RadioGraphics
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH
QUICK SEARCH:   [advanced]


     

Published online October 3, 2003, 10.1148/rg.e18

(RadioGraphics 2004;24:e18.)

A more recent version of this article appeared on January 1, 2004
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow MPEG movies
Right arrow All Versions of this Article:
e18v1
24/1/e18    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Geenen, R. W. F.
Right arrow Articles by Krestin, G. P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Geenen, R. W. F.
Right arrow Articles by Krestin, G. P.
© RSNA, 2003

Online Only

CT and MR Colonography: Scanning Techniques, Postprocessing, and Emphasis on Polyp Detection1

Remy W. F. Geenen, MD, Shahid M. Hussain, MD, PhD, Filippo Cademartiri, MD, Jan-Werner Poley, MD, Peter D. Siersema, MD, PhD and Gabriel P. Krestin, MD, PhD

1 From the Departments of Radiology (R.W.F.G., S.M.H., F.C, G.P.K.) and Gastroenterology and Hepatology (J.W.P., P.D.S.), Erasmus Medical Center, Dr Molenwaterplein 40, 3015 GD Rotterdam, The Netherlands, and the Department of Radiology, University Hospital Parma, Parma, Italy (F.C.). Presented as an educational exhibit at the 2002 RSNA scientific assembly. Received April 29, 2003, revision requested July 11, revision received and accepted August 13. Address correspondence to R.W.F.G. (e-mail: r.geenen@erasmusmc.nl).


   Abstract
Top
 Abstract
Introduction
 CT and MR Colonography
 Postprocessing
 Discussion
 Acknowledgement
 References
 
In the last decade, computed tomographic (CT) and magnetic resonance (MR) colonography, two new cross-sectional techniques for imaging of the colon, emerged. Both techniques show promising initial results in the detection of polyps equal to or greater than 1 cm in diameter in symptomatic patients. Imaging protocols are still mostly under development and prone to change. Both CT and MR colonography generate a large number of source images, which have to be read carefully for filling defects and, if intravenous contrast material is used, enhancing lesions. An important postprocessing technique is multiplanar reformatting, which allows the viewer to see potential lesions in an orientation other than that of the source images. Virtual endoscopy, a volume rendering technique that generates images from within the colon lumen, is used for problem solving. CT and MR colonography have potential advantages over colonoscopy and double-contrast barium enema examination: multiplanar capabilities, detection of enhancing lesions that make the distinction between fecal residue and true lesion possible, and ante- and retrograde virtual colonoscopy. Currently, a number of studies suggest that patients have a preference for CT colonography over colonoscopy. Patients consider bowel cleansing the most uncomfortable part of any colon examination; hence, from the acceptance point of view, fecal tagging techniques are promising. Before CT and MR colonography can be implemented in daily practice, they must show approximately the same accuracy as colonoscopy for polyp detection in both symptomatic and screening patients.

© RSNA, 2003

Index Terms: Colon, MR, 75.121411, 75.121412, 75.121419, 75.12143 • Colon, CT, 75.12115, 75.12117 • Colon neoplasms, 75.30 • Magnetic resonance (MR), three-dimensional, 75.12143, 75.12149


   Introduction
Top
 Abstract
 Introduction
CT and MR Colonography
 Postprocessing
 Discussion
 Acknowledgement
 References
 
Colorectal carcinoma is the fourth most common incident cancer in the United States after breast, lung, and prostate cancer. In the United States, colon carcinoma is responsible for 111,000 new cases and 51,000 deaths each year (1). It is second in cancer-related causes of death, responsible for nearly 12% of the total deaths due to cancer (2). High frequencies of colon cancer, 25–35 per 100,000, are seen in North America, Western Europe, and Australia. The lowest rates, 1–3 per 100,000, are found in India (1). The lifetime risk of developing a colorectal carcinoma in a person without specific risk factors living in the western world approaches 6%, with a 3% risk of dying of the disease (3).

Colonic carcinogenesis is a multistage process involving a number of morphologic, molecular, and genetic steps (4,5). There is substantial evidence that most colorectal carcinomas arise from preexisting adenomas, the so called adenoma-to-carcinoma sequence (Fig 1) (6,7). A polyp is defined as any circumscribed tumor or protuberance that projects above the surface of the surrounding normal flat mucous membrane (6). The adenoma-to-carcinoma sequence takes many years to develop. The time interval from normal colon to invasive cancer has been estimated to be 10 years and from normal colon to adenoma 5 years (8). This long interval implies that colorectal carcinoma is a potentially preventable disease, as long as polyps are discovered and removed before they become malignant.



