MRI of the Small Bowel: Enteroclysis

Key Points

MR enteroclysis is an emerging technique for small bowel imaging combining the advantages of conventional enteroclysis with those of cross-sectional imaging. MR enteroclysis is equal to conventional enteroclysis in detecting, localizing, and estimating the length of involved small bowel segments. Early lesions such as thickening and distortion of the val-vulae conniventes and superficial-type ulcers are clearly demonstrated on conventional enteroclysis but they are not consistently depicted by MR enteroclysis, due to its inadequate spatial resolution. The valvulae conniventes are shown in their best advantage and distortion of the mucosal folds are easily detected by MR enteroclysis. The characteristic discrete longitudinal or transverse ulcers of Crohn's disease can be demonstrated on MR entero-clysis, guaranteed by satisfactory distention and opacification of the bowel. Cobblestoning is caused by the combination of longitudinal and transverse ulceration, and is easily shown by MR enteroclysis. Bowel wall thickening is clearly shown by all MR enteroclysis sequences. Bowel wall thickness and the length of small bowel involvement can be measured on MR enteroclysis images. Narrowing of the lumen and associated prestenotic dilation are easily recognized on MR enteroclysis images by all sequences. Exoenteric manifestations of the disease are demonstrated in detail on true-FISP images due to the high contrast generated from the bright mesenteric fat. Disease activity can be best appreciated on postgadolinium fat-saturated 2D/3D T1-weighted sequence by the characteristic enhancement patterns of thickened small bowel wall and degree of enhancement of mesenteric lymph nodes compared to adjacent vessel enhancement.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

Advances in MRI hardware and software have allowed rapid acquisition of high-resolution images of the gastrointestinal (GI) tract, upgrading the diagnostic role of MRI for the imaging evaluation of patients with intestinal diseases (Gourtsoyiannis et al. 2000, 2001; Umschaden et al. 2000; Prassopoulos et al. 2001). Imaging of the entire small bowel may be challenging and important in view of problematic access of endoscopic methods. Optimal luminal distention is of paramount importance for accurate diagnosis and currently the only technique that can fulfill this requirement is enteroclysis. The combination of ultra-fast MRI sequences and duodenal intubation is known as MR enteroclysis (Gourtsoyiannis et al. 2000; Umschaden et al. 2000). Morphologic evaluation of small intestinal diseases, as well as functional information, can be obtained with state-of-the-art MR enteroclysis examinations. The most important advantages of MR enteroclysis include: (a) absence of radiation exposure, (b) rich soft tissue contrast through the utilization of multiple contrast mechanisms, and (c) three-dimensional imaging capabilities. The technical aspects, clinical applications, and limitations of MR enteroclysis are summarized in this chapter.

MR Enteroclysis Technique

Duodenal Intubation

Duodenal intubation ensures adequate small bowel distention, which is considered a prerequisite for iden-tificat ion of intest inal abnor malities (Gourtsoyiannis et al. 2000; Umschaden et al. 2000). In our institution we currently use the 13 French Maglinte enteroclysis catheter (MEC, Cook, Bloomington, IN). Positioning of the nasojejunal catheter is monitored fluoroscopi-cally. The range of intubation time is 2.5–3.5 min, while radiation dose is kept to a minimum by good collimation, low fluoroscopy current and by using short periods of intermittent fluoroscopy. Although in the beginning MR enteroclysis was performed in association with conventional enteroclysis (CE) (Papanikolaou et al. 2002b), nowadays the standalone MR enteroclysis examination is the mainstream approach (Gourtsoyiannis et al. 2006; Masselli et al. 2009). When MR enteroclysis follows CE, certain adjustments in the selection of the appropriate contrast agent, the amount and route of its administration are important for both examinations to be successful. The patient's tolerance to intubation is related to the examiner's experience and can be considered a drawback of MR enteroclysis. Conscious sedation is advocated in USA (Kohli and Maglinte 2009) to overcome the patient's discomfort during duodenal intubation for performance of CT entero-clysis. Local anesthesia of the pharynx and spending time explaining the procedure in detail to the patient have been proven in our everyday practice equally useful measures to minimize discomfort. Oral contrast administration without intubation is performed to reduce the patient's discomfort (Lomas and Graves 1999) (see Chap. 8). However, small bowel distention is usually not sufficient, primarily at the proximal small bowel. A collapsed or not adequately distended small bowel loop may mask minimal abnormalities or may result in false-positive findings. Without fail, a detailed evaluation of the small intestine requires luminal distention that can be guaranteed by intubation.

