Consensus Recommendations for Evaluation, Interpretation, and Utilization of Computed Tomography and Magnetic Resonance Enterography in Patients With Small Bowel Crohn’s Disease

Imaging Findings

Table 1 defines and describes imaging findings of mural inflammation at CTE and MRE, along with important diagnostic considerations and practical conclusions. Figure 1 pictorially illustrates an imaging-based morphologic construct that demonstrates the role of mural inflammation in driving Crohn’s disease exacerbations and response as seen at cross-sectional enterography, and which will be explained in greater depth after individual imaging findings have been reviewed. The pictorial representation of a single bowel loop is used to facilitate a unified understanding of how mural inflammation can change independent of signal properties of cross-sectional imaging modalities. Multiple studies have shown that in patients with Crohn’s disease, imaging findings of inflammation are strongly associated with the presence of histologic inflammation (36–40). Evidence describing and supporting the use of these imaging findings for small bowel inflammation is provided in references within Table 1. By extension and inference, similar findings can reflect enteric inflammation in the stomach and colon.

Table 1 Imaging Findings Associated With Small Bowel Crohn’s Disease Inflammation

Note.—Items in boldface are required descriptive terms that should be used when present. Conclusions are based on criteria identified in the methods, with the level of evidence summarized accordingly as very low, low, moderate, or high. DDX, differential diagnosis; NSAID, nonsteroidal anti-inflammatory drug.

Figure 1:

Download as PowerPoint

While the co-existence of segmental hyperenhancement and wall thickening are used in combination as imaging findings reflecting Crohn’s disease inflammation (40,41), a number of other conditions can result in these imaging findings even when segmental involvement is multifocal (42,43). Additionally, other imaging findings often seen in small bowel Crohn’s disease inflammation, such as mural stratification and intramural edema, can also be seen in a number of other conditions. Asymmetric inflammation in the bowel wall in Crohn’s disease is commonly more severe along the mesenteric border and is probably a specific feature in Crohn’s disease (Figure 2) (44). The co-existence of mural inflammation and penetrating complications should also suggest Crohn’s disease, in the absence of other known causes of penetrating complications, such as appendicitis, diverticulitis, tumor, and tuberculosis. Given these considerations, radiologists should diagnose inflammatory small bowel Crohn’s disease either in known Crohn’s disease patients when the nonspecific imaging findings of inflammation are present, or when enteric inflammation is asymmetric or co-exists with the typical penetrating complications of Crohn’s disease. In the absence of a clinical diagnosis of Crohn’s disease, asymmetric inflammation, or typical penetrating complications, radiologists should describe the location and length of nonspecific small bowel inflammation. Gastroenterologists can then correlate these nonspecific radiologic findings with endoscopic and other clinical data to guide further management.

Figure 2:

Download as PowerPoint

In addition to describing the length of intestinal inflammatory involvement, radiologists should describe imaging findings of severe inflammation based on wall thickness, presence of luminal ulcerations, and increased intramural T2 signal (4,23,45,46) (Figure 2). Luminal ulcerations appear as small focal breaks in the intraluminal surface of the bowel wall with focal extension of air or enteric contrast into the inflamed bowel wall (Figures 3 and 4). Because Crohn’s disease itself, as well as physiologic factors and technical factors affecting acquisition, can affect the degree of bowel wall contrast-enhancement, hyperenhancement is a sign of active inflammation, but is not used to describe severity unless quantitative measures are utilized (47). Mild inflammation is described when segmental hyperenhancement is present with minimal wall thickening of 3−5 mm and rarely causes luminal narrowing. Severe inflammation is present if ulcerations or high T2 intramural signal are identified (Figures 3 and 5). Restricted diffusion is a nonspecific sign of Crohn’s disease mural inflammation, but when other typical findings of mural inflammation are present on contrast-enhanced and/or T2-weighted images, restricted diffusion is a complementary and supportive finding that has been shown to correlate with severe inflammation at endoscopy (25). Restricted diffusion is present when intramural hyperintensity is present on high b-value images (often similar to reactive lymph nodes), and should prompt a careful assessment for other signs of severe inflammation. However, radiologists should be aware that false positives on diffusion-weighted imaging can be due to many factors, including suboptimal fluid distention. For example, the normal jejunum demonstrates increased relative nonfocal restricted diffusion in comparison to the normal ileum. The diagnosis of active Crohn’s disease should not be made on the basis of restricted diffusion alone (46,48).