View larger version (53K):
[in this window]
[in a new window]
  		Figure 1.  The adenoma-to-carcinoma sequence.

 
The two major types of polyp found in the colon are hyperplastic and adenomatous. Hyperplastic polyps are nonneoplastic proliferations of epithelium that are located predominantly in the rectosigmoid colon and constitute more than 10% of polypectomy specimens (9,10). Fifty percent of polyps equal to or smaller than 5 mm in diameter and 30% 6–9 mm are hyperplastic polyps (4,6). These polyps almost never undergo malignant transformation (4,6,9).

Adenomatous polyps may be classified as tubular, tubulovillous, or villous adenomas. Tubular adenomas constitute 87.1% of the endoscopically removed adenomas, tubulovillous adenomas 8.2%, and villous adenomas 4.7% (10). All of these may undergo malignant transformation into invasive adenocarcinoma. In general, the greater the villous component of the polyp, the greater the risk of malignant transformation (10). The risk of developing an adenocarcinoma is directly related to polyp size; only about 1% of adenomas less than 1 cm in diameter harbor adenocarcinoma, whereas 10% of adenomas 1–2 cm in diameter and more than 40% of those greater than 2 cm harbor adenocarcinoma (4). High-grade dysplasia is found in about 5% of polyps less than 1 cm in diameter, in 15% of polyps 1–2 cm, and in 25% of polyps greater than 2 cm (10). It has been proved that colonoscopic polypectomy decreases the incidence of colorectal cancer by 76%–90% (11). When it comes to colorectal cancer prevention strategies, the issue is whether it is important to find and remove small (<1 cm) colonic polyps (12). More than 70% of all discovered adenomas are <1 cm in diameter, and, as stated, 99% of them will not harbor adenocarcinoma (4,12). The terms significant polyp and advanced adenoma are becoming increasingly popular, applied to polyps equal to or larger than 1 cm in diameter or containing villous or dysplastic elements at histology (12). Practically, this means that any colon examination technique should help detect polyps equal to or larger than 1 cm with high accuracy, and would preferably also show polyps 6–9 mm in diameter, as polyps of this size can contain carcinoma or villous and dysplastic elements. The emphasis, however, should be on detecting polyps equal to or larger than 1 cm.

Currently, there are four methods for investigation of the entire colon. These are double-contrast barium enema (DCBE), colonoscopy, CT colonography, and MR colonography. Fischer described the DCBE technique in 1923 (13). It was refined in the late 1960s and became the radiologic technique of choice for colon imaging in the mid-1970s (14,15). Recently, the DCBE technique was reviewed (16). It was concluded that performing a high-quality DCBE study requires tailoring of the examination to the clinical history, patient, and fluoroscopic findings. Each colonic segment should be viewed in detail with spot radiographs or magnified digital images. The order in which these are obtained is flexible, as long as each loop of colon has adequate barium coating and distention and is demonstrated en face. Overhead views such as left and right side–down decubitis views and a prone-angled view of the rectosigmoid junction are helpful in piecing together the spot images (16).

Colonoscopy was first described in 1965 by three independent Japanese groups in the same journal (17–19). Since then, technical developments made scopes smaller, easier to manipulate around angles, and improved the quality of the visualization methods.

Compared to DCBE studies and colonoscopy, CT and MR colonography have a short history and are still being developed. CT colonography was described in 1994 by Vining et al (20) and MR colonography in 1997 by Luboldt et al (21). Both are cross-sectional methods that generate numerous images in the axial plane (CT) or any desired plane (MR imaging), preferably during one breath-hold. To efficiently read these images, postprocessing on a workstation is necessary. Such workstations should be able to handle the data quickly and therefore should have adequate hardware and software to allow fast interaction with the data set. These data sets can consist of up to 700 images with relatively high spatial resolution.

Reading the source images is the first step. These images need to be viewed carefully for filling defects and, if applicable, enhancing lesions. Postprocessing is an important feature of image interpretation. The most simple and important postprocessing technique for CT and MR colonography is multiplanar reformatting (MPR) (22,23). Furthermore, volume rendering techniques, such as tissue transition projection or endoscopic three-dimensional (3D) viewing (virtual endoscopy), can be performed. These require a great deal of computer power; endoscopic 3D viewing is especially time-consuming. Other 3D rendering techniques such as maximum-intensity projection (MIP) and shaded surface display (SSD) are easy to perform but only a small part of the entire data set is used in these techniques. Thus, much important information is lost and this makes these techniques unsuitable for polyp detection. Postprocessing techniques will be discussed in detail. The purpose of this article is to describe scanning techniques in CT and MR colonography, discuss the currently available postprocessing methods, and discuss the accuracy of these techniques for polyp detection compared with colonoscopy and DCBE.