Intraluminal Contrast Agents

Intraluminal contrast agents are essential for both lumen opacification and distention. A suitable contrast agent should provide homogeneous opacification throughout the entire small bowel lumen, clear differentiation between the lumen and the bowel wall, and be characterized by minimal mucosal absorption, absence of artifacts formation, no severe adverse effects, and low cost (see also Chap. 3). Various media have been proposed (Table 9.1) as optimal intraluminal contrast agents for small intestinal imaging that can fulfill the previous criteria, but there is no consensus on the ideal one, at present. Positive contrast agents render the intestinal lumen with high signal intensity. Negative intraluminal contrast agents result in low intraluminal signal intensity, while biphasic agents are affecting signal intensity depending on the pulse sequence used, usually low on T1- and high on T2-weighted images. These agents are considered more suitable for small bowel imaging, to our experience. Polyethylene glycol water solution with electrolytes is an iso-osmotic, biphasic contrast agent with excellent performance in MR enteroclysis (Gourtsoyiannis et al. 2000, 2001).

Table 9.1. The most important intraluminal contrast agents classified according the endoluminal signal intensity changes they induce

A controlled infusion is important for a successful MR enteroclysis. Initially a flow rate of 80–120 mL/ min is utilized until the contrast reaches the terminal ileum. Subsequently, the flow rate increases up to 200 mL/min to achieve reflex atony that facilitates acquisition of images with minimal motion artifacts. A home-made pump consisting of plastic pneumo-colon adapted with plastic tubes can be used to avoid problems with the magnetic field interferences. Alternatively a dedicated enteroclysis pump can be used positioned outside the MRI room.

Patient's Position

Patients can be examined either in prone or supine position. The former is suggested because it exerts mild pressure to the anterior abdominal wall facilitating separation of the small bowel loops, while it decreases the volume of peritoneal cavity to be imaged. As supine position is more comfortable, MR enteroclysis can be performed this way in older and frail patients.

Pulse Sequences

A comprehensive MR enteroclysis examination protocol should include both T1- and T2-weighted images as for example spoiled gradient echo (i.e., FLASH) and single shot turbo spin echo (SSTSE; i.e., HASTE) sequences, respectively (Table 9.2). Incorporating different MR sequences is the most effective way to demonstrate anatomy, identify and characterize abnormalities, and disclose associated extraintestinal manifestations. Sequences should be fast enough to permit comfortable breath-holding; high-performance gradient systems are important for this purpose. Image quality, irrespectively of pulse sequences applied, can significantly be improved when using abdominal phased array RF coils by increasing the signal-to-noise ratio.

Table 9.2. Scan parameters of different sequences utilized in MR enteroclysis protocol

Acquisition of SSTSE images (slab thickness 7–10 cm, TR: infinite, TE: 1,200 ms, scan time: 3 s) precedes the main examination to monitor the infusion process. Projectional SSTSE images of the small bowel have been applied for demonstrating small bowel obstruction (Umschaden et al. 2000). Faster versions of SSTSE sequence (Makki et al. 2002) may be used for MR fluoroscopy thus providing information about small bowel motility (see also Chap. 14).

True-FISP sequence (TR: 6 ms, TE: 3 ms, flip angle: 70°, slice thickness: 4 mm, scan matrix 256 or 512 and scan time 1.5 s per slice) is a cardinal sequence for MR enteroclysis. It provides motion-free, high-resolution, “T2-like” images of the small intestine, and the mesenteries in a few seconds. The normal bowel wall and valvulae conniventes exhibit moderate signal intensity on true-FISP images, while the intraluminal fluid and the extraluminal fat show high signal intensity favoring the delineation of the bowel wall (Fig. 9.1). Demonstration of the mesenteries is excellent using true-FISP sequence. The high signal intensity of the mesenteric fat provides the ideal background for the depiction of small anatomic structures; small lymph nodes and thin mesenteric vessels including the vasa recta are clearly seen with low signal intensity, especially when a 512 matrix is utilized. The sequence is prone to specific type of artifacts; susceptibility artifacts from trapped air and black boundary artifacts along to the external small bowel wall surface due to chemical shift phenomena may be seen (Fig. 9.2a). However, black boundary artifacts can be clearly differentiated from abnormal bowel wall thickening. The capability to add fat saturation pre-pulses in the true-FISP sequence may also be used to overcome black boundary artifacts (Fig. 9.2b). The acquisition of multiple images over time at the same slice in a dynamic manner permits the evaluation of local and global small bowel motility patterns. This sequence is termed cine-true-FISP and it has shown to be of help in the differentiation between “soft”, inflammatory type of stenosis and fibrostenotic strictures in patients with Crohn's disease.