Figure 3:

Download as PowerPoint

Figure 4:

Download as PowerPoint

Figure 5:

Download as PowerPoint

Multiple MR-based scoring systems that describe inflammatory severity have been developed based on measures of histologic or endoscopic inflammation and rely on the described visual observations (eg, ulcers, intramural hyperintense T2 signal) and/or quantitative measurements (eg, wall thickness, relative contrast enhancement). Scoring systems, such as MaRIA, Clermont score, and MEGS, differ in the imaging findings evaluated and the potential weighting given to each finding or measurement, and are used in clinical studies to quantitate improvement or worsening of active inflammatory Crohn’s disease (4,24,49–51). The advantage of the severity scoring systems is that they integrate imaging findings in a systematic and reproducible manner. When transmural healing occurs or penetrating complications develop at CTE/MRE after therapy, however, cross-sectional images clearly demonstrate information that can be quickly assessed and conveyed without performing systematic scoring. Moreover, scoring systems do not reflect how inflammatory severity can vary over an inflamed bowel segment or convey adequate information regarding length and location of disease, which is needed for clinical decision-making. Further refinement of specific imaging criteria that can be readily incorporated into clinical practice for mild, moderate, and severe inflammation will be a subject for future interdisciplinary investigation.

Crohn’s disease strictures result from complex interactions between inflammatory cells, cytokines, mesenchymal cells, and enteric flora, and result in variable degrees of luminal narrowing (52). The majority of Crohn’s disease strictures have both an inflammatory and a fibrotic component due to repeated inflammation and reparative damage (53,54). Estimating the relative contribution of inflammation, fibrosis, and smooth muscle hypertrophy in dominant strictures has been an area of active imaging investigation (55,56). However, there is no universally accepted clinical or histologic scoring system for stricture-related fibrosis (52). Gastroenterologists and radiologists generally refer to different physical findings when identifying a stricture. Endoscopists generally think of luminal narrowing as a stricture. Radiologists generally rely on the presence of proximal dilation (often defined as >3 cm), as many bowel segments with Crohn’s-related inflammation demonstrate luminal narrowing, and cross-sectional imaging cannot assess luminal compliance or readily differentiate between spasm or fixed narrowing at a single time point. Moreover, both predominantly fibrotic and predominantly inflammatory strictures can fail to respond to medical therapy and ultimately require surgical intervention. Several imaging techniques and findings, such as magnetization transfer, ultrasound elastography, diffusion-weighted imaging, and relative contrast enhancement on delayed MR imaging with gadolinium, are actively being investigated for their ability to estimate fibrosis in Crohn’s disease strictures, but none of them have been fully validated. However, multiphase cinematic thick slab imaging with balanced steady-state free precession (eg, true-FISP, FIESTA, or balanced FFE) can be helpful in detecting and increasing confidence in stricture presence at MRE (57,58). Until prospective studies validating the relationship of imaging findings to histologic fibrosis are completed and a consensus emerges, Crohn’s disease strictures can be reliably identified by both luminal narrowing and unequivocal upstream dilation in order to minimize false-positive findings (Table 1) (54). Fixed luminal narrowing without upstream dilation cannot reliably be diagnosed as a stricture on a single image, but when multiple pulse sequences, fluoroscopic observation, or serial imaging examinations demonstrate fixed narrowing without upstream dilation, it is appropriate for radiologists to describe that a probable stricture is present. Enteroclysis assessment can be helpful in equivocal cases, as it is more sensitive for stricture presence. Radiologists and clinicians should be aware that when strictures are in close proximity to each other, the ability to radiographically detect downstream small bowel strictures is compromised, as an upstream stricture is already causing an obstruction.

