Contrast studies such as barium follow-through (BaFT) have traditionally been the criterion standard for the investigation of small bowel Crohn disease (CD) in children (1). Newer imaging modalities, such as magnetic resonance imaging (MRI), are being used for reasons including reduction of radiation exposure, gaining greater anatomical detail, assessment of extraluminal disease, improved pathology detection, and tolerability. MRI of the abdomen has been shown to be superior to BaFT for detection of small bowel pathology in adults with recurrent CD (2). A meta-analysis has confirmed the high sensitivity and specificity of MRI in adult inflammatory bowel disease (IBD) and suggests its use to avoid radiation exposure (3). It has been shown that patients with IBD diagnosed in childhood are at higher risk of receiving potentially dangerous radiation exposure during their lifetime (4). This has been a major factor in the increasing use of MRI in children, despite the relative paucity of evidence.
Different protocols have been used to optimise the imaging sequences of the small bowel, but there is no consensus as to which methodology provides both adequate anatomical and pathological information, and is also practical to use in children. Institutions rely on patients drinking large volumes of an oral contrast agent or insertion of a nasojejunal tube to achieve adequate small bowel distension for imaging (5,6). Our institution developed a protocol that uses a small volume of oral preparation, and therefore should be more easily tolerated in a paediatric population. The main aim of the present study was to demonstrate how the local protocol adopted for abdominal MRI in patients with IBD compared with the previous routine investigation, BaFT.
All consecutive patients at a single regional paediatric gastroenterology centre from September 2008 to December 2010 with a new endoscopic diagnosis of IBD or a previous diagnosis requiring reassessment underwent both abdominal MRI and BaFT. The abdominal MRI was performed on a GE Signa Excite 1.5 T MRI scanner (GE Healthcare, Chalfont St Giles, UK) according to the protocol described in Table 1. Small bowel distension was achieved by drinking 50 mL of lactulose diluted to 100 mL total volume with water (flavoured to taste with lemon squash) 1 hour before the scan. Patients were then able to drink water freely (up to 1 L, but not required) before imaging. Intravenous (IV) hyoscine butylbromide was injected before the images were acquired, and IV gadolinium was given during the MRI.
The BaFT was performed by the patients drinking approximately 750 mL of Baritop solution. Images of the abdomen were then taken using intermittent screening (Ultimax FDP, Toshiba Medical Systems, Tokyo, Japan) until the barium was seen to enter the caecum. The studies were reported by nonblinded radiologists with an interest in gastrointestinal imaging. Comment was made of technical difficulties with the imaging, including imaging artefacts, underdistension of bowel loops, or poor tolerance of the oral contrast agent.
Data on demographics, diagnosis of CD, ulcerative colitis or indeterminate colitis (IC), duration of disease, Montreal classification for CD, endoscopic findings, and histology were obtained by case note review. Macroscopic and histological findings were correlated with the radiology reports if the endoscopy was performed within 1 month of either radiological investigation. All of the patients who had radiological investigations were included in the analysis, even if the study failed. Patients were excluded if there was no final diagnosis of IBD, or if MRI and BaFT were performed >1 month apart. The 2 radiology reports were compared by an independent assessor (E.G.) for any discrepancies between the 2 reports, and whether that difference was significant. Ethical approval was not required for the present study because it was initiated to assess and reconfigure service provision, in accordance with the guidelines of the UK National Research Ethics Service (7).
Eighty-eight patients with IBD were referred for both BaFT and MRI, 71 with a final diagnosis of CD, 11 with ulcerative colitis, and 6 with IC. Ages ranged from 5 to 18 years (median 13). Sixty-seven patients had newly diagnosed IBD and 21 patients were being reassessed (20 with CD, 1 with IC). The Montreal classifications of the patients are shown in Table 2. Fifty-one of 71 patients with CD had ileal or ileocolonic disease that was nonstricturing, nonpenetrating disease without perianal involvement (example Figs. 1 and 2). Stricturing disease was unusual and there were no patients with penetrating disease. A summary of the pathology found is seen in Table 3.
In 41 patients (47%), the MRI scan was performed with no technical difficulties. In 31 patients, (36%) the scans were reported to have suboptimal distension of the jejunum, but in only 1 was the test nondiagnostic. In 15 patients (17%), there was motion artefact that affected the quality of the images, but all were reported as of adequate diagnostic quality. One patient had the study abandoned because the patient would not take the lactulose. No patients required a nasoenteric tube. A summary of technical difficulties is seen in Figure 3. The terminal ileum was not identifiable in 1 BaFT because of overlapping bowel loops with no other reported technical difficulties.
