In patients with Crohn disease, magnetic resonance enterography (MRE) scores are minimal when no treatment is needed and significantly increase when surgical or medical intervention is mandated. *Indicates outliers from the mean in this patient group. The number next to the asterisk corresponds to a random identification number assigned to these patients. †P < .01. ‡P < .05. Values are given as mean (SD).
Customize your JAMA Network experience by selecting one or more topics from the list below.
Messaris E, Chandolias N, Grand D, Pricolo V. Role of Magnetic Resonance Enterography in the Management of Crohn Disease. Arch Surg. 2010;145(5):471–475. doi:10.1001/archsurg.2010.68
To assess the impact of magnetic resonance enterography (MRE) on therapeutic decision making for patients with Crohn disease.
Tertiary care medical center.
One hundred twenty patients who had either a history of or high suspicion for Crohn disease with onset of new symptoms underwent MRE over 18 months at our institution. All patients with Crohn disease were classified according to the Montreal system.
Magnetic resonance enterography and medical vs surgical therapy.
Main Outcome Measure
Changes in management after MRE findings.
Magnetic resonance enterography demonstrated active Crohn disease in 57.5% of patients, chronic changes of Crohn disease without active inflammation (eg, stricture, fistula, or abscess) in 12.5% of cases, and no evidence of Crohn disease in 30% of cases. After MRE, 37 (31%) had no change in medical therapy, 64 (53%) had additional medical management for active inflammation, and 19 (16%) underwent an operation for complicated Crohn disease or medical intractability. In all surgical patients, the intraoperative findings were consistent with the MRE diagnosis. The mean (SD) MRE score was 1.6 (0.5) for patients who had no change in their treatment plans, 5.8 (1) for patients who underwent surgery, and 8 (0.4) for patients who had their drug regimen changed (P < .001). The MRE score independently correlated with need for intervention (P = .001).
Magnetic resonance enterography shows promising ability to characterize the presence of active Crohn disease as well as chronic complications (eg, differentiate between stricture due to edema vs fibrotic scarring). Magnetic resonance enterography is fast becoming a useful adjunct in the management algorithm of patients with Crohn disease.
Crohn disease (CD) is a chronic inflammatory bowel disease characterized by transmural and segmental inflammation of the intestinal wall and can be associated with extraluminal complications.1 Several imaging techniques are used to detect CD and its complications. Traditionally, small-bowel follow-through (SBFT) examination has been the standard radiologic approach used to assess patients with active CD, as more than 70% of patients have involvement of the small intestine.2 Small-bowel enteroclysis has been reported to be more accurate than SBFT at detecting early mucosal lesions but requires nasojejunal intubation. Both methods provide only limited and indirect information in regard to the state of the bowel wall and extraluminal extension of CD.2,3 Cross-sectional imaging techniques such as computed tomography (CT) and magnetic resonance (MR) imaging are advantageous in their ability to visualize superimposed bowel loops and to improve visualization of extraluminal findings and complications.4,5 Thus, CT enterography has become accepted as a noninvasive imaging technique to evaluate CD. Computed tomographic enterography allows direct visualization of extraenteric structures and can reliably identify active inflammation in the small intestine. However, because patients with CD often require multiple imaging studies over their lifetime, MR enterography (MRE) has been introduced as a radiation-free alternative method to evaluate patients with CD.6-8 Investigators in several reports have shown MRE to be useful for the identification of active ileitis, assessment of disease activity, and evaluation of extraenteric complications in patients with CD.9-11 However, there is very limited literature on the application of MRE findings in clinical practice for patients with CD. The goal of the present study was to retrospectively assess the impact of MRE on the therapeutic decision making for patients with CD.
The study protocol was approved by our institutional review board. The study population derived from eligible subjects who were known to have CD or subjects who were suspected of having CD who required radiologic investigation of the small intestine. Subjects were enrolled in the study from April 2007 to October 2008. Four hundred fifty-seven consecutive patients underwent MRE over 18 months at our institution. Our study focused on 120 patients of any age who had either a history or high suspicion for CD and onset of new symptoms (eg, pain, nausea, diarrhea, fever). Institutional exclusion criteria for performing an MRE included renal insufficiency, documented adverse reaction to iodinated contrast material, or any contraindication to MR imaging (eg, cardiac pacemakers or metallic implants).
