The advent of powerful gradient systems resulted in substantial improvement of image quality in ultrafast MR imaging and therefore novel clinical applications emerged including gastrointestinal (GI) tract imaging. Within this context, MR imaging of the small and large bowel were developed, providing luminal, transmural and extramural diagnostic information of intestinal diseases. Current clinical applications of MR imaging of the small intestine include diagnostic evaluation and follow-up of patients with inflammatory, systemic and neoplastic diseases, wheras MR Colonography can be performed in patients with incomplete colonoscopies. Contrast Agents Various contrast agents have been proposed for bowel MRI applications (1, 11-22). The most important characteristics of an intraluminal contrast agent suitable for gastrointestinal (GI) applications may be summarized into the following: uniform and homogeneous lumen opacification, high contrast resolution between the lumen and bowel wall, no significant adverse effects and low cost. In addition, minimal mucosal absorption and absence of artifact formation are highly desirable for an optimal contrast agent. GI contrast agents can be classified according to their effect on MR images into positive, negative and biphasic. Positive contrast agents such as ferrous ammonium citrates, manganese chloride or iron phytate, produce increased intraluminal signal intensity, while negative contrast agents such as superparamagnetic iron oxide (SPIO), result in decreased signal intensity of the bowel lumen. On T2-w images the uniform distribution of the negative contrast within the small and large bowel allows easier visualization of the dark bowel loops within the hyperintense mesenteric fat tissue, thus improving the overall image quality. Biphasic contrast agents, such as methylcellulose, mannitol, sorbitol or polyethylenoglycol (PEG) water solutions, behave as positive or negative, depending on the applied pulse sequence. Small Bowel Route of contrast administration MR Enteroclysis is the only technique that provides sufficient luminal distention to guarantee accurate individual lesion detection. According to the latter technique, patients are examined in the prone position, utilizing a phase-array coil. A total amount of 1,500 to 2,000 ml of an isoosmotic water solution (PEG) is administered through a nasojejunal catheter, by means of an MR compatible pump. A controlled infusion is employed and the contrast is administrated in two phases. At first, a flow rate of 80 150 ml/min is utilized untill the contrast reaches the terminal ileum. In the second phase, the flow rate is increased up to 300 ml/min in order to create reflex atony. Pulse sequences Optimum small bowel imaging should include fast and ultrafast pulse sequences. The spatial resolution of these sequences should be high enough to permitt demonstration of small lesions, such as ulcers or mucosal nodularity, usually present in bowel diseases. Inherent poor signal to noise ratio (SNR) of these sequences must be increased to result in clinically acceptable image quality. All these requirements can be fulfilled by using high-end MR scanners, with field strength of at least 1.5 Tesla, that can provide higher SNR. Short repetition and echo times, which are of great importance in ultra fast imaging, can be only achieved by using advanced gradient systems. Dedicated abdominal phased-array RF coils should be utilized to further increase the limited SNR of the ultrafast pulse sequences. MR imaging examination protocols of the small bowel usually comprise T1- and T2-w sequences in axial and coronal planes. Both T1- and T2-w sequences should be fast enough to allow comfortable breath-hold acquisition times and reduce the motion related artifacts. For T1-w images, most authors are using gradient echo sequences in 2D and 3D acquisition modes with or without fat saturation prepulses, wheras for T2- w images, TSE and HASTE sequences are commonly employed. More recently, the true FISP sequence has been successfully applied in bowel imaging, providing high resolution images of the bowel wall and additional information from the mesentery. Fat-suppressed TSE or STIR sequences have been also applied to assess the activity in Crohns disease. A 3D version of Spoiled Gradient Echo (SGE) sequences was recently introduced. As opposed to 2D SGE, 3D SGE sequences provide increased through-plane and in-plane spatial resolution by obtaining acquisition of thin partitions (2mm) and high matrices (512), respectively. In addition, they offer higher SNR comparing to the 2D SGE sequences. The acquisition time for covering the whole small bowel is 22-25 sec. and it can be further reduced by