P190 Systematic review: Gastrointestinal ultrasound scoring indices for inflammatory bowel disease

T. Goodsall1, T. Nguyen2, C. Ma2, V. Jairath2, R. Bryant3

1Department of Gastroenterology, John Hunter Hospital, Newcastle, Australia, 2Department of Clinical Trials, Robarts Clinical Trials Inc., London, Canada, 3Department of Gastroenterology, The Queen Elizabeth Hospital, Adelaide, Australia

Background

The management of inflammatory bowel disease (IBD) requires frequent monitoring and assessment of disease activity. Endoscopic assessment with biopsy remains the gold standard for disease activity. Gastrointestinal ultrasound (GIUS) is a non-invasive, accessible and affordable test used to assess and monitor IBD and has been shown to be similar to MRI for detecting disease. The aim of this study was to systematically review the literature to identify scoring indices used for GIUS measurement of disease activity in IBD and to appraise their operating characteristics.

Methods

A systematic search of Embase, Medline, Pubmed, Cochrane Central and Clinical Trials.gov from inception to July 2019 was conducted according to PRISMA guidelines. Included were all study types reporting GIUS indices used for grading activity of severity of IBD in comparison to an objective reference standard. Studies using an exclusive clinical reference standard were excluded. All study types and abstracts were considered. Study quality was assessed using the QUADAS tool.

Results

27 eligible studies were identified investigating 1647 patients. Disease phenotype was Crohn’s disease (CD) (n = 13), ulcerative colitis (UC) (n = 10) and IBD (n = 4). The most common reference standard was colonoscopy (n = 23), histology (n = 2), and imaging (n = 2). Bowel wall thickness was an index parameter in 26 studies. The most frequent cut off was 3mm (n = 10), 4mm (n = 9), 5mm (n = 1), and not specified (n = 6). There was no noticeable difference in magnitude of cut off when stratified by disease phenotype. Colour Doppler was an index parameter in 16 studies and was based on the Limburg score (n = 7), binary (n = 7) or categorical (n = 2). Bowel wall stratification was an index parameter in 15 studies and was more frequently used in UC (70%) and IBD (75%) indices than in CD indices (38%). Other index parameters included bowel wall compressibility, presence of complications such as abscess or fistula, bowel wall echogenicity, mesenteric inflammatory, lymphadenopathy, contrast enhancement, ulceration, peristalsis, strictures, absence of haustra coli, and tissue sonoelastography. Twenty-three studies were identified as at risk of bias. Overall concordance was substantial to excellent and accuracy was good to excellent. Two studies demonstrated substantial inter-observer agreement. No studies reported intra-observer agreement.

Conclusion

The identified GIUS scoring indices demonstrate applicability to both CD and UC with good accuracy and concordance. Current evidence does not adequately address concerns about the intra- and inter-observer variability of GIUS. There is a need for robust validation of an evidence-based GIUS index before more widespread use in IBD as a surrogate for colonoscopy and in clinical trials.