The role of Imaging-enhanced endoscopy for the assessment of endoscopic activity in inflammatory bowel disease: a diagnostic test accuracy network meta-analysis

Le résumé fourni par la source

Aims Accurate endoscopic assessment is crucial to manage inflammatory bowel disease effectively (IBD). Advanced imaging-enhanced endoscopy (IEE) offers significant advantages over standard and high-definition white-light endoscopy (SD-WLE & HD-WLE) in assessing subtle inflammatory changes, bringing endoscopy closer to histology and predicting clinical outcomes (1). Therefore, comprehensive data comparing the diagnostic performance of different advanced IEE over standard techniques for IBD assessment are needed. This study aims to perform a network meta-analysis (NMA) to compare the diagnostic accuracy of SD-WLE and HD-WLE versus different IEE techniques, including virtual chromoendoscopy (VCE), probe-based confocal laser endomicroscopy (pCLE) and endocytoscopy (ECS) for the assessment of endoscopic activity, in terms of histological activity and major adverse outcomes (MAO) prediction. Methods The NMA protocol was registered with PROSPERO (No. CRD42024585229). A systematic search was conducted through PubMed and Scopus from inception to November 2024 to obtain studies comparing at least two endoscopic techniques for IBD assessment. Duplications were automatically removed using the EPI platform (No. 51645). Collected data included study design, population age, endoscopic technique, endoscopic score, and ground truth. Data extraction restored contingency tables for calculating sensitivity, specificity, positive and negative predictive value (PPV & NPV) and observed agreement. Results After removing 77 duplicates, 164 full manuscripts were screened. We analysed 2844 patients from 29 manuscripts: 26/29 UC, 25/29 prospective, and 3/29 multicentre, all carried out in adults. The most frequent pairwise comparisons were between HD-WLE vs pCLE (7/29), HD-WLE vs i-scan (6/29), and HD-WLE vs linked colour imaging (LCI) (5/29). The ground truth was based on histological activity in 14/29 studies, MAO in 10/29, and both in 5/29 studies. For predicting histological activity, the pooled observed agreement was 69% for SD-WLE and 75% for HD-WLE; lower than 90% for i-scan, 88% for ECS, 87% for pCLE, 83% for blue laser imaging (BLI), and 79% for autofluorescence imaging (AFI) (p<.001). For the prediction of MAO, HD-WLE showed an observed agreement of 65%, lower than i-scan (80%; p<.001). Considering HD-WLE as the statistical reference of the NMA, the random effects model demonstrated a significantly higher sensitivity for endoscopic assessment using i-scan (Relative Risk 1.1793, p=.0213), pCLE (RR 1.286, p=.0007), ECS (RR 1.307, p=.044), LCI (RR 1.317, p=.0007), AFI (RR 1.785, p=.013), and significantly lower for SD-WLE (RR 0.603, p=.043). Compared to HD-WLE, a significantly higher observed agreement was found with pCLE (RR 1.135, p=.047) and ECS (RR 1.2618, p=.025) [ 1 ]. Conclusions IEE, including pCLE and ECS, significantly improved diagnostic accuracy for predicting histological activity and MAO compared to SD-WLE or HD-WLE. Additionally, i-scan, LCI, and autofluorescence imaging demonstrated substantial improvements in sensitivity. Publication History Article published online: 27 March 2025 © 2025. European Society of Gastrointestinal Endoscopy. All rights reserved. Georg Thieme Verlag KG Oswald-Hesse-Straße 50, 70469 Stuttgart, Germany

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.