employing slice interpolation techniques, such as VIBE, which is becoming a promising one in abdominal imaging. The combination of such sequences with positive intraluminal contrast agents, results in images, which may be used to generate virtual endoscopic views. The major disadvantage of the 3D FLASH sequence is the increased sensitivity to motion artifacts that may cause blurring of the intestinal wall; administration of antiperistaltic drugs can overcome this drawback. The single shot variant of TSE sequence with half Fourier technique, the so called HASTE sequence, generates heavily T2-w images maintaining signals from solid tissues, although with lower resolution. The acquisition time can be as short as 1 sec per slice resulting in minimal respiratory related artifacts. Normal intestinal wall exhibits low SI, while inflammatory or neoplastic lesions exhibit high SI. The long echo train used in HASTE sequence makes it less sensitive to susceptibility artifacts, which may appear in gradient echo sequences due to presence of intraluminal air. Moreover it is not sensitive to chemical shift artifacts and therefore it can be used for accurate quantification of intestinal wall thickness. Sufficient reduction of the endoluminal SI, provided by the use of a negative contrast agent, results in depiction of bowel wall abnormalities with high conspicuity. In case of positive endoluminal contrast agents, HASTE sequence is sensitive to intraluminal flow voids related to peristaltic motion. This problem may be reduced when acquiring HASTE images after spasmolitic drug administration. Another limitation of the HASTE sequence is the poor demonstration of the mesenteric structures due to k-space filtering effects. Tissues with short T2 relaxation constant, such as lymph nodes and fibrous tissue, are missing the high order spatial frequencies thus resulting in a blurring effect due to the unique way that k- space is filled in HASTE sequence. Large Bowel MR imaging of the large bowel should fullfill the same technical requirements as for imaging the small intestine. The need for a larger field of view requires the use of at least two abdominal array coils. Patients should undergo bowel cleansing prior to the examination. Novel approches, like feacal tagging, show that this prerequisite may become unecessery. Usually the large bowel is distended using a maximum amount of 2-3 liters of pure water or gadolinium-spiked water solution, through rectal administration. Antiperstaltic drugs, like scopolamine, should be used to minimize peristaltic artifacts. The so called dark-lumen MR Colonography technique is a combination of administration of pure water that renders the colonic lumen with low signal intensity and i.v. gadolinium injection, which results in high SI of the colonic wall, leading to this double contrast appearance. Bright lumen MR Colonography incorporates the administration of gadolinium-spiked (<2%) water solution. In such an approach i.v. gadolinium administration is not necessary. However, two acquisitions should be performed in supine and prone position to differentiate between residual stool or air and polyps, since both present as filling defects. CLINICAL APPLICATIONS Small Bowel Crohns Disease The role of imaging has nowdays expanded to incorporate classification of Crohn disease subtypes. Accurate classification based on MR imaging findings can be achieved when using a technically demanding MRI examination protocol. MR Enteroclysis (MRE) has shown to be highly sensitive to demonstrate superficial, mural and extramural lesions in patients with Crohn disease. Subbtle lesions such as mucosal nodularity, superficial ulcerations and thickening of the folds may be depicted by MRE, although to a lesser extent as compared to conventional enteroclysis, due to its lower spatial resolution. Using true FISP images, MRE can demonstrate the characteristic descrete ulceration of Crohns disease; deep linear ulcers appear as thin lines of high signal intensity, longitudinally or transversely (fissure ulcers) oriented within the thickened bowel wall. Cobblestoning can also be appreciated on MRE images, as patchy areas of high signal intensity, sharply demarcated, along affected small bowel segments. True FISP images are superior to HASTE in demonstrating linear ulcers or cobblestoning and intramural tracts, wheras 3D FLASH images are less sensitive. Wall thickening is clearly shown by all MRE sequences, provided that the small intestinal lumen is adequately distended. Thickened wall in the absence of extensive edema exhibits low to moderate signal intensity on true FISP and HASTE images. Accurate measurements of bowel wall thickening and estimation of the length of