Regarding the barium studies, 37 of 88 patients had normal reports. Fifty patients had disease in the ileum reported, with 7 of those reporting a stricture and 1 reporting a possible fistula. Seventy-six of these 88 patients had complete ileocolonoscopy within 1 month of the BaFT. In 11 of these patients (14%), there were abnormal macroscopic findings at endoscopy that were not seen on the BaFT. In 58 patients (76%), the findings on BaFT and endoscopy were in agreement, and in 8 patients (11%), there were abnormal findings on the BaFT not seen at endoscopy.
Comparing BaFT to histology when available, which we consider to be the criterion standard for detecting ileal inflammation, there were 11 false-negative and 4 false-positive reports. Three BaFT studies reported jejunal disease. Strictures were reported in 6 patients and suspected in 1 other on BaFT. This only correlated with endoscopic findings in 1 case and with the MRI report in 3 cases. The summary of BaFT findings when compared with terminal ileal histology is shown in Figure 4. Using histology when available as a criterion standard, BaFT had a sensitivity of 76% and a specificity of 67% for detecting terminal ileal disease.
Twenty-five of the 87 scans were normal. In comparing endoscopic findings of the small bowel, there were 76 patients with available comparisons (as above). In 68 patients (89%), the findings on MRI and endoscopy were the same; this compares with 76% with BaFT. In 7 patients (9%), there were abnormal macroscopic findings at endoscopy that were not seen on the MRI, and in 5 patients, (7%) there were abnormal findings on the MRI not seen at endoscopy.
In comparison with histology for ileal inflammation, there were 6 false-negative and 1 false-positive MRI report. Eleven patients had changes on MRI in which the ileum was not reached endoscopically, and therefore comparison with macroscopic appearance and histology was not possible. Eight MRI reports described jejunal disease. There were 8 patients with MRI reports of strictures. The summary of MRI findings in comparison to histology is shown in Figure 4. Using histology as the criterion standard, MRI had a sensitivity of 83% and a specificity of 95% for detecting terminal ileal disease.
Comparison Between MRI and BaFT
In comparing the 2 modalities, in 24 patients (27%) both studies were reported as normal. In 28 patients (32%) the findings were abnormal but concordant. Six patients (7%) had similar reports, but the MRI was more specific and gave greater detail of disease. One patient had more detail in the BaFT report. Overall, 59 patients (67%) had similar reports from both modalities. Twenty-seven patients (31%) had additional findings on the MRI of colitis or ileal or jejunal disease, which were not seen on BaFT (example Figs. 5 and 6). Four BaFTs detected gastro-oesophageal reflux. Four BaFTs reported terminal ileal abnormalities, which were not demonstrated on the patients’ MRIs or endoscopies (histology normal).
The MRI reports also gave additional information not available from BaFT. Bowel wall enhancement postgadolinium was reported, which helped distinguish inflammatory from fibrotic strictures in 5 patients (although the radiologists stated that they cannot reliably do this on enhancement pattern alone). MRI detected incidental extraintestinal pathology in 3 patients.
MRI is a tool that is being increasingly used in the diagnosis and reassessment of both paediatric and adult CD. This practice is being driven by increasing concern about radiation exposure, particularly in children; however, practice varies widely and there is little evidence on its methodology and use. The present study demonstrates the feasibility of our institution's protocol for use in young people without the need for nasojejunal intubation. It provides diagnostic images that certainly compare favourably, particularly in the terminal ileum, to what has previously been considered the criterion standard radiological investigation in paediatric IBD, BaFT.
The Porto criteria were developed through the European Society of Paediatric Gastroenterology, Hepatology, and Nutrition to aid in diagnosing IBD in children (1). These guidelines suggest contrast studies as the modality of choice for small bowel imaging. The newer guidelines of the European Crohn's and Colitis Organisation suggest that MRI may be used in centres with expertise (8), but this is based on few comparative data. There have been several publications using MRI in children with IBD, but most of these do not provide direct comparison with the previous criterion standard, BaFT (5,6,9).
Perhaps the most attractive reason for the increasing use of MRI in paediatrics is the lack of radiation exposure. This is particularly true in a chronic relapsing disease such as CD, in which repeated imaging is often required during a lifetime of illness. The mean dose for a BaFT in the paediatric population in the United Kingdom has been measured and varies according to age, but for ages 10 and 15 years, it is 4.5 and 7.2 mSv (10), respectively, giving a lifetime adjusted risk of up to 0.3% for developing a malignancy. More significantly, Desmond et al (4) showed that in CD, a cumulative dose of radiation of >75 mSv increased cancer mortality by 7.3% and one of the significant risk factors for this occurring was diagnosis before age 17 years.