Following stricture identification, radiologists should state whether findings of inflammation are present or absent within the stricture (Figure 6). Findings of inflammation within a stricture are critical, as current medical treatments can alleviate inflammation and avoid or delay surgery, while true fibrotic strictures are likely to require strictureplasty, excision, or endoscopic bowel dilation. Additionally, strictures should be evaluated for symmetry, nodularity, or extension of soft tissue into the adjacent mesentery that may signal development of a neoplasm (59). Radiologists should report the number, location and length of Crohn’s disease strictures in patients so that gastroenterologists and surgeons can decide on the best therapeutic option and approach. While it is understood that the degree of bowel dilation proximal to a stricture is a result of many factors, including chronicity and ingested material, the degree of upstream dilation is often useful to endoscopists and surgeons in deciding if treatment is warranted, or which strictures to treat if multiple strictures are present. The combination of presence/absence and severity of inflammation, stricture length, and degree of upstream dilation and fistulas can provide clinicians with necessary information for treatment decisions (60).

Figure 6:

Download as PowerPoint

Table 2 summarizes imaging findings in penetrating complications and mesenteric findings in Crohn’s disease. Penetrating complications result from transmural inflammation and include sinus tracts, fistulas, inflammatory masses, abscesses, and, rarely, free intraperitoneal perforation. Sinus tracts can be blind-ending in the mesentery, terminate at fascial planes, or extend longitudinally within the bowel wall. Fistulas should be described by the 2 epithelial structures they connect (eg, enteroenteric, enterocolic, enterocutaneous, rectovaginal, or enterovesical). Enteric fistulas within the abdominal cavity should be described as simple or complex similar to perianal fistulas (61). Complex, asterisk-shaped fistula complexes are often seen that tether multiple loops of small bowel and/or colon (Figure 7). Inflammatory mass describes dense mesenteric inflammation adjacent to severe mural inflammation or penetrating complications that is not an abscess and does not have a well-defined fluid component. The term phlegmon is discouraged due to its ambiguous definition, as it does not describe if there is a drainable component as in an abscess, or nondrainable, as in localized inflammation or inflammatory mass. It should be noted that clinical experience and the pathologic literature supports the strong association between stricture formation and penetrating disease (29,30). Thus, when penetrating disease is present, visual inspection should be directed at the site of fistula origin for an inflamed and stenotic bowel segment with upstream dilation, as these are nearly always present. Conversely, the proximal end of an inflamed and stenotic bowel segment should be scrutinized for detection of penetrating complications, as most arise from that part of the involved segment. We acknowledge that a weakness of the current proposal is that some inflamed small bowel segments giving rise to fistulas will not cause proximal small bowel dilation, as the upstream pressure gradient causes decompression through the fistula rather than dilation of the proximal bowel; these segments would not be identified or termed strictures based on a strict interpretation of our proposed scheme. However, because the evidence is overwhelming when a complex fistula is seen to arise from a small bowel segment with active inflammation and no upstream dilation is present, one might consider an impression in the clinical report, such as “complex penetrating disease with active inflammatory small bowel Crohn’s disease with luminal narrowing; stricture with imaging findings of active inflammation highly likely.”

Table 2 Imaging Findings of Penetrating Disease and Mesenteric Inflammation in Crohn’s Disease

Note.—PV, portal vein; SMV, superior mesenteric vein.

