with temporal smoothing and non-informative filtering, enabled both segmental and global MES estimation. For independent testing, 915 colonoscopy videos were annotated for each segment for MES scoring by expert gastroenterologists using a 2+1 reading paradigm. Colon segment reference annotations were automatically extracted from endoscopic videos based on recording software annotations. System performance was evaluated using the Quadratic Weighted Kappa (QWK) for both segmental and overall MES agreement. Classification accuracy in detecting endoscopic granular remission (MES 0, 1 vs 2, 3) was reported for all colon segments.

Results

The AI system demonstrated strong agreement (QWK=0.752 [95% CI: 0.719, 0.782]) with gastroenterologist assessments for overall MES on the test dataset. Good agreement was confirmed for SI (QWK=0.647 [95% CI: 0.602, 0.687]) and RM (QWK=0.687 [95% CI: 0.646, 0.724]), moderate agreement was found for DC (QWK=0.592 [95% CI: 0.541, 0.638]). Frame-level colon segment classification accuracy was 41% for DC, 69% for SI, and 78% for RM; with lower performance in DC linked to high variability in reference annotations. Colon segmentation enabled a 37.72% reduction in mean colonoscopy video time by filtering out-of-body and insertion-phase sections. The system showed very high accuracy in detecting both overall (ACC=90.38%) and segmental (DC: 78.80%; SI: 82.84%; RM: 85.03%) endoscopic remission. Mean accuracy in detecting the number of colon segments per video with active disease was 59.13%.

Conclusion

We introduced a novel AI system for automated, granular MES scoring in ulcerative colitis patients, hierarchically grading colonoscopy videos. Our framework demonstrated strong agreement with gastroenterologists in both overall and segmental endoscopic activity assessments. By integrating continuous scoring with anatomical landmarking, the AI system can offer a more comprehensive view of total disease burden in UC patients which can lead to improved precision of disease estimation and treatment response evaluation in clinical settings.

Disclosure

All authors are employees of Clario Inc., Philadelphia, USA.

Introduction

The Tokyo Guidelines recommend a CRP cutoff >10 mg/L for diagnosing acute cholangitis. However, this threshold was defined in lithiasic populations and may not be applicable to malignant biliary obstruction, where CRP elevation often reflects tumor-driven inflammation rather than true infection. Applying this universal threshold in oncologic settings risks antibiotic overuse and promotes resistance.

Aims & Methods

This study aimed to reassess the diagnostic value of the current CRP cutoff in tumor-related cholangitis and to identify a more appropriate threshold to enhance specificity and guide targeted antibiotic use.
We conducted a retrospective monocentric study of consecutive patients at a tertiary center between May 2016 and October 2024. Among 571 patients undergoing ERCP, 268 patients with confirmed malignant biliary obstruction and available bile sampling were included. Bacteriological status was defined as positive when either direct bile examination or culture isolated microorganisms. CRP levels, leukocyte counts (hyperleukocytosis >10,000/mm³), and tumor etiology were recorded. ROC curve analysis was performed to determine the optimal CRP cutoff. Comparative analyses employed Mann-Whitney U, chi-square, and multivariate logistic regression.

Results

The mean age was 65.9 ± 12.4 years, with an equal sex distribution (50% male). Tumor etiologies included pancreatic cancer (56%), cholangiocarcinoma (22%) — 40 hilar and 18 distal cases, gallbladder cancer (13%), ampulloma (7%), and extrinsic or secondary compression (3%). The median CRP was significantly higher in culture-positive patients (47 mg/L [IQR 21–98]) compared to culture-sterile patients (22 mg/L [IQR 11–36]; p<0.001). ROC analysis identified an optimal CRP cutoff of 21 mg/L (AUC = 0.64), achieving a sensitivity of 67% and specificity of 54%. Hyperleukocytosis was present in 66% of culture-positive cases versus 31% of sterile cases (p<0.001). In multivariate analysis, CRP ≥21 mg/L (OR 2.1; 95% CI [1.3–3.4]; p=0.002) and hyperleukocytosis (OR 3.4; 95% CI [1.9–5.9]; p<0.001) were independently associated with bile bacteriological positivity.

Conclusion

In malignant biliary obstruction, the universal CRP threshold of 10 mg/L lacks specificity and risks antibiotic overuse. A higher threshold around 21–25 mg/L, combined with leukocyte count assessment, improves diagnostic precision. Our findings advocate for a context-specific adaptation of the Tokyo Guidelines to better differentiate tumor-related inflammation from true infection, ultimately enhancing patient safety and antimicrobial stewardship in oncologic care.

