| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Developments |
1 From the Departments of Radiology (D.J.L., R.R.S., M.J.G., A.K.D.) and Surgery (R.M., N.R.H.), University of Cambridge and Addenbrooke’s Hospital, Box 219, Hills Rd, Cambridge, CB2 2QQ, England. Supported in part by a grant from Smith’s Charity. Received April 6, 2000; revision requested May 31; revision received August 23; accepted September 19. Address correspondence to D.J.L. (e-mail: djl15@radiol.cam.ac.uk).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and MethodsResultsDiscussionREFERENCES |
|---|
Index terms: Carbon dioxide • Colon, MR, 75.121411 • Colon neoplasms, 75.311, 75.321
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResultsDiscussionREFERENCES |
|---|
Both these approaches have well-recognized limitations that have driven clinicians to search for improved methods in the detection of and screening for colon cancer. This search has led to several imaging developments, including virtual colonoscopy (based initially on x-ray computed tomography [CT] and the use of intraluminal gas [1,2]), and, more recently, magnetic resonance (MR) imaging and the use of an intraluminal aqueous gadolinium solution (3). These newer imaging techniques allow demonstration and localization of luminal colonic lesions and have particular advantages when compared with the established techniques, but they also have potential disadvantages that may undermine their widespread application in both symptomatic and screening populations.
In particular, the CT method is limited to the transverse acquisition plane and by concerns about the use of ionizing radiation, which tends to minimize data acquisition. This method also depends mainly on interpretation after the examination by using image-processing developments, such as virtual colonoscopy, to improve interrogation of the available data.
The initially proposed MR imaging approach (3) is not limited by concerns about ionizing radiation but by the practical aspects of maintaining adequate bowel distention with the use of a fluid medium. Within the confined environment of current MR imaging systems, this makes the procedure relatively uncomfortable for patients and difficult for the staff to supervise. In addition, a three-dimensional angiographic approach that is designed to generate maximal signal from material with a short T1 (gadolinium-based contrast medium) within the lumen (vascular or colonic) and suppress all the other tissue signal has been used. This approach requires an effective breath hold of up to 30 seconds, as true three-dimensional data acquisition is performed, and directly demonstrates the lumen but not the bowel or vascular wall. These factors, combined with the current limiting design of MR systems, indicate that the potential advantages of MR imaging have yet to be fully exploited.
In this article, we present our initial results with an alternative MR imaging approach to imaging colonic tumors that further exploits the range of contrast that MR imaging can provide and that we consider likely to be more robust than the initially described method. The purpose of this study was to assess the initial technical performance with the use of the technique according to its ability to demonstrate known colonic tumors in a group of patients prior to surgery.
|
Materials and Methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResultsDiscussionREFERENCES |
|---|
Six lesions were eccentric or annular and involved the bowel wall, and one was a polyp; all were subsequently proved malignant at histologic study. The five lesions demonstrated at barium enema examination varied between 23 and 45 mm in maximal length measured along the long axis of the bowel. This was defined as the distance between the proximal and distal points of distortion or contraction of the lumen. The measurements were made with a ruler to millimeter accuracy on the image by using a correction factor for magnification effects on the basis of a typical source-to-intensifier distance. The measurements are the mean of two radiologists’ observations. The polyp demonstrated at endoscopy was estimated to be 25 mm in diameter. There were two cecal, one ascending colon, two hepatic flexure (right colic flexure), and two sigmoid colon tumors. One patient had non–insulin-dependent diabetes mellitus, and another had epilepsy.
The study was approved by the local ethical review committee, and written informed consent was obtained from each patient.
Patients underwent the standard bowel preparation for colon surgery with an orally administered cleansing agent (Picolax; Ferring, Middlesex, England) prior to surgery. MR imaging was performed during the 24 hours prior to surgery, with exact timing being dependent on MR system availability. The examinations were performed by using a commercial whole-body clinical 1.5-T MR system (LX Echospeed; GE Medical Systems, Milwaukee, Wis) with a torso multicoil receiver array. An enema tube (Lafayette, Yorba Linda, Calif) was placed in the rectum prior to the imaging procedure, and the patient was positioned supine on the imager couch.
All imaging was performed by using a half-Fourier single-shot fast spin-echo (SE) sequence. After initial imaging was performed to confirm the correct positioning of the patient, the colon was gently insufflated with CO2 (in prefilled enema bags) by using hand pressure. Colonic filling was observed at imaging every 5–10 seconds with the repeated acquisition of a single coronal 40–60-mm section through the cecum by using the following parameters: 5,000–10,000/96 (repetition time msec/effective echo time msec), 40–60-mm section thickness, 256 x 256 matrix, and 0.5 signal acquired. This was continued until the cecum was seen to fill or the patient complained of discomfort. Coronal and transverse imaging of the distended colon was then immediately performed with the following parameters: 
/28, 5-mm section thickness, no gap, 256 x 192 matrix, 0.5 signal acquired, 34–38-cm field of view, and 12 sections acquired in 15 seconds. Blocks of images of contiguous sections were obtained by using up to five breath holds to image the colon.