   CT and MR Colonography
Top
 Abstract
 Introduction
 CT and MR Colonography
Postprocessing
 Discussion
 Acknowledgement
 References
 
Both CT and MR colonography can be performed after bowel cleansing or fecal tagging is used. The latter means that patients must add a contrast agent to their diet to give the stool a different attenuation value (CT) or signal intensity (MR imaging) than that of polyps. In CT colonography, residual cleansing liquid limits interpretation. It has been shown that a 48-hour low-residue diet, combined with phosphosoda and bisacodyl gives less residual liquid than polyethylene glycol solution (24). In MR colonography, the amount of residual liquid after bowel cleansing is of little importance, since water or a water-based enema solution is usually used.

Bowel relaxing agents are used to relax the colonic wall and minimize peristalsis. In the United States, glucagon hydrochloride is used for this purpose. Its usefulness in CT colonography is not clear. It does not substantially improve polyp detection or colonic distention, and it is unclear whether it diminishes discomfort (25). In Europe, butyl scopolamine (Buscopan; Boehringer Ingelheim, Alkmaar, Netherlands) is recommended for CT and MR colonography (26). It has a better safety profile than glucagon, is much cheaper, and may reduce pain associated with colonic distention (26). For MR colonography, all investigators use bowel-relaxing agents to reduce peristalsis-related artifacts, which are integrated over the entire imaging time in 3D sequences (26). To expose the entire bowel mucosa to the enema solution and to not miss polyps that may not protrude, prone and supine imaging are needed. Imaging should start with the patient in prone position. This avoids breathing excursions if the patient is not able to hold his breath, to compress the field of view for optimal voxel resolution, and to improve distention of the sigmoid colon by compression from the small bowel (26).

For CT colonography, the colon is filled with either room air or carbon dioxide (Fig 2). The latter is quickly absorbed through the colon wall and blood (26). It therefore diminishes discomfort; however, using it adds additional cost and complexity to the procedure (12). Thin-section helical CT of the abdomen and pelvis is then performed. A single-detector CT protocol consists of the following parameters: tube current, 70–110 mA; voltage, 120 kVp; collimation, 3–5 mm; table feed or rotation, 3.75–10 mm; reconstruction increment, 1.5–3 mm (27). Owing to its ability to produce isotropic voxels and scan a larger volume with high spatial resolution within one breath hold, CT colonography should preferentially be performed with a multisection CT scanner (26). A multidetector CT colonography protocol depends on the type of scanner, as different manufacturers use different detector designs and different connections (28). Protocols for different types of multidetector scanners have been described in the literature (26–28). Compared with single-detector CT colonography, multidetector CT colonography improves demonstration of colonic distention and has fewer respiratory artifacts (29).



View larger version (49K):
[in this window]
[in a new window]
  		Figure 2.  Segmental collapse of the colon. (Left) Axial source CT scan and (right) tissue transition projection.

 
Stool rests are a major problem in reading CT and MR colonographic images, as they can mimic polyps. This problem can be overcome in one of three ways: 1. The use of intravenous contrast material; polyps enhance, stool rests do not. 2. Fecal tagging. 3. A combination of the two. Morrin et al (30) used 150 mL of a nonionic monomer infused intravenously at a rate of 3.5 mL/sec and a delay of 45 seconds for CT colonography. They found significantly improved reader confidence in assessment of bowel wall conspicuity and a significant improvement in the depiction of polyps 6–9 mm in diameter. Oto et al (31) used 150 mL of a nonionic monomer infused intravenously at a rate of 3–4 mL/sec and a delay of 70 seconds. On unenhanced CT scans, polyps had an attenuation value of 32.4 HU, and carcinomas had a value of 42.6 HU. After contrast material administration, these values increased substantially to 78.9 HU and 90.7 HU, respectively. Lefere et al (32) compared standard bowel cleansing with polyethylene glycol with a fecal-tagging diet that contained 250 mL of barium solution administered three times per day. They found increased specificity in the fecal-tagging group and a comparable sensitivity. Furthermore, in the fecal-tagging group, patients reported less discomfort. Thomeer et al (33) added an ionic contrast agent to the bowel preparation solution to tag residual liquid and fecal rests. They found a high sensitivity (91.7%) and specificity (95%) for the detection of polyps equal to or larger than 1 cm in diameter. Comparable figures for polyps equal to or larger than 1 cm (sensitivity, 90%; specificity, 94.6%) were recently reported by Pineau et al (34), who used a single-section protocol in which fecal tagging was performed with an ionic contrast agent.