Apart from the lack of radiation, in the present study, MRI has been shown to be superior to BaFT for detecting small intestinal pathology. Using our protocol, the sensitivity and specificity of the MRI in detecting terminal ileal disease are similar to other studies (5,9,11,12). These studies have used different protocols, in particular different oral contrast solutions. Probably the most commonly used agent is polyethylene glycol (9,11), but this along with others requires a large volume to be consumed. Our protocol was adopted because of the small volume of liquid required, palatability, and ability to give the contrast agent orally. The lactulose acts as an osmotic agent, allowing improved visualisation of the small bowel loops. The lactulose is nonabsorbable and (in the absence of small bowel overgrowth) remains largely unmodified in the small bowel, links with water molecules, fills the lumen, and thus distends the loops (13). The tolerability is particularly important in a largely adolescent patient group, who may require repeat examinations. It may be that this small volume compromises proximal jejunal distension, but this does not appear to significantly affect the diagnostic quality of the investigation. Another important advantage of MRI is the ability of gadolinium contrast enhancement to help differentiate between inflammatory and fibrotic narrowing (14). Finally, our study also demonstrated incidental lesions, although the clinical relevance of these types of findings is unclear. An adult study suggests that there is no clinical benefit from the detection of such extraintestinal lesions (15).
Apart from MRI, other alternative modalities have been considered in IBD. Computerised tomography (CT) has also been used to image the small bowel and has been shown to be effective in paediatric IBD (16); however, it has a much higher radiation exposure than BaFT (17), and a recent study has shown an equivalent accuracy of magnetic resonance enterography compared with CT in a paediatric cohort (18). CT also accounted for the majority of the radiation in the patients in the study of Desmond et al (4).
Wireless capsule endoscopy may be considered in symptomatic children with suspected small bowel involvement, after exclusion of strictures or stenoses. Capsule endoscopy may be superior in sensitivity to the conventional techniques in detecting mucosal lesions in the small bowel (19–21), although a recent publication showed that MRI was superior (22); however, studies comparing the different techniques excluded patients with strictures, which limits their conclusions. In addition, bowel wall thickness, stenoses with prestenotic dilation, as well as extraluminal complications cannot be identified with capsule endoscopy. Wireless capsule endoscopy is considered safe, with a low risk of capsule impaction beyond infancy (23). This may suggest a role for this investigation in children who are unable to tolerate other imaging modalities; however, the low risk of impaction remains, which a small bowel follow-through or a pretest with ingestion of a dissolvable patency capsule are not sensitive enough to exclude (21). In patients with IBD, the risk for capsule impaction is higher than that for other indications (eg, bleeding, protein-losing enteropathy), because of the incidence of stricturing disease. Finally, young children may not swallow the capsule and therefore may require endoscopic replacement of the capsule in the duodenum (21) under general anaesthetic. Given these practical difficulties alone, it is likely that wireless capsule endoscopy will remain a reserve investigation tool and nonroutine for imaging in paediatric IBD.
Ultrasound is also being used increasingly to image the bowel in IBD and other conditions and has the shared advantage with MRI of no radiation exposure. It has been shown to be sensitive and specific in detecting small bowel disease (24,25), but there has been paediatric evidence that it is inferior at detecting disease that is not in the terminal ileum (26). The conflicting nature of the evidence in this field implies that the results are still extremely operator dependent and that perhaps with time this may become an increasingly used tool; however, given that its place in adult practice is not yet established, it is unlikely to be considered a criterion standard investigation in children in the short term.
This is the largest series of children with IBD undergoing abdominal MRI for detecting small bowel pathology to our knowledge, and one of the few studies that compares directly to BaFT. The data confirm the feasibility of MRI as a routine investigation in children with IBD. Only 1 patient in 88 failed to tolerate the procedure. The technical difficulties have had minimal adverse effect on image quality and may continue to reduce with time and experience of the protocol. The numbers of MRI sequences performed for paediatric small bowel MRI have already been reduced to make the examination more child friendly. We use a small volume of palatable contrast without the need for a nasogastric tube. The sensitivity and especially specificity of MRI for detecting small bowel pathology were superior to BaFT, with no radiation burden. Although there may be issues around availability, expertise, and cost, MRI should replace BaFT as the criterion standard in investigating paediatric IBD.
The authors acknowledge the help of the diagnostic imaging department of the Royal London Children's Hospital, particularly the MRI staff, for helping with the new protocol.
