Offir Ukashi, Shelly Soffer, Eyal Klang, Rami Eliakim, Shomron Ben-Horin, and Uri Kopylov. Gut and Liver 2023;17:516-28. https://doi.org/10.5009/gnl220507

Background and Aims: The aim of our study was to create an accurate patient-level combined algorithm for the identification of ulcers on CE images from two different capsules. Methods: We retrospectively collected CE images from PillCam-SB3′s capsule and PillCam-Crohn’s capsule. ML algorithms were trained to classify small bowel CE images into either normal or ulcerated mucosa: a separate model for each capsule type, a cross-domain model (training the model on one capsule type and testing on the other), and a combined model. Results: The dataset included 33,100 CE images: 20,621 PillCam-SB3 images and 12,479 PillCam-Crohn’s images, of which 3582 were colonic images. There were 15,684 normal mucosa images and 17,416 ulcerated mucosa images. While the separate model for each capsule type achieved excellent accuracy (average AUC 0.95 and 0.98, respectively), the cross-domain model achieved a wide range of accuracies (0.569–0.88) with an AUC of 0.93. The combined model achieved the best results with an average AUC of 0.99 and average mean patient accuracy of 0.974. Conclusions: A combined model for two different capsules provided high and consistent diagnostic accuracy. Creating a holistic AI model for automated capsule reading is an essential part of the refinement required in ML models on the way to adapting them to clinical practice.

INTRODUCTION: Patency capsule (PC) is a recommended procedure to rule out small bowel stenosis before video capsule endoscopy (VCE). We examined future clinical outcomes among patients with a failed PC vs patients in whom the PC had passed (passed PC). METHODS: A post hoc analysis of 2 prospective cohort studies of adult patients with quiescent small bowel Crohn's disease (CD) who underwent PC between 2013 and 2020. The primary composite outcome was the need for intestinal surgery or endoscopic dilation during follow-up in patients with or without a failed PC. RESULTS: A total of 190 patients were included (47: failed PC and 143: passed PC, median follow-up 34.12 months). Patients with a failed PC had higher rates of the primary composite outcome (21.3% vs 1.4%, hazard ratio [HR] 20.3, 95% confidence interval [CI] 4.4–93.7, P < 0.001) and also secondary outcomes including intestinal surgery (14.9% vs 0.70%, P < 0.001), endoscopic dilation (14.9% vs 0.70%, P < 0.001), admissions (23.3% vs 5.7%, P < 0.001), and clinical flares (43.9% vs 27.7%, P = 0.005) during follow-up compared with controls. Failed PC was the only statistically significant factor for surgery and/or endoscopic dilation, regardless of a B2/B3 phenotype at baseline. In sensitivity analyses restricted only to patients with a stricturing phenotype (n = 73), a failed PC still predicted the long-term composite outcome (HR 8.68, 95% CI 1.72–43.68, P = 0.002). Of the 190 patients ingesting a PC, only 1 patient with a failed PC had 48 hours of self-limiting mild symptoms. DISCUSSION: Patients with clinically stable CD with a failed PC have worse long-term clinical outcomes than those without, independently of the CD phenotype. Standalone PC may serve as a novel, safe, and affordable prognostic examination to identify patients with quiescent CD who have a higher risk for future worse clinical outcomes.

Background and Aims: Fecal calprotectin (FC) is known to be a sensitive biomarker of colonic inflammation but to a lesser degree of small bowel (SB) inflammation. Moreover, data on FC’s diagnostic levels in different SB segments are scarce. We aimed to examine FC’s diagnostic levels along the SB axis in CD. Methods: This was a post hoc aggregated analysis of 5 prospective studies of adult CD patients who underwent FC testing and SB video capsule endoscopy. Lewis score (LS) inflammation in different SB segments was tested for correlation with FC level after the exclusion of colonic disease. The diagnostic levels of FC for SB inflammatory topographical gradient were assessed using a receiver operating characteristic. Results: Two hundred and fourteen patients were included (age: 30 [24-43] year-old, males-57%). For a similar SB inflammatory activity (LS ≥ 135), FC levels incrementally increased from proximal to distal SB segments (63 [30-121] vs 190 [78-549], p = 0.005) and from distal SB segment to the colon (190 [78-549] vs 542 [185-1000], p = 0.010). The best FC cutoffs to identify isolated mild proximal/distal SB inflammation (LS ≥ 135) were 77 µg/g and 123 µg/g, respectively. A cutoff of 234 µg/g was best to detect more significant proximal inflammation (LS ≥ 350) when only mild distal SB inflammation was present. In sensitivity analyses, this proximal-to-distal FC gradient was maintained when LS ≥ 350 and LS ≥ 790 were used as the inflammatory reference values. Unlike FC, the magnitude of CRP elevation was unrelated to the topography of inflammation along the SB axis. Conclusions: FC may serve as a topographical biomarker of CD-activity, with its sensitivity to identify mucosal inflammation increases from proximal to distal SB segments.