MR colonography is an even more recent method than CT colonography. A final protocol has not been defined yet. First reports described breath-hold 3D gradient-echo (GRE) sequences that covered the entire enema solution–filled colon in the coronal plane (21). Although some reports show that MR colonography is possible with 1-T imagers, most reports describe use of 1.5-T imagers (35,36). Air and carbon dioxide have been used as colonic distention agents, but most reports address liquid enema techniques (37–39).

Currently, three different liquid enema techniques are used: bright lumen, black lumen, and fecal tagging (40–42) (Table 1). Both bright and black lumen methods are used after bowel cleansing. The bright lumen is based on a water-gadolinium enema with a gadolinium concentration of at least 10 mmol/L (eg, 40 mL of 0.5 M gadolinium in 2 L of water) (26). Owing to the lack of radiation, an MR fluoroscopy sequence can be applied to monitor filling of the colon. In the case of a water-gadolinium enema, this is a T1-weighted GRE sequence (repetition time [TR], 3.7 msec; echo time [TE], 1.1 msec; flip angle, 40°–60°). After complete filling, a thin-section 3D spoiled T1-weighted GRE sequence (3.7/1.1; flip angle, 40°–50°) is performed. With this sequence, the lumen appears bright because of the presence of gadolinium, and polyps are visible as filling defects (40). Hereafter a two-dimensional half-Fourier single-shot turbo spin-echo sequence (TR, infinite; TE [effective], 90–180 msec; turbo factor, 90–120) can be used for delineation of the colonic wall (40).


View this table:
[in this window]
[in a new window]
  		TABLE 1. Three MR Colonography Techniques (40-42)

 
The black lumen technique is based on a water enema for luminal distention and an intravenous infusion of gadolinium for enhancement of the colonic wall (41). Filling can be monitored by applying a T2-weighted sequence such as a balanced GRE sequence (33). We use a T2-weighted single-shot turbo spin-echo sequence (TR, infinite; TE [effective], 600–800 msec; turbo factor, 20–25). After complete filling, a 3D T1-weighted spoiled GRE sequence (3.7/1.1; flip angle, 20°–30°) is performed before and after the intravenous administration of gadolinium. The lumen appears dark with this T1-weighted sequence, and enhancement of the colonic wall and possible inflammatory changes and polyps can be seen (41). We also add a two-dimensional or 3D balanced GRE sequence (3.7/1.1; flip angle, 80°) before the administration of gadolinium. This sequence is T2/T1-weighted, and the lumen appears bright, allowing filling defects such as polyps to be seen. A combination of 3D T1-weighted spoiled GRE sequence and balanced GRE sequence makes bright and black lumen colonography possible with one protocol (Fig 3).



View larger version (136K):
[in this window]
[in a new window]
  		Figure 3.  MR colonography. A, Bright lumen technique: 3D GRE sequence with water-gadolinium enema. B, Virtual colonoscopy of normal ascending colon. C, D, Combined bright and black lumen technique: C, contrast-enhanced 3D spoiled T1-weighted GRE image and, D, nonenhanced 3D spoiled and balanced GRE image.

 
The fecal tagging technique is based on a diet that contains barium to give stools the same signal intensity as water on T1-weighted GRE images. The diet consists of a 1 mg/mL barium sulfate solution administered in a dose of 200 mL with all four meals, beginning 36 hours before MR colonography. Patients are instructed to avoid fiber-rich foods and foods with a high concentration of manganese, such as fruit and chocolate (42). MR colonography starts with filling of the colon with water. Filling is monitored with a two-dimensional balanced GRE sequence (3.7/1.1; flip angle, 80°). Intravenous gadolinium is administered, and a 3D T1-weighted spoiled GRE sequence (3.7/1.1; flip angle, 20°–30°) is performed. The lumen appears black, and enhancement of the colonic wall and possible polyps is seen (42).


