| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Institute of Diagnostic Radiology (W.L., S.W., B.M., J.F.D.) and the Department of Internal Medicine, Section of Gastroenterology (P.B., M.F.), University Hospital Zurich, Switzerland. Received June 21, 1999; revision requested August 10; final revision received November 9; accepted November 16. Supported in part by German Research Society grant DFG Lu 687/1-1. Address correspondence to J.F.D., Department of Radiology, University Hospital Essen, Hufelandstrasse 55, D-45122 Essen, Germany (e-mail: debatin@uni-essen.de).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: MR colonography was performed in 132 patients referred for colonoscopy because of the possible presence of a mass. After rectal filling with a gadopentetate dimeglumine and water enema, T1-weighted three-dimensional gradient-echo MR studies were acquired with the patient in the prone and supine positions. Water-sensitive single-shot fast spin-echo MR images were also obtained. Surface-rendered virtual endoscopic endoluminal views, orthogonal sections in three planes, and water-sensitive MR images were interactively assessed for presence of colorectal masses by two radiologists.
RESULTS: MR colonography was well tolerated without sedation or analgesia. MR image quality was sufficient for diagnosis in 127 (96%) patients. Most small (
5-mm-diameter) masses were overlooked at MR colonography, but 19 of 31 6–10-mm lesions and 26 of 27 large (>10-mm) lesions were correctly identified. For these large masses, MR colonography had a sensitivity of 93%, specificity of 99%, positive predictive value of 92%, and negative predictive value of 98% for detection of masses.
CONCLUSION: MR colonography is a promising modality for help in detecting colorectal mass lesions larger than 10 mm in diameter.
Index terms: Colon, MR, 75.121411, 75.121412, 75.12143, 75.12149 • Colon neoplasms, 75.311, 75.32 • Colon neoplasms, diagnosis • Magnetic resonance (MR), three-dimensional, 75.121411, 75.121412, 75.12143, 75.12149
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The apparent discrepancy between theoretic potential and clinical reality points to the lack of an ideal modality for colorectal polyp screening. Currently, a number of methods are used to detect colorectal mass lesions: tests for occult blood in fecal material (4), tumor markers tests (5), sigmoidoscopy (6), single- or double-contrast barium enema studies (7), hydrocolonic ultrasonography (US) (8), and colonoscopy (9). None fulfill all of the essential requirements needed for a screening test (10): high diagnostic accuracy, particularly with regard to the negative predictive value; low cost; and high patient acceptance and compliance.
This has stimulated interest in the development of cross-sectional imaging as a tool for the detection of colorectal masses. Proposed modalities include computed tomographic (CT) colonography (11) and magnetic resonance (MR) colonography (12). Both techniques are based on the acquisition of high-spatial-resolution three-dimensional (3D) data sets that encompass the entire colon. Advanced image postprocessing allows analysis of the colon in a multiplanar, as well as a virtual endoscopic, format. Lack of ionizing radiation and excellent soft-tissue contrast capabilities appear to favor the use of MR imaging.
The purpose of this study was to assess the ability to detect colorectal masses by using MR colonography, with conventional colonoscopy as the standard of reference.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Exclusion criteria included absolute contraindications to MR imaging, such as the presence of a pacemaker or metallic cerebral aneurysm clips, as well as a clinical diagnosis of acute abdomen. Furthermore, patients with severe claustrophobia were excluded. Owing to limitations in available research imager time, the number of possible studies per week was limited to three. For each slot on a particular day, the study patient was selected from among those examined with colonoscopy at the gastrointestinal clinic on the same day on a "first to consent" basis. All patients underwent standard preparation for colonoscopy (oral ingestion of 3 L of a bowel preparation solution) the day before the examination.
MR Colonography
MR colonography was performed with a 1.5-T MR system (Signa EchoSpeed; GE Medical Systems, Milwaukee, Wis). No sedative or analgesic agents were administered. To permit coverage of the entire colon, the body coil was used for signal transmission and reception.