Fig. 9.1.

High-quality images of the small intestine can be acquired by combining the true-FISP sequence with PEG as an intraluminal contrast agent

Fig. 9.2.

Coronal true-FISP images in a patient with Crohn's disease without (a) and with fat-saturation prepulses (b). In the native true-FISP image the presence of black boundary artifact is evident in voxels were both fat and water protons are coexisting (arrows). By adding fat-saturation prepulses (b), the black boundary artifacts may be eliminated, permitting accurate wall thickness analysis

The most commonly used sequence in small bowel protocols is the HASTE sequence (Prassopoulos et al. 2001; Lee et al. 1998). It can provide heavily T2-weighted images with high contrast resolution in less than 1 s per slice (Fig. 9.3). Imaging parameters may include TE of 90 ms, infinite TR, 4–6 mm slice thickness, 256 × 256–512 scan matrix and scanning time appropriate for breath-hold acquisition of images. HASTE images does not suffer from susceptibility or chemical shift artifacts but are prone to highorder motion usually manifested as intraluminal flow voids (Fig. 9.4a) or blurring of small bowel wall edges (Papanikolaou et al. 2002a). To overcome this limitation administration of antiperistaltic drugs – as for example 1 mg of glucagon – should precede acquisition of HASTE images. The normal bowel wall is clearly seen on HASTE images with low signal intensity (Fig. 9.3), as opposed to the high signal intensity of the thickened bowel wall that may be observed in patients with active Crohn's disease. This sequence is not appropriate for imaging evaluation of the mesenteries. K-space filtering effects may obscure small anatomic structures and result in decreased definition of larger vessels or lymph nodes (Fig. 9.5a).

Fig. 9.3.

Coronal HASTE image demonstrating T2 contrast and clear depiction of the intestinal wall. Notice the loss of information from the mesenteries due to selective spatial blurring effects

Fig. 9.4.

MR enteroclysis: coronal HASTE (a) and true-FISP (b) images. HASTE is suffering from high-order motion due to peristaltic waves presented as signal voids (arrows), while true-FISP sequences are rather insensitive to such artifacts due to flow compensation capabilities in all three axes

Fig. 9.5.

Coronal HASTE (a) and true-FISP (b) images in a patient with Crohn's disease. Mesen-teric lymph nodes (arrow) and vessels are poorly depicted with the HASTE sequence due to K-space filtering effects

High spatial resolution T1-weighted images of the small bowel can be obtained applying 3D or 2D spoiled gradient echo sequence (Gourtsoyiannis et al. 2001; Gourtsoyiannis and Papanikolaou 2008). Fat-saturation prepulses facilitate the demonstration of the bowel wall in combination with a negative intraluminal contrast agent. Thin sections and small pixel size allows for increased definition of the bowel wall. Intravenous administration of gadolinium increases the conspicuity of the normal bowel wall and may permit lesion characterization by evaluating different enhancement patterns. A 3D FLASH sequence with fat saturation may applied with 4.8 ms TR, 1.8 ms TE, 45° flip angle, 2.5 mm slice thickness, 256 × 512 scan matrix and 23 s scan time. Normal bowel wall on contrast-enhanced FLASH images with fat saturation exhibits high signal intensity due to gadolinium uptake and is perfectly delineated between the low signal intensity of the mesenteric fat and the negative intraluminal contrast agent (Fig. 9.6). A scanning delay of 60–80 s post-IV contrast results in maximal contrast between the intestinal wall and the lumen (Fig. 9.7). 3D FLASH sequence is sensitive to motion; consequently, antiperistaltic drugs administration should precede the application of the sequence (Fig. 9.8).

Fig. 9.6.

Normal appearance of the small bowel in postgadolinium coronal FLASH images with fat-saturation pre-pulses. The combination of parallel imaging algorithms and state of the art matrix coils permits the acquisition of 1,024 matrices with sufficient signal-to-noise ratio

Fig. 9.7.