Introduction

The development of peroral endoscopic myotomy (POEM) has led to a paradigm shift in the treatment of esophageal motility disorders^1-5. Hiatal hernia (HH) has been reported in approximately 4-20% of patients with achalasia^6-8. Due to the altered position of the proximal stomach entering thoracic cavity, careful anatomical inspection at the completion of myotomy is essential to minimize the risk of adverse events⁷. Although two observational studies suggested that the presence of HH did not significantly affect clinical outcomes^9-10, these studies were limited by small sample sizes from single centers, with 13 and 56 cases of HH, respectively. Further study involving larger cohorts is needed to enable meaningful comparisons and to elucidate the impact of HH on POEM outcomes.

Aims & Methods

We hypothesized that the coexistence of a HH may be associated with adverse events, such as mediastinitis and esophageal perforation, or incomplete myotomy due to the technical challenges of approaching the esophagogastric junction^6-7. This retrospective cohort study aimed to compare outcomes between patients with and without HH undergoing POEM. Using the Nationwide Inpatient Sample in the United States from 2016 through 2022, we identified 6572 patients undergoing POEM for achalasia or non-achalasia spastic disorders. One-to-one propensity score matching was performed to compare adverse events, defined as any adverse event associated with an extended length of stay (>3 days), reintervention, 30-day mortality, and length of stay between the groups.

Results

	HH (n=1276)	Non-HH (n=1276)	P value
Adverse events	107 (8.4)	80 (6.3)	0.04
Infection (peritonitis, mediastinitis, sepsis)	62 (4.9)	42 (3.3)	0.04
Bleeding (gastrointestinal bleeding, hematoma)	32 (2.3)	26 (2.0)	0.49
Perforation (esophageal perforation, laceration)	9 (0.7)	8 (0.6)	0.52
Trauma (pneumothorax, injuries to other organs)	14 (1.1)	5 (0.4)	0.03
Reintervention	92 (7.2)	67 (5.3)	0.04
Pneumatic dilatation	88 (6.9)	67 (5.3)	0.08
Laparoscopic Heller myotomy	18 (1.4)	6 (0.5)	0.01
Second POEM	0 (0)	4 (0.3)	0.05

In the propensity score-matched analysis involving 1276 pairs, the HH group had significantly higher rates of adverse events (8.4% vs 6.3%; p = 0.04) and reintervention (7.2% vs 5.3%; p = 0.04) compared to the non-HH group (Table). The most common adverse events were infections, such as peritonitis and mediastinitis (4.9% vs 3.3%; p = 0.04). A longer length of stay (median: 4 vs 3 days; p < 0.01) was observed in the HH group. No significant difference in 30-day mortality was noted between the groups. On logistic regression analysis, HH (OR 1.33; 95% CI, 1.06–1.68) and achalasia (vs. non-achalasia spastic disorders; OR 1.54; 95% CI, 1.17–2.02) were associated with adverse events.

Conclusion

The risk of adverse events and reintervention were higher in the HH group. These findings highlight the need for tailored management in POEM for HH cases.

References

[1] Inoue H, et al. POEM for esophageal achalasia. Endoscopy. 2010;42:265-71.
[2] Ramchandani M, et al. Recent Advancements and Future Perspectives of POEM. Clinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association. 2024;22:1983-96.e2.
[3] Inoue H, et al. Statement for GERD after POEM from an international multicenter experience. Esophagus. 2020;17:3-10.
[4] Ponds FA, et al. Effect of POEM vs Pneumatic Dilation on Symptom Severity and Treatment Outcomes Among Treatment-Naive Patients With Achalasia: A Randomized Clinical Trial. Jama. 2019;322:134-44.
[5] Inoue H, et al. Clinical practice guidelines for POEM. Digestive endoscopy. 2018;30:563-79.
[6] Ushimaru Y, et al. Occult Hiatal Hernia in Achalasia Patients: Its Incidence and Treatment Options. Dig Surg. 2019;36:418-25.
[7] Fletcher PR. Acute hiatal hernia with oesophageal perforation following Heller's operation. Br J Surg. 1978;65:486-8.
[8] Sanei-Moghaddam A, et al. Cricopharyngeal myotomy for cricopharyngeus stricture in an inclusion body myositis patient with hiatus hernia: a learning experience. BMJ case reports. 2013;2013.
[9] Xu P, et al. Peroral endoscopic myotomy is an effective treatment for achalasia combined with hiatal hernia. Surgical endoscopy. 2024;38:4543-9.
[10] Tutuian G, et al. Achalasia and Hiatal Hernia: A Rare Association and a Therapeutic Challenge. J Neurogastroenterol Motil. 2023;29:455-9.

Disclosure

No disclosures for this study