After placement of a disposable rectal enema tube (E-Z-Em, Westbury, NY), patients were placed in the prone position on the imager table. The enema tip was blocked with an inflated balloon. To minimize peristalsis and alleviate colonic spasm, 20 mg of scopolamine butylbromide (Buscopan; Boehringer Ingelheim, Germany) was injected intravenously. Subsequently, the enema, which consisted of 3 L of water with 60 mL of 0.5 mol/L gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany), was administered via the rectal tube by using up to 100 cm of hydrostatic pressure. The prefilled enema bag contained 3,000 mL of fluid. The administered volume was 1,800–3,000 mL, tailored to the capacity of the patient’s colon.
To ensure safe and optimal filling of the colon, the filling process was monitored with a non–section-selective MR sequence (repetition time msec/echo time msec = 6/1.3, 60° flip angle) that provided an update image every second (Fig 1). Once complete filling and adequate distention of the entire colon had been documented, 3D spoiled gradient-echo (3.8/2.5, 40° flip angle) (Fig 2a) and two-dimensional single-shot fast spin-echo (
/65 [effective], 90° flip angle) pulse sequences (Fig 2c) were performed with the patient in the prone position. Subsequently, the 3D spoiled gradient-echo sequence was repeated with the patient in the supine position (Fig 2b). Each imaging sequence was performed in the coronal plane within a single breath hold that lasted 30 seconds or less. The imaging parameters included a field of view of 34–42 cm, section-thickness of 2–3 mm for the 3D spoiled gradient-echo sequence and 6 mm for the two-dimensional single-shot fast spin-echo sequence, and matrix of 384 x 192 or 256 x 160. The parameters were chosen to maximize spatial resolution within the confines of breath holding and coverage of the entire colon.
|
|
|
|
MR colonographic studies were interpreted by two radiologists (W.L., S.W.) at a consensus reading. Both readers were blinded to the colonoscopic findings and patient history. For analysis, the colon was divided into its six anatomic segments: rectum; sigmoid, descending, transverse, and ascending colon; and cecum. The presence of a mass in each segment was rated with a five-point scale: 1 for definitely no mass, 2 for probably no mass, 3 for nondiagnostic study, 4 for mass probably present, and 5 for mass definitely present.
The initial analysis was based on the virtual colonoscopic endoluminal view of the contrast agent–filled colon as seen on images from the prone 3D data sets. In regions that contained larger pockets of air, these were supplemented by views from the supine data set. The colon was explored in both antegrade (cecum to rectum) and retrograde (rectum to cecum) directions to ensure full visualization of both sides of the haustral folds. If a possible endoluminal protrusion was identified on the virtual colonoscopic views, the corresponding region was analyzed on the orthogonal reconstructed cross sections, as well as on images obtained with the other sequences (Fig 2). Differential diagnostic considerations included the presence of air, fecal material, or a polypoid mass. Criteria for differentiating air from a mass included a gravity-dependent position, presence of an associated air-fluid level, movement of the finding between the prone and supine data sets, and lack of a correlated finding on T2-weighted MR images. The differentiation of fecal material was based on motion between the prone and supine data sets and low signal intensity on the T2-weighted images. A polyp, on the other hand, remained unchanged in position between the prone and supine data sets and had high signal intensity on T2-weighted images (Fig 2). Polyp size was determined on the basis of its largest diameter as seen on images from multiplanar 3D data sets.
Subsequently, the colon was analyzed again by systematically scrolling through each of the three orthogonal reformatted image sets from the prone and supine 3D MR data followed by detailed analysis of the water-sensitive two-dimensional single-shot fast spin-echo MR images. The water-sensitive images were also used for screening for other abdominal and pelvic abnormalities contained in the field of view.
Conventional Colonoscopy
Colonoscopy was performed in the standard fashion with commercially available equipment (model CFQ 140; Olympus Europa, Hamburg, Germany). When necessary, sedation was accomplished with intravenous administration of 2–12 mg of midazolam maleate (Dormicum; Roche, Basel, Switzerland), and analgesia was accomplished with 10–50 mg of meperidine hydrochloride (Dolantin; Hoechst, Bad Soden, Germany). The largest diameter of each detected mass was estimated by means of endoscopic placement next to the lesion of an open biopsy forceps with a known width of 8 mm. In cases where multiple masses were apparent in one segment, only the five largest lesions were measured for size and used for correlation with MR colonographic results.