Coronal post-gadolinium FLASH images on different patients acquired at three time points after gadolinium injection at 25 s (a), 75 s (b) and 180 s (c). Optimal contrast between the intestinal wall and surrounding tissues is achieved at 75 s

Fig. 9.8.

Coronal postgadolinium FLASH sequence before (a) and after intravenous administration of glucagon (b). Peristaltic motion present in (a) is resulting in sig-nificant blurring

Incorporation of different MR sequences – such as SSTSE, true-FISP, HASTE and gadolinium-enhanced FLASH – into a comprehensive MR entero-clysis imaging protocol is important to obtain all the information the method can provide, in a balanced fashion, where disadvantages of one sequence are overcome by the advantages of the other (Table 9.3). For example, susceptibility artifacts on true-FISP images may prevent demonstration of a small portion of small bowel wall that can be evaluated on HASTE or FLASH images, while motion artifacts downgrading image quality on FLASH images do not interfere with true-FISP images. Information from images obtained using different contrast mechanisms increases the confidence in lesion detection and characterization, and provides a completed anatomic demonstration of the intestinal lumen small bowel wall and the mesenteries.

Table 9.3. Comprehensive MR enteroclysis examination protocol

Normal Appearences

The use of a biphasic intraluminal contrast agent such as PEG results in high signal intensity of the lumen on true-FISP images and low signal intensity on post-gadolinium FLASH images (Figs. 9.1 and 9.6). A very powerful contrast mechanism that MR is offering is the T2 contrast that is based on differences between the T2 relaxation times of tissues. In T2-weighted images, such as HASTE, normal intestinal wall exhibits moderate signal intensity. Normal bowel wall is uniformly thin not exceeding 3 mm in thickness. Normal intestinal folds are visualized as dark linear structures. HASTE provides high contrast between bowel wall and intraluminal fluid. One should be aware that some ultrafast pulse sequences that are routinely utilized for small bowel imaging, such as true-FISP sequence, deviate in terms of contrast due to T1 interferences. True-FISP is capable of demonstrating normal anatomy and morphology equally well to HASTE; however, the mixed contrast may decrease its potential in tissue characterization. In addition the superb contrast between the bright mesenteric fat and the dark mesenteric vessels or lymph nodes, make it exceptional for the demonstration of extraintestinal anatomy.

In case of postgadolinium FLASH images, normal intestinal wall exhibits moderate homogeneous enhancement while abnormal wall might be expressed by three different enhancement patterns as will be discussed in detail in the next paragraph. Postgadolinium FLASH sequences allow also improved lesion characterization in case of small bowel neoplasms.

Interpretation of Imaging Findings

Crohn's Disease

Crohn's disease is a preferential field for the application of MR enteroclysis. Different pulse sequences may offer different kind of information that can be intergraded to provide an overall view of mucosal, mural, and extramural abnormalities associated with the disease and complementary information about disease activity and complications. These MR entero-clysis features of Crohn's disease are to a large extent identical to the features of MR enterography.

In a recent study (Gourtsoyiannis et al. 2006), a comparison between MR enteroclysis and CE was performed on a sign-by-sign basis in patients with Crohn's disease. A total number of 49 involved segments were disclosed by MR enteroclysis, all of which were con-firmed by CE. There was full agreement between MR enteroclysis and CE in determining the length and the site of involvement of the diseased segments.

Early lesions of Crohn's disease, such as thickening and distortion of the valvulae conniventes and superficial types of ulcers that are clearly demonstrated on CE are not consistently depicted by MR enteroclysis (Gourtsoyiannis et al. 2006). This results in an overall sensitivity of less than 50% for MR enteroclysis detecting superficial ulcerations, most probably due to inadequate spatial resolution. Direct imaging techniques such as video capsule endoscopy (VCE) are obviously performing better in detecting such early mucosal abnormalities. It is speculated that dedicated ultrafast, high-resolution sequences and stronger gradients will further improve the detection of such early, but not specific, manifestations of the disease in the near future.

The characteristic discrete, longitudinal, or transverse ulcers of Crohn's disease can be demonstrated on MR enteroclysis, following optimal distention and homogeneous opacification of the bowel lumen (Fig. 9.9). A sensitivity of about 90% in detecting deep ulcers was achieved in a recent study of 52 patients with suspected or established Crohn's disease, and this is probably related to the optimized technique that was utilized (Gourtsoyiannis et al. 2006).