These images were reviewed immediately by the radiologist supervising the examination (who was aware of the earlier diagnostic examination results) to see whether the known colonic lesion could be identified and to confirm visualization of the whole colon. Further transverse or coronal images were obtained, if required, to include any part of the colon omitted on the first set of images. In addition, if a segment of colon was not adequately distended, more CO2 was insufflated, or, if residual fluid was present, the patient was turned prone, and then further transverse and/or coronal image sets were obtained. These additional image sets were required in four patients (three for prone imaging and one for additional CO2 insufflation); for prone imaging, additional CO2 was insufflated to maintain distention of the colon.
Specific local imaging of the abnormal colon was performed in all patients by using transverse or coronal contiguous sections with the following parameters: /28, 3-mm section thickness, no gap, 256 x 128 matrix, 0.5 signal acquired, 34–38-cm field of view, and 12 sections acquired in 12 seconds. At the end of the procedure, any remaining gas was evacuated through the rectal tube, which was then removed. The mean procedure time was 45 minutes. The patients were examined by a radiologist (D.J.L.) immediately after the procedure for evidence of any serious complications, such as intestinal perforation, prior to surgery.
All the image sets were then reviewed independently on a workstation by the two radiologists, one of whom had supervised the procedure. Both were aware of the findings of the earlier diagnostic examinations regarding tumor location but were unaware of the subsequent surgical or histologic findings. Planar and multiplanar reformatting was used to detect and assess the extent of the known lesions. The maximal tumor length in the long axis of the bowel was measured by using electronic calipers.
Tumor invasion beyond the muscularis propria was diagnosed if the lesion mass extended into the adjacent intraabdominal fat beyond the boundary along the long axis of the bowel defined by the interface of the contiguous outer bowel wall of the uninvolved segments proximal and distal to the lesion and the adjacent intraabdominal fat. When the lesion caused contraction and reduction of the outer wall diameter, the presence of irregularity at the interface between the intermediate signal intensity of tumor and the high signal intensity of adjacent fat was interpreted as positive for tumor invasion. The presence of any other lesions and luminal contaminants was also assessed. Because the focus of this study was the ability of the technique to depict the tumor itself, liver and nodal status were not assessed.
The surgical notes and pathology reports about the excised tumors were reviewed (by one of the radiologists who reviewed the MR images and one of the surgeons [N.R.H.]) and were compared with the MR findings with regard to the location and extent of the lesions. The surgical notes were reviewed for evidence of gross tumor invasion of adjacent organs by the known lesion and with regard to any other colonic lesions found at surgery. The histologic results for each specimen were reviewed, and breach of the muscularis propria (ie, T3 stage) was interpreted as positive for tumor invasion.
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResultsDiscussionREFERENCES |
|---|
|
|
|
|
|
|
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResultsDiscussionREFERENCES |
|---|
In this work, we have used a different MR approach in the form of a single-shot rapid acquisition and relaxation-enhancement SE sequence that permits the use of gas (CO2) as a negative contrast medium. This sequence can be used to image each section in less than a second (avoiding respiration and peristalsis artifacts) and to provide soft-tissue contrast that allows visualization of the normal and pathologic bowel wall and the adjacent intraabdominal tissues. A particular advantage of this sequence is the train of refocusing SE pulses that confer resistance to susceptibility-induced geometric distortion and dephasing artifacts. This allows successful bowel imaging by using a gas-based medium without the artifacts (7) that are present by using gradient-echo and, in particular, echo-planar sequences.
The use of a negative contrast medium, unlike the use of positive contrast medium, virtually eliminates the risk of image artifacts related to motion within the medium. Gas media are simpler than fluid media to use in the constrained MR system environment, and experience with barium enema examination, endoscopy, and laparoscopic procedures has shown that CO2 is more comfortable for patients than room air because it is relatively rapidly and harmlessly reabsorbed into the blood.
The success of current imaging techniques, including endoscopy, depends on adequate bowel distention and prior preparation of the bowel to prevent inadequate visualization of the lumen or contamination by feces or fluid that may lead to false-positive and false-negative findings. Bowel preparation in this study was not ideal because the MR examinations were performed during preparation for surgery to minimize the inconvenience to the patients. In the future, better preparation and use of orally administered tagging agents may allow improved discrimination between tumors and contaminants. Experience with conventional barium enema examination and CT colonography has shown the value of changing patient posture and of obtaining multiple views of the colon, because overlapping loops of bowel, inadequately distended segments of colon, and barium may obscure the lumen. The same limitations and need for additional views is likely to apply to the newer CT and MR imaging techniques, if they are to be applicable in the general population, given the wide variation of normal anatomy, ability of patients to cooperate with the procedure, and the likely presence of contaminants. In the majority of patients in this study, we obtained additional images of the abnormal area and data sets with patients in both prone and supine positions. In the future, the ability to obtain images with patients in decubitus positions is feasible, although modification of the current MR system coil arrays and couch systems is required to facilitate this.