Figure 7:

Download as PowerPoint

Because approximately one-quarter of Crohn’s disease patients present with an anorectal fistula, complete imaging of the anal sphincters and perineum is imperative for every CTE and MRE examination (Figure 8). Artifacts often occur over the anus due to the placement of exterior phased-array coils at MRE, but adequate anal imaging can be performed in such cases using the body coil that is intrinsic to the magnet itself. Anatomic classification, multimodality treatment, and imaging endpoints and limitations relating to perianal fistulas are outside of the focus of this work focusing on CTE and MRE in small bowel Crohn’s disease and can be found elsewhere (62); however, it should be the expectation that every CTE and MRE examination image the entire anus, and that the presence or absence of perianal disease be evaluated by the radiologist along with other imaging findings. In clinical care, gastroenterologists are often most interested in the presence or absence of a perianal fistula or abscess; detailed fistula anatomy is often not required. In the absence of an abscess, therapy with immunosuppressive or biologic medications can proceed, whereas an abscess will require antibiotic treatment and/or drainage before the initiation or continuation of therapy, depending on its size. The question of the presence or absence of an anorectal abscess can typically be answered with CTE or MRE. Because dedicated imaging of the anus for multimodality treatment of perianal Crohn’s disease requires additional pulse sequences centered about the anus (and not included in the MRE examination), dedicated perianal MR imaging should be performed when clinically indicated as noted in a recent global consensus statement (63). Some institutions offer combined MRE and dedicated perianal examinations in patients with known small bowel and perianal disease. It should be noted that perianal disease is not considered penetrating disease in either this guideline or the Paris classification (27). The mechanism of perianal disease is distinctly different than that of classic penetrating disease (64). In addition to the anus and colorectum, radiologists should carefully inspect the appendix, as it is frequently involved with ileocolonic Crohn’s disease (65,66), and appendicitis is rarely the first presentation of Crohn’s disease. Imaging findings of appendiceal Crohn’s disease involvement are similar to those in the small bowel, and ileal-appendiceal fistulas are consequently not uncommon.

Figure 8:

Download as PowerPoint

The spectrum of mesenteric vein thrombosis or occlusion has recently been described in Crohn’s disease patients (67,68). Radiologists should evaluate for and distinguish between acute mesenteric thrombosis and sequela from prior thrombosis, sometimes referred to as chronic mesenteric vein thrombosis, but more accurately termed chronic mesenteric venous occlusion. Acute portal and superior mesenteric vein thrombus can be seen in Crohn’s disease patients as a hypoattenuating thrombus, expanding the vein. These thrombi have been observed to generally resolve without anticoagulation. However, peripheral mesenteric venous thrombi frequently evolve into chronic peripheral mesenteric venous occlusion on follow-up imaging, with segmental pruning of the mesenteric arcade with development of collateral pathways or small bowel varices. Chronic peripheral mesenteric venous occlusions typically correspond anatomically to small bowel segments with active or prior Crohn’s disease inflammation (Figure 9). Coronal imaging with maximum intensity projections are especially helpful in visualizing the mesenteric venous arcade. Acute and chronic mesenteric venous thrombosis/occlusion have been correlated to increased risk for stricture or surgery in a retrospective series (68), but their impact on the natural history of disease is poorly understood.

Figure 9:

Download as PowerPoint

Table 3 lists extra-intestinal findings related to Crohn’s disease (or Crohn’s disease therapies) that should be searched for in every CTE and MRE examination. The most clinically important findings are sacroiliitis, primary sclerosing cholangitis, and avascular necrosis, most often involving the femoral heads. Many patients with Crohn’s disease complain of low back pain. Identifying the changes of sacroiliitis identifies the cause and facilitates therapy. Early primary sclerosing cholangitis is often first identified on enterography, and is manifest by the presence of discontinuous, intrahepatic bile ducts that do not connect to nondilated central ducts. Once identified, the patient can be followed more closely for complications of primary sclerosing cholangitis, typically with MR imaging/magnetic resonance cholangiopancreatography and endoscopic retrograde cholangiopancreatography. Lastly, identifying avascular necrosis will again assist in the care of a patient with hip pain and prompt avoidance of steroids when possible.