Analysis
Conventional colonoscopy served as the standard of reference. To determine the diagnostic performance of MR colonography relative to that of colonoscopy, "definite" and "probable" ratings were combined. Initially, the sensitivity of MR colonography with regard to detection of masses was evaluated independent of lesion size. Subsequently, the data were analyzed on segmental and patient levels. These analyses were based on the single largest mass detected at colonoscopy in either a segment or a patient. Sensitivity, specificity, and negative and positive predictive values were calculated by using 5-, 7-, and 10-mm cutoff values for mass size as determined at colonoscopy. Diagnostic confidence was assessed with receiver operating characteristic analysis of segmental MR colonographic ratings. The same 5-, 7-, and 10-mm cutoff values were used.
MR colonographic data sets were evaluated for the presence of other relevant abnormalities that had been previously documented with US (hepatic metastasis), CT in conjunction with 
-fetoprotein assay results (hepatocellular carcinoma), MR imaging (ovarian cancer), or surgery (ovarian and prostate cancer).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Image quality was sufficient to permit a definitive rating of "positive" or "negative" in 541 (74%) of 732 segments and a rating of "probable" in 176 (24%) of 732 segments; poor image quality resulted in a rating of "nondiagnostic study" in 15 (2%) of 732 segments in five patients. With regard to the cecum and ascending colon, image quality was predominantly affected by motion artifacts. These studies were excluded from further analysis, so that the final determination of diagnostic accuracy was based on 702 segments in 117 patients.
Colonoscopy revealed a total of 189 masses in 58 patients, with 129 (68%) lesions smaller than or equal to 5 mm in diameter and 60 (32%) larger than 5 mm. The size distribution of colonoscopically detected lesions and the correlation with MR colonographic results are shown in Figure 3. Although most lesions smaller than or equal to 5 mm were not detected at MR colonography, all but one lesion larger than 10 mm were correctly identified.
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Despite the operator dependence (13) and a miss rate of up to 27% for smaller (5-mm) lesions (14), colonoscopy must be considered the standard of reference for assessment of the colon. In addition to its unsurpassed diagnostic accuracy, colonoscopy provides a treatment option. Discomfort, resulting in frequent need for sedatives and analgesics, as well as high cost, limit the use of colonoscopy for screening purposes (15). MR colonography is characterized by considerably less discomfort. Although 53% of patients required a sedative and 8% requested analgesics during colonoscopy, the administration of the single contrast material enema for MR colonography did not trigger any request for analgesics. This reflects the low pressure needed for liquid-based distention of the colon. Aside from the reduction in pain, this form of distention also decreases the risk of perforation.
Once the colon is filled with contrast material, MR colonography requires fewer than 90 seconds for data acquisition, which is accomplished with three breath holds. The entire examination, including patient positioning and filling of the colon, which can be performed by a technologist, lasted an average of 20 minutes. Additional physician time must be added for analysis of the 3D data set. The time required for analysis of a colonic 3D data set is dependent on a number of factors, including the complexity of the underlying colonic morphology and the configuration of the user interface inherent to the software system. Most important, however, analysis times appeared to be influenced by reader experience. Thus, although not documented in a quantitative fashion, we noted a considerable decrease in required analysis time as the study progressed and the readers became more experienced. Evaluation of a complete data set was possible in 10–15 minutes.