An intrinsic feature of sequences based on rapid acquisition with relaxation enhancement, the generation of high signal intensity from fat owing to reduced J coupling was exploited to provide contrast differences between fat and the intermediate signal intensity of tumor and the lack of signal intensity from intraluminal gas. Interpretation of the extraluminal extent of tumor partly relied on this contrast difference between tumor and adjacent intraabdominal fat and proved possible in all the patients examined. Variations in body habitus that influence the amount of intraabdominal fat may limit interpretation, particularly in those with reduced intraabdominal fat.
A limitation of the technique presented here is the need to acquire images grouped as sequential blocks over several breath holds, which may lead to misregistration of adjacent blocks of images if breath holding is not consistent. In practice, this did not prove to be a marked limitation, since the data were interpreted by using primarily multiplanar reformatting, with which such misregistration becomes obvious. The use of a multicoil receiver array introduces additional signal intensity variation into the images, which, along with misregistration, may make processing of subsequent virtual endoscopic images more complex. In practice, virtual endoscopic images of the respective lesions could be generated (Navigator software; GE Medical Systems) directly from the image data sets obtained (eg, Fig 3b).
Despite these potential limitations, it was possible, in all the patients in this initial study, to obtain diagnostic images of the primary lesions and to determine their relationships to adjacent tissues and the remaining colon. This was of particular value in one patient in whom complete colonoscopy was not possible owing to a distal colonic stricture and in one in whom prior barium enema examination had been unsuccessful owing to an inability to retain the barium.
In this work, the examination of known and relatively large (Table) colonic tumors may have introduced a bias in favor of adequate visualization of the lesions. Further evaluation with a larger number of patients suspected of having lesions rather than those known to have lesions will be required to confirm and further evaluate these preliminary findings. This initial interpretation caused underestimation of the tumor stage in three of the seven patients; different imaging criteria for tumor invasion may allow this aspect to be improved. The overall morbidity of the technique is likely to be low and less than that of conventional double-contrast barium enema techniques, since the method is similar, but without the use of either barium or x rays. This would be advantageous if the technique were eventually used in a screening program, since there is a recognized morbidity and mortality for other techniques, such as colonoscopy, even with experienced operators.
In the future, the advantages of MR examination of the colon are likely to become more apparent when technical developments allow true MR fluoroscopic capability, more interactive control of the imaging parameters, and easier alteration of patient position. Improved display systems to allow faster interpretation during the examination, rather than retrospectively, will become available. The MR examination could easily be modified to include evaluation of the liver and the mesenteric and paraaortic nodes for staging purposes, maximizing the information derived from a single investigation.
Further research and development of this MR technique are required to address the optimal preparation, applicability, interpretation methods, value of posture change, role of fluoroscopic techniques, time required to achieve adequate visualization of the colon, and accuracy, compared with those of other techniques.
These study findings highlight the intrinsic potential advantages of MR imaging as a future relatively noninvasive method of diagnosing and staging colon carcinoma. The ability of the physician to image in any plane at any depth and to vary the image contrast and the lack of ionizing radiation, which allows serial and repeat imaging, are all important advantages, compared with those of conventional methods. These findings show that an MR imaging technique with the use of CO2 as a negative contrast agent can demonstrate colonic tumors and their relationship to the lumen, wall, and surrounding tissues with a single examination.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, D.J.L.; study concepts and design, D.J.L., R.R.S.; literature research, D.J.L.; clinical studies, D.J.L., R.R.S., M.J.G., R.M., N.R.H.; data acquisition, all authors; data analysis/interpretation, D.J.L., R.R.S., A.K.D.; manuscript preparation, D.J.L., A.K.D.; manuscript definition of intellectual content, D.J.L.; manuscript editing, revision/review, and final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResultsDiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  V L Jardine, E Sala, and D J Lomas MR colonography: baseline appearance of the unprepared rectosigmoid Br. J. Radiol., March 1, 2005; 78(927): 202 - 206. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Ajaj, T. C. Lauenstein, G. Pelster, G. Holtmann, S. G. Ruehm, J. F. Debatin, and S. C. Goehde MR Colonography in Patients with Incomplete Conventional Colonoscopy Radiology, February 1, 2005; 234(2): 452 - 459. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. W. F. Geenen, S. M. Hussain, F. Cademartiri, J.-W. Poley, P. D. Siersema, and G. P. Krestin CT and MR Colonography: Scanning Techniques, Postprocessing, and Emphasis on Polyp Detection RadioGraphics, January 1, 2004; 24(1): e18 - e18. [Abstract] [Full Text]
|
|
|
|
 |
|
 W Ajaj, G Pelster, U Treichel, F M Vogt, J F Debatin, S G Ruehm, and T C Lauenstein Dark lumen magnetic resonance colonography: comparison with conventional colonoscopy for the detection of colorectal pathology Gut, December 1, 2003; 52(12): 1738 - 1743. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J F Debatin and T C Lauenstein Virtual magnetic resonance colonography Gut, June 1, 2003; 52(90004): iv17 - 22. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. C. Lauenstein, S. C. Goehde, S. G. Ruehm, G. Holtmann, and J. F. Debatin MR Colonography with Barium-based Fecal Tagging: Initial Clinical Experience Radiology, April 1, 2002; 223(1): 248 - 254. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