involved segment can be performed on MRE images. Luminal narrowing and associated prestenotic small bowel dilatation are easily recognized with all sequences. MRE was in full agreement with conventional enteroclysis in detecting, localizing, estimating the length of all involved small bowel segments and in assessing thickening of bowel wall, luminal narrowing or high grade stenosis in one series. MRE has reported to have a clear advantage over conventional enteroclysis in the demonstration of exoenteric manifestations or complications of Crohns disease. The extent of fibrofatty proliferation and its fatty or fibrotic composition can be assessed on true FISP images, while it can be only suspected on conventional enteroclysis. Fibrofatty proliferation may present with space-occupying lesion characteristics, separating and/or displacing small bowel loops. The involved mesentery may contain small lymph nodes, mostly less than 8 mm in diameter, easily detected on true FISP images by their low signal intensity against the bright mesenteric fat. Such lymph nodes are not clearly demonstrated on HASTE images, due to k-space filtering effects or on 3D FLASH images, due to saturation of mesenteric fat signal. Sinus tracts and fistulas are demonstrated by the high signal intensity of their fluid content on true FISP and HASTE images, but they may be overlooked on the 3D FLASH images, due to limited contrast resolution with surrounding tissues. Abscesses can be recognized by their fluid content and wall enhancement. Large Bowel Although colorectal cancer screening has been advocated as a potential clinical application of MR Colonography, currently it is limited to patients who have undergone incomplete endoscopic colonoscopy (58). The absence of radiation exposure is an attractive feature of MR Colonography for screening applications. However, technical improvements in terms of spatial resolution and minimization of artifacts are still pending. MR Colonography may have a role in staging colorectal carcinoma. As shown in a recent report, tumors were demonstrated in all cases and in 57% of them breach of the muscularis propria was correctly predicted. Future applications of MR Colonography in staging colorectal carcinoma may be more appealing when combined with state-of-the-art MR imaging of the liver (59). Like for staging, MR Colonography can be also used for postoperative surveillance. In case of inflammatory bowel disease, MR Colonography might offer an alternative technique to colonoscopic evaluation for documenting the extent of disease and assessing disease activity (60). In patients with ulcerative colitis, estimation of disease extension using morphologic criteria known from conventional studies, has shown to be feasible (61). It has been additionally demonstrated that the use of a negative superparamagnetic oral contrast agent provides results comparable to endoscopy in the assessment of activity of ulcerative colitis (62). Moreover, wall thickness measurements are contributing to differentiate patients with mild to moderate disease from those with severe degree of clinical-endoscopic activity. CONCLUSIONS MR imaging is presently contributing to diagnostic assessment of GI diseases. The most important advantages of MR Imaging include superb soft tissue contrast, ability for functional information, direct multiplanar imaging and lack of radiation exposure. Adequate bowel distention, homogeneous lumen opacification, fast sequences with breath hold acquisition times, both T1- and T2-w imaging and contrast enhancement are cornerstones for an optimal MR imaging examination of the GI Tract. Especially, a comprehensive MR Imaging protocol for the small intestine should comprise SSTSE, true FISP, HASTE and fat suppressed 3D FLASH sequences. SSTSE is utilized for monitoring the infusion process and performing MR fluoroscopy, while true FISP and HASTE are mainly used for anatomic demonstration and detection of the pathology. 3D FLASH sequences after I.V. gadolinium injection may aid tissue characterization. Inflammatory or neoplastic diseases, including intestinal wall abnormalities, exoenteric disease manifestations and complications, disease activity and to a lesser extent, mucosal abnormalities can be appreciated on MR Imaging. There are strong indications that staging of gastric cancer and assessment of gastric motility can be reliably performed with MR Imaging. Currently, MR Colonography is considered as an alternative technique to CT Colonography and its most important clinical indication is in patients with incomplete endoscopy. Further technical improvements in terms of spatial resolution may increase its potential role in colorectal cancer screening programs.