In its current implementation, virtual endoscopic processing of the prone 3D data set was used as the primary means for polyp detection (Fig 2). Virtual endoscopy permitted appreciation of the circumferential morphology of haustral folds, which was difficult to appreciate on cross-sectional images (Fig 5) (16,17). Reformation images from the 3D MR data sets and the coronal two-dimensional water-sensitive MR images were used to confirm the presence of a space-occupying lesion in the colonic lumen and to differentiate polyps from air bubbles and residual stool (Fig 2). The efficiency of data analysis could be substantially improved by implementing an automated pathfinder (18) capable of simultaneously providing antegrade and retrograde views as the virtual camera is advanced through the colon. This would ensure that both sides of the haustrum are displayed and would eliminate the need for a reverse fly through. Finally, simultaneous display of the corresponding location on water-sensitive MR images would improve diagnostic efficiency.
|
The diagnostic performance of MR colonography was directly related to polyp size. Whereas clinically relevant lesions whose size exceeded 10 mm were reliably diagnosed (Fig 4), with sensitivity and specificity values of 94% and 99%, respectively, most polyps with a diameter that was smaller than or equal to 5 mm were overlooked. In view of the limited risk of malignant degeneration associated with such small lesions (24–30), the clinical effect of overlooking them is almost negligible in the context of a 3–5-year screening interval. Failure to detect 12 (39%) of 31 lesions with a diameter of 6–10 mm may have more serious consequences. Pending improvements in spatial resolution driven by developments in receiver coil technology (31–33), faster gradient systems, and more efficient pulse sequence design (34) will help further improve the performance of MR colonography, especially with regard to the water-sensitive single-shot fast spin-echo sequence, which is presently limited because of the relatively large (6-mm) section thickness. Thus, the beneficial effect of this sequence with regard to confirmation of the presence of a polyp applies only to larger (>10-mm) lesions. Accordingly, there was only one false-positive finding for a lesion larger than 10 mm (Fig 3). Flat lesions, which are easily recognized at colonoscopy (13), however, will likely remain obscure at MR colonography.
Artifacts related to colonic peristalsis resulted in nondiagnostic studies for 15 segments in five patients. This 4% failure rate was greater than that for colonoscopy, where the colonoscope could not be passed up the sigmoid colon in 3% of patients; the failure rate for colonoscopy in our study was similar to that reported in the literature (35). Diagnostic confidence, particularly with regard to the detection of small polyps (Fig 4), was reduced by readings of a "probable" polyp in 176 (24%) segments. Intraperitoneal segments (cecum, transverse and sigmoid colon) were affected more than were retroperitoneal segments (rectum, ascending and descending colon), which suggests that respiratory and peristaltic motion artifacts were a major contributor to the reduction in diagnostic confidence.
The cost of the enema used for MR colonography still needs to be reduced; this could be accomplished, for example, by means of partial or total substitution of gadopentetate dimeglumine with iron glycerophosphate (36). However, MR colonography appears to overcome many of the limitations of colonoscopy, barium enema examination, and even CT colonography, which to date have prevented realization of the well-documented benefits associated with screening for the presence of a colorectal mass (2). The noninvasive nature and lack of discomfort and ionizing radiation, in conjunction with the potential to simultaneously screen for other abdominal or pelvic malignancies, promise to improve patient acceptance—an important key variable (15) in the cost-effectiveness ratio of any screening program. We conclude that the performance of MR colonography, as currently implemented, for the detection of masses that exceed 10 mm in diameter warrants further consideration of this technique as a potent option in the diagnostic arsenal for colorectal mass screening.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, J.F.D.; study concepts and design, W.L., J.F.D., P.B.; definition of intellectual content, W.L., J.F.D.; literature research, W.L.; clinical studies, W.L., S.W., P.B., M.F.; data acquisition, W.L., P.B., M.F., S.W.; data analysis, J.F.D., W.L., P.B., S.W.; statistical analysis, J.F.D., W.L., S.W.; manuscript preparation and editing, W.L., J.F.D.; manuscript review, J.F.D., B.M., M.F.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  I. Yamada, S. Okabe, M. Enomoto, K. Sugihara, N. Yoshino, A. Tetsumura, J. Kumagai, and H. Shibuya Colorectal Carcinoma: In Vitro Evaluation with High-Spatial-Resolution 3D Constructive Interference in Steady-State MR Imaging Radiology, December 19, 2007; (2007) 2462070128. [Abstract] [Full Text]