Background: Inflammatory bowel disease (IBD) presents unique challenges in elderly patients due to comorbidities and treatment-related risks. Objectives: This study evaluates ustekinumab (UST) and vedolizumab (VDZ) efficacy and safety in elderly Crohn’s disease (CD) patients. Design: A retrospective cohort study at a tertiary medical center. Methods: CD patients aged ⩾60 years (elderly) treated with UST, compared to non-elderly (<60 years) patients treated with UST and elderly patients treated with VDZ. Clinical response was evaluated using the Harvey–Bradshaw index (HBI) and clinical biomarkers, alongside monitoring steroid use, hospitalization rates, treatment persistence, and surgical interventions. Results: The study included 166 CD patients: 32 elderly and 65 non-elderly patients treated with UST, and 69 elderly patients treated with VDZ. The mean duration of follow-up was 10.8 ± 2.8 months in the non-elderly group, 9.97 ± 3.28 months in the elderly UST group, and 10.0 ± 3.29 months in the VDZ group. Elderly UST patients were more likely to receive corticosteroids at initiation than non-elderly UST patients (44% vs 14%, p = 0.001). At 12 months, clinical response rates did not significantly differ between elderly and non-elderly UST groups, respectively (48% vs 40%, p = 0.5). However, elderly UST patients exhibited higher hospitalization rates over time compared to non-elderly UST patients (6-month: 19% vs 6.2%, p = 0.077; 12-month: 19% vs 4.6%, p = 0.055; log-rank p = 0.004). No significant differences were observed in clinical response and remission rates between elderly UST and elderly VDZ patients at 6 and 12 months. At 6 months, a higher hospitalization rate was observed in the UST group (19% vs 4.3% p = 0.027), but this difference did not persist over time. Conclusion: UST and VDZ are effective and safe treatments for elderly CD patients, despite higher hospitalization rates compared to non-elderly patients, likely due to age-related complications.

Background: The Montreal classification has been widely used in Crohn's disease since 2005 to categorize patients by the age of onset (A), disease location (L), behavior (B), and upper gastrointestinal tract and perianal involvement. With evolving management paradigms in Crohn's disease, we aimed to assess the performance of gastroenterologists in applying the Montreal classification. Methods: An online survey was conducted among participants at an international educational conference on inflammatory bowel diseases. Participants classified 20 theoretical Crohn's disease cases using the Montreal classification. Agreement rates with the inflammatory bowel diseases board (three expert gastroenterologists whose consensus rating was considered the gold standard) were calculated for gastroenterologist specialists and fellows/specialists with ≤ 2 years of clinical experience. A majority vote < 75% among participants was considered a notable disagreement. The same cases were classified using three large language models (LLMs), ChatGPT‐4, Claude‐3, and Gemini‐1.5, and assessed for agreement with the board and gastroenterologists. Fleiss Kappa was used to assess within‐group agreement. Results: Thirty‐eight participants from five countries completed the survey. In defining the Montreal classification as a whole, specialists (21/38 [55%]) had a higher agreement rate with the board compared to fellows/young specialists (17/38 [45%]) (58% vs. 49%, p = 0.012) and to LLMs (58% vs. 18%, p < 0.001). Disease behavior classification was the most challenging, with 76% agreement among specialists and fellows/young specialists and 48% among LLMs compared to the inflammatory bowel diseases board. Regarding disease behavior, within‐group agreement was moderate (specialists: k = 0.522, fellows/young specialists: k = 0.532, LLMs: k = 0.577; p < 0.001 for all). Notable points of disagreement included: defining disease behavior concerning obstructive symptoms, assessing disease extent via video capsule endoscopy, and evaluating treatment‐related reversibility of the disease phenotype. Conclusions: There is significant inter‐rater disagreement in applying the Montreal classification, particularly for disease behavior in Crohn's disease. Improved education or revisions to phenotype criteria may be needed to enhance consensus on the Montreal classification.