A fellowship-trained board-certified gastrointestinal radiologist independently reviewed all MRE images. To aid in interpretation of images, the small intestine and colon were divided into 6 segments that included the jejunum and proximal ileum, terminal ileum, cecum and ascending colon, transverse colon, descending colon, and rectosigmoid and anus. The terminal ileum was defined as the 20-cm segment of ileum nearest to the ileocecal valve or ileocecal anastomosis. Distinction among segments of the small bowel was based on the readers' visual assessment of their location (ie, the bowel located in the left upper abdomen was considered to be the jejunum; the bowel located in the pelvis and the right lower abdomen was considered to be the ileum), their fold pattern (ie, closely packed folds for the jejunum and sparse folds for the ileum), and their estimated distances from the duodenojejunal junction and the ileocecal valve. A scoring system that assigns numeric values to widely agreed indicators of active inflammation12-16 was used to quantify the radiographic findings. For MRE images, the reviewer assigned a grade to the degree of wall thickness (normal <3 mm), hyperenhancement (relative to the renal cortex), high signal intensity of the bowel wall or perienteric fat on T2-weighted images, mesenteric engorgement (hyperemia, fibrofatty proliferation), and local adenopathy (normal nodes <5 mm in short axis). A visual 4-grade scoring system was used: 0 = absent, 1 = mild, 2 = moderate, and 3 = severe. The reviewer also evaluated and graded (0 = absent, 1 = present) the presence or absence of complications of CD, such as abscess, fistula, stricture, and other CD-associated findings (nephrolithiasis, sacroiliitis, and sclerosing cholangitis). Finally, the scoring of all findings was summarized in an overall MRE score designed to attempt to distinguish active inflammation from chronic fibrotic changes of CD.12 For a variety of MRE pictures, see the video.
All demographic data, clinical history, operative history, indication for performing the MRE, and outcome of the patient were recorded in the database. Furthermore, all patients were classified according to the Montreal system, which includes age, location of disease, and disease phenotype. Note was made of medical decisions following MRE, as well as patient outcome after an intervention, if any was performed. Patients' records were reviewed 3 months following MRE to identify changes to the initial post-MRE decision to treat.
The χ2 test was used for analysis of categorical variables. The t test and 1-way analyses of variance were used to compare continuous variables between patient groups. A regression analysis model was implemented to examine the ability of MRE score to predict the need for intervention when the model was adjusted for age, sex, and history of surgical intervention. P values less than .05 were considered statistically significant.
Patients' ages ranged from 10 to 80 years (mean [SD], 37.1 [1.4] years) and the majority were female (n = 73 [61%] vs 47 males [39%]). Indications for MRE were new onset of pain or discomfort in 86% of the cases, suspected fistula (7%), abscess (4%) or stricture (2%), and bowel obstruction (1%). Forty-three patients had previous surgery for CD (36%), mostly for complicated disease either with a small-bowel stricture (60%) or a fistula (12%), while 12% had uncomplicated disease and underwent exploratory laparotomy with no intervention. Based on the medical history, the Montreal classification system was used to stratify the patients by age, disease location, and type of disease before the MRE (Table 1). In the review of the MRE images, bowel distention was higher than average in 94.6% of the cases and was found poor but adequate for reliable reading in 5 patients (5.4%).
Magnetic resonance enterography demonstrated active CD in 57.5% of patients, no active disease but a complication of it (eg, stricture, fistula, or abscess) in 12.5% of cases, and no disease in 30% of cases. Magnetic resonance enterography findings are presented in detail in Table 2. Magnetic resonance enterography detected cases of CD complicated with stricture (21%), fistula (12%), and abscess (6%) formation. On the basis of physical examination and MRE findings, 37 patients (31%) had no change in medical therapy, 64 (53%) had additional medical management, and 19 (16%) underwent an operation for complicated CD or medical intractability. In all surgical patients, the intraoperative findings were consistent with the MRE diagnosis as demonstrated in Table 3. The MRE score was found to significantly correlate with the need for intervention. The mean (SD) MRE score was 1.6 (0.5) (median, 0) for patients who had no change in their treatment plans, 5.8 (1) (median, 5) for patients who underwent surgery, and 8 (0.4) (median, 8) for patients who had their drug regimen changed (P < .001) (Figure). The presence of wall thickness, hyperenhancement (P = .001), high signal intensity of the bowel wall or perienteric fat on T2-weighted images (P = .001), mesenteric engorgement (P = .002), and local adenopathy (P < .001) were found to increase the need for medical or surgical intervention in patients with CD. In a regression model of analysis, MRE score was found to be an independent predictor for surgical or medical intervention (Table 4). A 1-unit increase in MRE score was found to augment by 1.5 times the odds of intervention in a patient with CD. At 3-month follow-up after MRE and the decision to treat were made, no changes in the treatment protocol were identified. None of the patients with normal MRE findings underwent an operation or were prescribed new medication. In addition, no crossover patients were noted in the patient groups who were initially treated with new medications or an operation.
There was no association between age or disease phenotype with presence of active disease on MRE (P .05). Patients with history of colorectal and perianal disease were more likely to have active disease on MRE (25%) than patients with previous ileocecal disease (7%) (P = .009).
Patients who had previous surgery for CD were examined as a subgroup. Overall MRE scores were similar in both patients with no previous operation (mean [SD], 6 [0.5]) and patients with history of an operation for CD (mean [SD], 5.1 [0.66]; P .05). Furthermore, prior surgery for CD did not significantly affect any of the individual components of the MRE score, such as wall thickness, hyperenhancement, high signal intensity of the bowel wall or perienteric fat on T2-weighted images, mesenteric engorgement, and local adenopathy (P .05). Prior surgery for CD was not found to alter the type of treatment administered to the patients or the location of the recurrence (most frequent terminal ileum or neoterminal ileum) or increase the risk for presenting with recurrent complicated CD (P .05). As expected, older patients (mean [SD] age, 44.6 [2.3] years) were more likely to have a history of surgery for CD compared with younger patients (mean [SD] age, 32.8 [1.68] years; P = .001).