|
|
|
|
|
|
J. Florie, E. Birnie, R. E. van Gelder, S. Jensch, B. Haberkorn, J. F. Bartelsman, A. van der Sluys Veer, P. Snel, V. P. M. van der Hulst, G. J. Bonsel, et al. MR Colonography with Limited Bowel Preparation: Patient Acceptance Compared with That of Full-Preparation Colonoscopy Radiology, October 1, 2007; 245(1): 150 - 159. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B Saar, A Meining, A Beer, M Settles, H Helmberger, E Frimberger, E J Rummeny, and T Rosch Prospective study on bright lumen magnetic resonance colonography in comparison with conventional colonoscopy Br. J. Radiol., April 1, 2007; 80(952): 235 - 241. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Florie, S. Jensch, R. A. J. Nievelstein, J. F. Bartelsman, L. C. Baak, R. E. van Gelder, B. Haberkorn, A. van Randen, M. M. van der Ham, P. Snel, et al. MR Colonography with Limited Bowel Preparation Compared with Optical Colonoscopy in Patients at Increased Risk for Colorectal Cancer Radiology, April 1, 2007; 243(1): 122 - 131. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Wessling, R. Fischbach, A. Borchert, H. Kugel, T. Allkemper, N. Osada, and W. Heindel Detection of Colorectal Polyps: Comparison of Multi-Detector Row CT and MR Colonography in a Colon Phantom Radiology, October 1, 2006; 241(1): 125 - 131. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
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]
|
|
|
|
 |
|
 S. C. Goehde, P. Hunold, F. M. Vogt, W. Ajaj, M. Goyen, C. U. Herborn, M. Forsting, J. F. Debatin, and S. G. Ruehm Full-Body Cardiovascular and Tumor MRI for Early Detection of Disease: Feasibility and Initial Experience in 298 Subjects Am. J. Roentgenol., February 1, 2005; 184(2): 598 - 611. [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. Wessling, R. Fischbach, N. Meier, T. Allkemper, J. Klusmeier, K. Ludwig, and W. Heindel CT Colonography: Protocol Optimization with Multi-Detector Row CT--Study in an Anthropomorphic Colon Phantom Radiology, September 1, 2003; 228(3): 753 - 759. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. C. Lauenstein, H. Schneemann, F. M. Vogt, C. U. Herborn, S. G. Ruhm, and J. F. Debatin Optimization of Oral Contrast Agents for MR Imaging of the Small Bowel Radiology, July 1, 2003; 228(1): 279 - 283. [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]
|
|
|
|
|
|
J. T. Ferrucci Colon Cancer Screening with Virtual Colonoscopy: Promise, Polyps, Politics Am. J. Roentgenol., November 1, 2001; 177(5): 975 - 988. [Full Text] [PDF]
|
|
|
|
|
|
T. Lauenstein, G. Holtmann, D. Schoenfelder, S. Bosk, S. G. Ruehm, and J. F. Debatin MR Colonography Without Colonic Cleansing: A New Strategy to Improve Patient Acceptance Am. J. Roentgenol., October 1, 2001; 177(4): 823 - 827. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. T. Keogan and R. R. Edelman Technologic Advances in Abdominal MR Imaging Radiology, August 1, 2001; 220(2): 310 - 320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Luboldt, O. Luz, R. Vonthein, M. Heuschmid, M. Seemann, J. Schaefer, D. Stueker, and C. D. Claussen Three-Dimensional Double-Contrast MR Colonography: A Display Method Simulating Double-Contrast Barium Enema Am. J. Roentgenol., April 1, 2001; 176(4): 930 - 932. [Full Text] [PDF]
|
|
|
|
 |
|
 Magnetic Resonance Colonography to Detect Polyps Journal Watch Dermatology, September 25, 2000; 2000(925): 15 - 15. [Full Text]
|
|
|
|
 |
|
 Magnetic Resonance Colonography to Detect Polyps Journal Watch (General), August 29, 2000; 2000(829): 3 - 3. [Full Text]
|
|
|
|
|
|
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]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
), 7 mm (
), and 10 mm (
) were applied only for readings of the standard colonoscopic (reference) images. Larger polyps were diagnosed with a high level of confidence almost independently of underlying image quality.