Fig. 9.9.

Coronal true-FISP with fat saturation after gadolinium administration. Normal mucosal layer is demonstrated as a black thin line due to T2 shortening effects from increased local concentration of gadolinium (black arrows). Discrete ulceration can be discriminated from fold thickening in the basis of preservation of continuity of mucosa. In the area of the discrete ulcer (white arrow) the mucosal layer is disrupted

Cobblestoning, a combination of longitudinal and transverse ulceration, can be easily demonstrated by MR enteroclysis. As in the case of discrete ulcers, a tailored MR enteroclysis examination can guarantee a high sensitivity and specificity of 92.3% and 94.4%, respectively, for the demonstration of cobblestoning (Gourtsoyiannis et al. 2006). A true-FISP sequence is known to be superior to HASTE in demonstrating linear ulcers (Fig. 9.10), cobblestoning, and intramural tracts, due to its superb resolution capabilities and its relative insensitivity to motion artifacts, while 3D FLASH sequences are less efficient in depicting such lesions when they are smaller than 3 mm (Prassopoulos et al. 2001). Bowel wall thickening can be clearly detected by all MR enteroclysis sequences. Mural thickening presenting with moderate signal intensity on true-FISP images can be easily differentiated from misregistration due to the black boundary artifacts (Gourtsoyiannis and Papanikolaou 2008). Bowel wall thickness and the length of small bowel involvement can be measured accurately on MR enteroclysis images. Narrowing of the lumen and associated prestenotic dilation (Fig. 9.11), as well as skip or multiple lesions (Fig. 9.12), are routinely recognized on MR enteroclysis images in all sequences. The addition of cine true-FISP sequences that can be used to demonstrate local motility patterns might reduce any potential false-positive results in stenotic lesions due to bowel spasm.

Fig. 9.10.

Linear ulcer (black arrows) is demonstrated on a patient with active inflammatory subtype of Crohn's disease, on a coronal true-FISP image

Fig. 9.11.

Axial true-FISP image in a patient with fibros-tenotic subtype of Crohn's disease. A fibrostenotic lesion is causing significant luminal narrowing (arrows) and prest-enotic dilation (asterisk). The mural signal intensity is moderate reflecting the presence of fibrosis

Fig. 9.12.

Skip lesions are identified (arrows) both on coronal (a) true-FISP and (b) postgadolinium FLASH with fat-saturation images in a patient with longstanding Crohn's disease

One of the most significant advantages that MR enteroclysis is offering over endoscopic techniques is its capability to demonstrate in the same imaging session exoenteric manifestations of the disease. This can be done in the best of its advantage on true-FISP images due to the high contrast generated from the bright mesenteric fat. Complications such as fistulae, phlegmons (Fig. 9.13) or abscesses may be more accurately diagnosed on T1-weighted FLASH images with fat saturation by the characteristic pattern of enhancement after gadolinium administration (Gourtsoyiannis et al. 2006). The latter sequence is additionally useful in active disease because of the marked contrast uptake in the thickened small bowel wall (Lohan et al. 2007; Makki et al. 2002) and mesenteric lymph node enhancement (Fig. 9.14) (Lee et al. 1998). MR enteroclysis imaging signs that are related to active disease have been proposed (Gourtsoyiannis et al. 2004, 2009; Koh et al. 2001; Florie et al. 2006), and this may represent one of the most important indications for the examination in the near future, however there are no universal agreed criteria for such assessment of disease activity as the protocols of MR enteroclysis or enterography are widely different. According to one study (Gourtsoyiannis et al. 2004), there are criteria, among all possible imaging findings that one encounters on MR enteroclysis in patients with Crohn's disease and in correlation with Crohn's disease activity index (CDAI) that provide a reference to discriminate active from non active disease. These are: (a) more than one transmural ulcer, (b) wall thickness >7 mm, and (c) enhancing mesenteric lymph nodes with a signal intensity enhancement ratio (signal intensity of lymph node/ signal intensity of adjacent vessel) >0.7.

Fig. 9.13.

Patient with active inflammatory Crohn's disease subtype. The presence of phlegmon mass is disclosed both in coronal postgadolinium FLASH image (a) and in coronal true-FISP with fat saturation (b). Note the increased gadolinium uptake in the periphery of the mass due to hyperemic conditions (arrows)

Fig. 9.14.