Table 3 Extra-Intestinal Findings Relevant to Crohn’s Disease and Seen at Computed Tomography and Magnetic Resonance Enterography

Characterization of Disease Activity

Table 4 lists recommendations for clinical practice based upon the evidence for specific imaging findings. Each recommendation is accompanied by a description of the strength of the recommendation (ie, strong vs weak), with strong recommendations having anticipated desirable effects on patient outcomes (35). These recommendations set forth imaging criteria for the imaging diagnosis of Crohn’s, as well as describing its severity and complications at CTE and MRE. Furthermore, they recommend cross-sectional enterography be performed at diagnosis to detect small bowel involvement that may not be identified by other methods (Figure 10), and recommend it be considered in disease monitoring when small bowel disease or penetrating complications are present (Figure 8). The selection of CTE or MRE will vary according to a variety of factors, including patient preference, age and clinical presentation and concerns, imaging availability, and local expertise, and have been addressed, in part, in practice parameters published jointly between radiology societies (9,69). Potential factors to consider in selecting CTE or MRE as the most appropriate examination for an individual patient are listed in Table 5. MRE is generally preferred in the pediatric population, although CTE is an acceptable alternative, and some practices perform CTE at time of diagnosis. The imaging findings of Crohn’s disease at CTE and MRE are identical between pediatric and adult patients (70,71).

Table 4 Recommendations for Use of Computed Tomography or Magnetic Resonance Enterography, and Incorporation of Imaging Findings Into the Clinical Report

Note.—Strong recommendation indicates confidence that the desirable effects of the test or interpretation will result in a positive impact on patient care. Weak recommendation indicates that uncertainty exists relating to the positive and negative impacts on patient care.

Figure 10:

Download as PowerPoint

Table 5 Potential Considerations for Selecting Computed Tomography or Magnetic Resonance Enterography in a Crohn’s Patient

Table 6 lists recommended impressions in radiology reports for summarizing imaging findings and grouping them into recognized patterns of disease in a manner that is useful to referring physicians, and accounts for exacerbations and response to therapy as seen at cross-sectional enterography (Figure 1). This imaging-based morphologic construct comes from an observation of the dynamic nature of Crohn’s inflammation. As observed by Cosnes and Lemann (26,72), active inflammation is thought to eventually progress to stricturing and penetrating disease complications in a high proportion of patients, with some patients presenting with penetrating or stricturing disease complications, which may portend a more aggressive course. With mild inflammation, wall thickening and hyperenhancement is often seen without luminal narrowing. As inflammation progresses and becomes more severe, enterographic images may display increased intramural T2 signal, restricted diffusion, and ulcer formation in conjunction with luminal narrowing. Adoption of a consistent and well-defined reporting mechanism that links imaging findings of inflammation, stricturing disease and penetrating complications with estimates of disease severity will facilitate selection of optimal therapies and communicate disease progression and reversibility (73), and directly parallel similar linkages provided in the Lemann index without the onerous per-segment analysis required by the research tool (^,Appendix 2) (73).

Table 6 Recommended Impressions Summarizing Imaging Findings of Small Bowel Crohn’s Disease at Computed Tomography and Magnetic Resonance Enterography

Note.—Colonoscopy is considered the reference standard for colorectal inflammation. Recommendations for CTE and MRE descriptions of colorectal inflammation are not provided, but can parallel descriptions of small bowel inflammation, stricture, and penetration.

AVN, avascular necrosis; MRI, magnetic resonance imaging; PSC, primary sclerosing cholangitis.