1. IBD Working Group of the European Society for Paediatric Gastroenterology HpaN. Inflammatory bowel disease in children and adolescents: recommendations for diagnosis—the Porto criteria. J Pediatr Gastroenterol Nutr
2. Bernstein CN, Greenberg H, Boult I, et al. A prospective comparison study of MRI versus small bowel follow-through in recurrent Crohn's disease. Am J Gastroenterol
3. Horsthuis K, Bipat S, Bennink R, et al. Inflammatory bowel disease diagnosed with US, MR, scintigraphy, and CT: meta-analysis of prospective studies. Radiology
4. Desmond AN, O’Regan K, Curran C, et al. Crohn's disease: factors associated with exposure to high levels of diagnostic radiation. Gut
5. Borthne AS, Abdelnoor M, Rugtveit J, et al. Bowel magnetic resonance imaging of pediatric patients with oral mannitol MRI compared to endoscopy and intestinal ultrasound. Eur Radiol
6. Dagia C, Ditchfield M, Kean M, et al. Feasibility of 3-T MRI for the evaluation of Crohn disease in children. Pediatr Radiol
8. Van Assche G, Dignass A, Reinisch W, et al. The second European evidence-based consensus on the diagnosis and management of Crohn's disease: special situations. J Crohns Colitis
9. Magnano G, Granata C, Barabino A, et al. Polyethylene glycol and contrast-enhanced MRI of Crohn's disease in children: preliminary experience. Pediatr Radiol
11. Laghi A, Borrelli O, Paolantonio P, et al. Contrast enhanced magnetic resonance imaging of the terminal ileum in children with Crohn's disease. Gut
12. Horsthuis K, de Ridder L, Smets AM, et al. Magnetic resonance enterography for suspected inflammatory bowel disease in a pediatric population. J Pediatr Gastroenterol Nutr
13. Bouhnik Y, Neut C, Raskine L, et al. Prospective, randomized, parallel-group trial to evaluate the effects of lactulose and polyethylene glycol-4000 on colonic flora in chronic idiopathic constipation. Aliment Pharmacol Ther
14. Masselli G, Brizi GM, Parrella A, et al. Crohn disease: magnetic resonance enteroclysis. Abdom Imag
15. Jensen MD, Nathan T, Kjeldsen J, et al. Incidental findings at MRI-enterography in patients with suspected or known Crohn's disease. World J Gastroenterol
16. Brown S, Applegate KE, Sandrasegaran K, et al. Fluoroscopic and CT enteroclysis in children: initial experience, technical feasibility, and utility. Pediatr Radiol
17. Gaca AM, Jaffe TA, Delaney S, et al. Radiation doses from small-bowel follow-through and abdomen/pelvis MDCT in pediatric Crohn disease. Pediatr Radiol
18. Gee MS, Nimkin K, Hsu M, et al. Prospective evaluation of MR enterography as the primary imaging modality for pediatric Crohn disease assessment. AJR Am J Roentgenol
19. Guilhon de Araujo Sant’Anna AM, Dubois J, Miron MC, et al. Wireless capsule endoscopy for obscure small-bowel disorders: final results of the first pediatric controlled trial. Clin Gastroenterol Hepatol
20. Thomson M, Fritscher-Ravens A, Mylonaki M, et al. Wireless capsule endoscopy in children: a study to assess diagnostic yield in small bowel disease in paediatric patients. J Pediatr Gastroenterol Nutr
21. El-Matary W. Wireless capsule endoscopy: indications, limitations, and future challenges. J Pediatr Gastroenterol Nutr
22. Casciani E, Masselli G, Di Nardo G, et al. MR enterography versus capsule endoscopy in paediatric patients with suspected Crohn's disease. Eur Radiol
23. Moy L, Levine J. Wireless capsule endoscopy in the pediatric age group: experience and complications. J Pediatr Gastroenterol Nutr
24. Martínez MJ, Ripollés T, Paredes JM, et al. Assessment of the extension and the inflammatory activity in Crohn's disease: comparison of ultrasound and MRI. Abdom Imag
25. Parente F, Greco S, Molteni M, et al. Oral contrast enhanced bowel ultrasonography in the assessment of small intestine Crohn's disease. A prospective comparison with conventional ultrasound, x ray studies, and ileocolonoscopy. Gut
26. Epifanio M, Baldisserotto M, Spolidoro JV, et al. Grey-scale and colour Doppler sonography in the evaluation of children with suspected bowel inflammation: correlation with colonoscopy and histological findings. Clin Radiol