The presence of intramural fat might simulate the appearance of submucosal edema. In HASTE images (a) both intramural fat and edema are exhibiting high signal intensity (arrow), while on post-gadolinium FLASH images (b) with fat saturation fat and edema presents with low signal intensity (arrows). True-FISP without or with fat saturation might help to discriminate between fat and edema, since fat is generating the typical black boundary artifact (arrow on c), or it presents with low signal on fat-saturated true-FISP images (d), contrary to edema which on the latter images is bright

Intramural fat can be accurately identified when combining features from true-FISP and gadolinium-enhanced 3D FLASH images with fat saturation. Discrimination between the deposition of fat and the presence of edema (Fig. 9.15) may be helpful for the disease classification. Collagen is known to result in late gadolinium enhancement (Semelka et al. 2001). In this context, MR has the potential to differentiate fibrostenotic from edematous lesions on the basis of different gadolinium enhancement patterns.

Fig. 9.15.

Patient with active inflammatory subtype of Crohn's disease. Multiple, enhancing mesenteric lymph nodes (arrows) are easily depicted in a coronal post-gadolinium FLASH image

By assessment of Crohn's disease via MR entero-clysis, certain diagnostic tasks can be accomplished including the demonstration of extent of mural and extramural disease, the calculation of disease activity, and the identification of clinical subtypes, a critical point for further patient management. Maglinte et al. (2003) introduced an imaging-based classification of Crohn's disease subtypes constituting of active inflammatory (AI), fibrostenotic (FS), and fis-tulizing-perforating (FP) subtype. In a recent study (Gourtsoyianni et al. 2009), mesenteric lymph nodes of patients with AI disease presented with the highest enhancement ratio when compared to mes-enteric lymph nodes of patients with FP disease, who presented with moderate enhancement ratio and those in patients with FS who presented with the lowest enhancement ratio. The difference in mean values of enhancement ratio between mesenteric lymph nodes of AI and FS subtypes was statistically significant. Thus, according to this study degree of homogeneous uptake is not the same for all different disease subtypes, and therefore quantification of enhancement ratio was proven useful in clinical practice for disease subtype classification.

MR enteroclysis also allows the assessment of Crohn's disease recurrence after ileocolic resection. Sailer et al. (2008) developed a reproducible MR score which showed high agreement with the endoscopic Rutgeerts score, and according to it medical therapy or surgical intervention can be indicated.

MR enteroclysis is an emerging technique for the assessment of small bowel pathology and its clinical applications is so far limited to centers of reference. The method is complementary to CE in detecting superficial or early Crohn's disease, but it is of equal diagnostic accuracy in disclosing transmural disease. In addition, MR enteroclysis can adequately depict mesenteric involvement and extraintestinal complications of Crohn's disease which is an important advantage over CE. Specific MR enteroclysis imaging features may be used to assess disease activity. Furthermore, accurate individual lesion detection, provided by MR enteroclysis, may successfully address clinical questions related to the classification of Crohn's disease subtypes.

Small Bowel Neoplasms

MR enteroclysis may combine the effectiveness of cross-sectional MRI (Semelka et al. 1996) with the advantages of CE in the detection and characterization of small bowel neoplasms. Small bowel tumors usually exhibit moderate signal intensity on true-FISP images (Fig. 9.16a) as opposed to the high signal intensity of the distended lumen and the mesenteric fat. Post-gadolinium 3D FLASH sequence with fat saturation may be the most important one (Fig. 9.16b) for the identification and characterization of small bowel tumors. The degree of prestenotic dilation, the peritoneal extension of the neoplasm and the associated lymphadenopathy can be appreciated by all MR enteroclysis sequences. The role of MR enteroclysis in small bowel tumors has not fully established, at present, due to the limited experience on that issue. However the technique is very promising and may be considered as the effective approach for imaging of small bowel neoplasms. Due to the excellent disten-tion and the possibility for repeated imaging with different sequences provided by MR enteroclysis it possible to differentiate with greater certainty between collapsed bowel loops and potential masses. In a more recent study (Masseli et al. 2009), the accuracy of MR enteroclysis for identification of small bowel neoplasms in symptomatic patients with a size ranging from 8 mm to 7 cm was found to be 97% when compared to histopathological examination. Another study (Schmid-Tannwald et al. 2009) concentrated in the ability of MR enteroclysis to identify and localize only primary carcinoid tumors of the small bowel which was found to reach 93% when using contrast enhanced T1-weighted fat-saturated GRE sequences as compared to surgical findings. Primary carcinoid tumors appeared as nodular intraluminal masses in 40% of the cases, as focal wall thickening in 33% and in 20% as both (Schmid-Tannwald et al. 2009).