Several terms should be used in describing the pathophysiological significance of imaging findings associated with current or prior small bowel inflammation. Active Crohn’s disease inflammation should be identified based on the predefined criteria, as should nonspecific inflammation. Active inflammation may respond to medical therapy. When no imaging findings of active inflammation are identified in patients with suspected Crohn’s disease, this should be explicitly stated in the radiologic report. Complete resolution of small bowel or colonic inflammatory findings can occur in Crohn’s disease patients, with the bowel returning to a normal appearance. In these cases, it is also correct to report that no small bowel inflammation is seen. Partial response to medical therapy may be indicated by a decrease in the severity of imaging findings within an inflamed segment, or evolution to much shorter and patchy areas of involvement over the length of the involved segment (Figure 1) (13). Alternatively, inflammation may resolve with residual findings, such as asymmetric fat deposits within the small bowel wall, residual pseudosacculation and scarring, or mild wall thickening without luminal narrowing, or other morphologic or signal changes reflecting active inflammation (ie, absent T2 signal hyperintensity, hyperenhancement, restricted diffusion). When sequelae of prior inflammation are present without active inflammation, “Crohn’s disease with no imaging signs of active inflammation is present” should be stated in the conclusion of the report. Terms such as quiescent or chronic are discouraged because their meaning may be erroneously interpreted, especially by patients who now, in many institutions, have access to their imaging reports. Gastroenterologists and patients making clinical decisions based on imaging findings should be aware that active vs inactive disease based on imaging criteria does not always equate to histologically, endoscopically, or clinically active or inactive disease. There is a relationship between these assessment modalities, but the properties assessed with different modalities vary.

Stricture formation occurs when there is focal or segmental luminal narrowing with unequivocal upstream dilation. Imaging findings of concomitant active inflammation are most often present (53), and we have termed this pattern stricture with findings of active inflammation (Figures 1 and 4). Strictures without imaging findings of inflammation may also exist. In this situation, the bowel wall is thickened without other imaging findings of inflammation. Adler et al (31) found that strictures without imaging findings of inflammation had less inflammation and less fibrosis, but lack of imaging findings of inflammation did not imply that histologic inflammation was absent. While there is a paucity of published data on the subject, in the experience of the radiologist coauthors, penetrating disease has not been seen to arise in the setting of a stricture without inflammation. Imaging criteria for fibrosis are currently being developed and evaluated (54,56).

Internal penetrating disease (not perianal disease) may occur at any time point during the course of the disease, but occurs overwhelmingly in patients with strictures associated with active inflammation. Sinus tract and fistula formation, abscess, and free perforation are all findings of penetrating disease. Fistulas may be simple or complex. Simple fistulas are comprised of a single tract connecting a bowel loop to adjacent bowel or other structures, such as the urinary bladder. Complex fistulas connect multiple adjacent bowel loops or structures. With both simple and complex fistulas, the bowel loops affected are often angulated and appear tethered by the fistula tract (Figure 7). Furthermore, it is common to see small, interloop abscesses along the course of these complex fistulas. If no active inflammation is associated with a fistula, this should also be stated. Postoperative fistulas are often not associated with inflammation, but obviously arise at or near the site of anastomoses.

Colonoscopy is considered the reference standard for colorectal inflammation. This guideline only addresses small bowel Crohn’s disease and complications frequently seen on CTE and MRE in these patients. A comprehensive guide for describing colorectal inflammation at cross-sectional imaging is beyond the scope of this work, as we considered CTE and MRE as part of an imaging strategy combined with clinical assessment and ileocolonoscopy. Transabdominal ultrasound (with or without intravenous contrast) and video capsule endoscopy are used at many institutions in the diagnosis and surveillance of Crohn’s disease, and their role in clinical management continues to evolve; however, integration of their imaging findings is also beyond the scope of this work, which focuses exclusively on CTE and MRE for small bowel Crohn’s disease.

Structured Reporting

Structured reporting templates are used by many radiologic practices for specific clinical scenarios to insure important clinical information is always captured in a systematic fashion. They have been shown to improve the quality of information conveyed to referring clinicians (74). Several groups have advocated for structured reporting for CTE and MRE. Table 7 demonstrates a structured cross-sectional enterography report and is adapted from Baker et al (7).

Table 7 Suggested Reporting Template Adapted From Baker et al

Note.—Adapted From Baker et al (7). AVN, avascular necrosis; PSC, primary sclerosing cholangitis.