Editorial: can MRI enterography be an efficient tool for patient selection in clinical trials?

Abstract

Trials of new drug therapies to treat active Crohn's disease are usually complex and expensive. Historically the Crohn's disease activity index (CDAI) has been used to identify patients with presumed active disease prior to trial recruitment, but is rapidly falling from favour due to its relatively poor discriminatory ability. Coimbra et al.1 in a small but important study investigate whether MR enterography with a colonic water enema could be a more efficient method to select patients for trial inclusion in comparison to CDAI and colonoscopic derived Crohn's disease endoscopic index of severity (CDEIS). Specifically they applied a standardised MRI protocol across six sites recruiting 19 patients, investigated the repeatability of two scans separated by 24 h, and tested intra- and inter-reader agreement when applying the MaRIA score, a validated MRI score of disease activity. The researchers report reassuringly high scan quality across sites, good repeatability of scans performed a day apart, and reasonable reader agreement. Furthermore, by virtue of its ability to identify extra enteric complications and small bowel disease beyond the reach of colonoscopy, the authors suggest MRI would have correctly identified seven patients eligible for theoretical therapeutic trial inclusion (with two false positives), compared to four for colonoscopy and just two for CDAI. MRI is increasingly disseminated for evaluation of Crohn's disease and recommended by expert consensus groups.2 There is robust evidence that quantitative MRI activity scores such as MaRIA and MR enterography global score are effective,3, 4 have reasonable inter-reader agreement5 and are responsive to the effects of therapy.6-8 The ability of MRI to evaluate the full burden of disease, including extra-enteric complications and bowel inaccessible to colonoscopy, are significant advantages. These new data from Coimbra et al. tentatively suggest that MRI may be an effective triage tool for multi-centre clinical trials. Of course there must be a word of caution. The study included only 19 patients and therefore the findings must be reproduced in a larger cohort; large multi-centre and multi-reader trials of MRI are underway.9 Reader agreement was lower in the proximal small bowel, which requires attention, and the strength of correlation of the MaRIA score with CDEIS was lower than in the original single site validation studies3 (as would perhaps be expected with multi-site dissemination). The need for a routine colonic water enema requires clarification as this negatively impacted on patient tolerance and adds to the complexity of the MRI protocol. Cost effectiveness data must also be robust if the pharmaceutical industry is to invest in MRI as a routine part of clinical trials. Finally, one interesting observation was the presence of abnormal oedematous colonic wall on MRI in the absence of mucosal ulcers at endoscopy in two patients. Whilst Coimbra et al. considered these to be MRI false positives, the natural history of such lesions is unclear. It may be the ability of MRI to assess the complete bowel wall will hold advantages over the limited mucosal assessment of endoscopy and that ‘mural’ healing on MRI may even supersede endoscopic mucosal healing as the ultimate goal in Crohn's therapy. Declaration of personal interests: Stuart Taylor has served as a consultant to Robarts plc. Declaration of funding interests: Stuart Taylor is an NIHR senior Investigator and has received research funds from the NIHR and from University College London Hospitals Biomedical Research Centre.