Fig. 9.16.

Coronal true-FISP (a) and post-gadolinium FLASH (b) images are showing a well-defined mass (arrows) with moderate signal intensity on true-FISP and increased contrast uptake. The presence of the mass was confirmed on conventional enteroclysis (c), and proved to be a benign stromal tumor (with permission from Gourtsoyiannis et al. 2001)

Intestinal Obstruction

MR enteroclysis is very effective in determining the presence and the level (Fig. 9.17) of low-grade small bowel obstruction (Lienemann et al. 2000). Initial evaluation of a patient with suspected small bowel obstruction with dynamic SSTSE images is useful for the prompt disclosure of this condition, meanwhile providing information about small bowel motility. In addition, MR enteroclysis may reveal the cause of obstruction. Sequential true-FISP images in a functional cine MRI mode proved highly accurate in small bowel obstruction due to post-surgical adhesions (Lienemann et al. 2000), while post-gadolinium 3D FLASH images may demonstrate other than adhesions causes of obstruction. In patients with obvious signs of high-grade obstruction, an MRI examination without enteral contrast administration can be considered.

Fig. 9.17.

Patient with colonic carcinoma that has invaded a jejunal small bowel loop. (a) Coronal projection single shot TSE image demonstrating dilation of jejunal loops and absence of opacification of distal small bowel loops. (b) Axial true-FISP image disclosing a large mass (T), invading the small bowel, responsible for bowel obstruction

Malabsorption Disorders

Malabsorption states, and especially celiac disease, are another indication for MR imaging of the small bowel (Laghi et al. 2003). The role of diagnostic imaging in celiac disease is usually limited, since the final diagnosis is obtained by jejunal biopsy. Decreased number of jejunal folds, reversed jejunal folds, i.e., decreased number of jejunal folds and increased folds in ileum is indicative of celiac sprue due to total villous atrophy, a finding that can be better appreciated by MR entero-clysis than by MR enterography due to excellent dis-tention of the jejunum. MR enteroclysis might be a useful tool both for diagnosis and follow-up, since it can provide, noninvasively, morphologic information such as bowel dilation, ileal jejunization, jejuno-ileal reversal pattern and, at the same time, extraintestinal findings, such as mesenteric vascular congestion, lymphadenopathy, hyposplenism or the presence of intussusception (Laghi et al. 2003). One should aware of lymphomas in patients with celiac disease.

Limitations and Disadvantages of MR Enteroclysis

MR enteroclysis has not been adequately compared to other techniques for small bowel imaging. Duodenal intubation and inherent very limited radiation exposure are considered drawbacks of MR enteroclysis as compared to MR enterography. Optimal and homogeneous distention throughout the entire small bowel however allow for easier depiction of individual lesions, like discrete ulcerations, avoid misregistration of collapsed or undistended bowel segments, increases confidence in diagnosis for the average practitioner and guarantees better comparison and measurements in follow-up studies after treatment due to consistent optimal image quality. Adequate distention is important to verify the presence of edema in a diseased bowel segment, and therefore aid in patients management in differentiating inflammatory from fibrotic disease. A disadvantage of MR enteroclysis is the limited spatial resolution as compared to CE and multi-detector CT. Temporal resolution, even with functional cine MR mode or dynamic SSTSE, is inferior to the real-time imaging provided by CE and ultrasound. The cost-effectiveness of the method has not been evaluated, yet. A technically successful MR enterocly-sis examination and correct interpretation of its find-ings requires familiarity with the technique.

Conclusions

MR enteroclysis appears very promising for the evaluation of small bowel anatomy, motility, and pathology. Accumulated experience during the past decade shows that MR enteroclysis is well performing in Crohn's disease, tumorous lesions, and small bowel obstruction. MR enteroclysis is a highly specific examination for demonstration of the transmural nature and extent of inflammation. It can depict lesions beyond severe luminal stenosis, can characterize further stenotic lesions (edema vs. fibrosis differentiation), and can identify intraperitoneal extension or extraintestinal manifestations. Various imaging features have been shown to correlate with disease activity, while MR enteroclysis may permit accurate classification in different disease subtypes, affecting directly patient management.