P201 Reproducibility of magnetic resonance enterography assessment of disease activity in Crohn's Disease using central readers
I. Ordas*1, J. Rimola1, G. Zou2, 3, C. Santillan4, K. Horsthuis5, S. Taylor6, J. Panes1, R. Khanna2, 7, J. Stoker8, L. Stitt2, G. D'Haens2, 9, W.J. Sandborn2, 10, B.G. Feagan2, 3, 7, M.K. Vandervoort2, B.G. Levesque2, 10
1Hospital Clinic de Barcelona, IDIBAPS and CIBERehd, Gastroenterology Unit, Barcelona, Spain, 2Robarts Clinical Trials Inc., Robarts Research Institute, University of Western Ontario, London, Ontario, Canada, 3University of Western Ontario, Department of Epidemiology and Biostatistics, London Ontario, Canada, 4University of California, UCSD Inflammatory Bowel Disease Center, San Diego CA, United States, 5Academic Medical Center, Inflammatory Bowel Disease Center, Amsterdam, Netherlands, 6University College London, Department of Medical Imaging, London, United Kingdom, 7University of Western Ontario, Department of Medicine, London , Canada, 8Academic Medical Center Amsterdam, Department of Internal Medicine, Amsterdam, Netherlands, 9University of Amsterdam, Academic Medical Center, Amsterdam, Netherlands, 10University of California San Diego, Division of Gastroenterology, La Jolla, United States
Reproducibility is a critical property of magnetic resonance enterography (MRE) Crohn's disease activity indices if they are to be used as outcome measures in clinical trials. We evaluated the reproducibility of two MRE disease activity instruments, the MaRIA and London indices, when centrally read by expert gastrointestinal body imaging radiologist readers in a multi-center trial setting.
Four central reader radiologists at centers in Europe and North America reviewed 50 MRE sequences of patients with a spectrum of Crohn's disease activity and location. Readers assessed the MaRIA and London indices, pre-specified individual MRE findings, and a global rating of severity based on a visual analogue scale (VAS). Intraclass correlation coefficients (ICCs) for intra- and inter-rater agreement were calculated for each assessment.
Intra-rater ICCs (95% confidence intervals) for the MaRIA, London, London Extended indices and the VAS were 0.89 (0.84 to 0.91), 0.84 (0.76 to 0.88), 0.81 (0.71 to 0.85) and 0.86 (0.81 to 0.90). Corresponding inter-rater ICCs were 0.71 (0.61 to 0.77), 0.50 (0.32 to 0.62), 0.56 (0.40 to 0.64), and 0.71 (0.62 to 0.77) (Table 1).
“Table 1. Reproducibility of magnetic resonance enterography indices for Crohn’s disease”
The correlation between each reader’s VAS and the MaRIA, London, and London Extended indices were 0.79 (0.71 to 0.85), 0.68 (0.58 to 0.77) and 0.67 (0.58 to 0.76), respectively.
These Results indicate that there was "almost perfect" intra-rater reproducibility of centrally read MaRIA and London indices. Inter-rater agreement was "substantial" for the MaRIA and "moderate" for the London indices. The MaRIA index appears to have the best operating characteristics and is thus the currently preferred instrument for use in clinical trials.