Magnetic resonance enterography has been compared with more conventional imaging modalities in several studies. In a study of active CD in the terminal ileum with endoscopy and histology results as the criteria, MRE had a sensitivity of 89% compared with 72% for conventional enteroclysis.8 Other reports have compared CT and MR imaging in the assessment of CD. In a prospective study, CT enterography, MRE, and SBFT were found to be equally accurate in the detection of active inflammation in the small intestine.10 Although the sensitivity values of CT enterography (89%) and MRE (83%) are slightly higher than those of SBFT (67%-72%) with regard to active terminal ileitis, these differences are not significant.10,13-15,17 Researchers in comparative studies have also demonstrated that CT and MRE are more sensitive than SBFT in detecting extraintestinal complications. Despite the previous data, there is no literature to our knowledge to describe the impact of MRE on the decision making for treatment of patients with CD. Our study describes the MRE findings and the treatment based on these findings in 120 patients with CD with onset of new symptoms. The treatment algorithm included implementation of new pharmaceutical regimen, operative intervention, or no medical or surgical intervention at all. There was a 3-month follow-up to assure no immediate changes were made in the treatment regimen or intervention. In all patients, the MRE results were in complete accordance with the operating room findings. Furthermore, in the 3-month follow-up no changes were made in the treatment protocol that was initiated after the MRE. In patients with known CD and onset of new symptoms, MRE can add significant information to the clinical evaluation to determine if any therapy is needed and if it is, whether it should be surgical or medical.
An MRE scoring system designed to increase with increasing imaging indicators of active inflammation was used in our investigation. We demonstrate that patients with no or little inflammatory changes can be managed with no treatment or no change to their current treatment regimen protocol, patients with findings of chronic CD are likely to be treated surgically, and patients with imaging scores indicative of extensive active inflammation require initiation of medical treatment.
It has been reported that colonoscopy and MRE are of similar value to predict the risk of clinical recurrence in postoperative patients with CD.14,18 In the current investigation, it was demonstrated that MRE was able to detect disease activity irrelevant to the surgical history of the patient. In addition, the components of the MRE score did not differ between patients who had previous surgery and patients who did not, suggesting that MRE has the capability to differentiate fibrotic from inflammatory tissue even in cases where previous tissue manipulation has occurred.19
Patients with CD have a lifelong disease that may require multiple imaging studies over a lifetime, thus exposing the individual to increased doses of radiation. Desmond et al20 reported that patients with CD diagnosed at an early age and those with upper tract inflammation, penetrating disease, and requirement of intravenous steroids, infliximab, or multiple surgeries are exposed to higher doses of radiation. Magnetic resonance enterography is a radiation-free imaging study that offers results equivalent to conventional diagnostic modalities, as previously described.
There are several advantages and limitations in the design of our study. This is a retrospective review of data collected from patients who underwent MRE in our institution. It does not include patients who presented with new symptoms of CD and underwent a different diagnostic study or were initiated in a treatment protocol or were taken to the operating room based on physical examination only. In the patient population that we investigated, the decision making for treatment was made by 6 different medical physicians, either gastroenterologists or surgeons. The physician did not make the decision to treat solely on the results of the MRE but from overall evaluation of the patient; however, no other imaging modalities were used after MRE to assist in the treatment algorithm.
Patients with a history of CD and onset of new gastrointestinal symptoms remain difficult to diagnose and treat. Our study demonstrates that MRE shows promising results in identifying disease activity and successfully enhancing the ability of the clinician to implement new surgical or medical treatment if required. Magnetic resonance enterography has been proven to have high sensitivity and specificity in detecting CD in the gastrointestinal tract, hopefully prompting larger prospective studies to investigate whether the results of MRE can successfully alter the treatment algorithm of these patients and avoid unnecessary medical or surgical interventions.
Correspondence: Evangelos Messaris, MD, PhD, Rhode Island Hospital, 593 Eddy St, APC 415, Providence, RI 02903 (email@example.com).
Accepted for Publication: December 30, 2009.
Author Contributions:Study concept and design: Messaris, Grand, and Pricolo. Acquisition of data: Messaris, Chandolias, Grand, and Pricolo. Analysis and interpretation of data: Messaris, Grand, and Pricolo. Drafting of the manuscript: Messaris, Chandolias, and Pricolo. Critical revision of the manuscript for important intellectual content: Grand and Pricolo. Statistical analysis: Messaris. Administrative, technical, and material support: Chandolias, Grand, and Pricolo. Study supervision: Grand and Pricolo.
Financial Disclosure: None reported.
Previous Presentation: This paper was presented at the 90th Meeting of the New England Surgical Society; September 11, 2009; Newport, Rhode Island; and is published after peer review and revision.