   Postprocessing
Top
 Abstract
 Introduction
 CT and MR Colonography
 Postprocessing
Discussion
 Acknowledgement
 References
 
Before any postprocessing is performed, the source images should be evaluated for image quality and presence of abnormalities. It should be kept in mind that the source images must have the best possible quality. The better the quality of the source images the easier it is to see the difference between folds and polyps and the better the quality of the postprocessing images will be. The source images should be examined carefully for filling defects and, if applicable, enhancing lesions. Currently, many postprocessing methods are commercially available on workstations. These include two-dimensional MPR, volume rendering techniques such as virtual colonoscopy and tissue transition projection, and older 3D rendering techniques such as shaded surface display and MIP. The difference between 3D rendering techniques and volume rendering is that in the latter the entire data set is used for the 3D image, while with the 3D rendering techniques only about 10% of the data set is used for rendering (43). A great deal of information is not used and lost. The value of these methods for CT and MR colonography is shown in Table 2.


View this table:
[in this window]
[in a new window]
  		TABLE 2. Postprocessing Techniques (20,22,43-50).

 
MPR
MPR (Fig 4, Movies 1–6) displays the images produced from the original reconstruction process in an orientation other than that of the source images (44). If, for example, axial CT images are the source images, the x direction is along the patient's left-right axis, the y direction along the anteroposterior axis and the z direction along the superoinferior axis. A sagittal view is then formed by sampling in the y and z directions and a coronal view by sampling in the x and z directions (44). Although MPR is a two-dimensional technique, no information is lost (45). Up-to-date workstations are able to display the source images into axial, coronal, and sagittal planes. Many workstations today are also capable of performing reformatting in any desired plane. MPRs are easy and fast to perform. Because of the nearly isotropic voxels produced in multisection CT colonography, these source images are able to produce MPRs of good quality. MPRs obtained after single-section CT colonography or MR colonography are of lower quality because the voxels are not entirely isotropic. In our experience, the MPR quality of single-section CT colonography and MR colonography data sets is fairly good in terms of providing the necessary diagnostic information concerning the distinction between haustral folds and polyps. Reformatting is mandatory, for haustral folds can mimic polyps in one direction. Haustral folds traverse the colonic wall from one side to the other, while polyps stop inside the air- or fluid-filled colon. Polyps show an abrupt ending in three directions. If a potential lesion is present in at least two of the three image planes, the confidence that this represents a true lesion increases (22). It has been validated that scrolling through MPR images is sufficient if no lesions or obvious lesions are seen. If doubt still exists, volume rendering techniques should be used for problem solving (22,26).



View larger version (103K):
[in this window]
[in a new window]
  		Figure 4.  MR colonography. (A) Axial (B) coronal, and (C) saggital MPRs show polyp near splenic flexure.

 
Commonly used volume rendering techniques for CT and MR colonography data sets are endoluminal views (virtual colonoscopy) and tissue transition projections. Volume rendering uses information from all the voxels in the data set, so no information is lost (43,45,46). By assigning a specific color and opacity value to every attenuation (CT) or signal intensity (MR imaging), groups of voxels are selected for display. The information from every voxel is integrated into the resulting image, and both brightness and depth information are presented (45).

Virtual Colonoscopy
Virtual colonoscopy (Figs 5, 6; Movies 7–9) is a combination of endoscopic viewing and cross-sectional volumetric imaging (46). The clinical application of this technique was described in 1996 by Rubin et al (47). By applying a threshold to the data set (high in bright lumen and low in black lumen, fecal tagging, and CT), it is possible to isolate the inside of the colon from the rest of the data set (47). Images are then rendered from a point source (the virtual endoscope) within the colon lumen (46). As a result, parts of the colon close to the virtual endoscope appear larger than those of identical size at a greater distance (47). Virtual colonoscopy should be performed in antegrade and retrograde directions, to look in front of and behind all haustral folds. If virtual colonoscopy is performed in only one direction, up to 25% of the colon lumen may be missed. This is reduced to 7% when bidirectional viewing is performed (48). When the colon is folded out and spread on a flat plane, the so-called unfold-cube representation, the entire luminal wall is exposed for examination (48). This technique was described in 2001 (48) and is not available on all workstations. On many workstations, it is possible to perform automatic path tracking through the colon lumen. This is possible only if the data set is of good quality, with the colon well distended and with little or no wall movement. If this is not the case, path tracking must be done manually, which is time-consuming. It then takes 20–30 minutes to create an endoluminal view from cecum to rectum. It is especially difficult to maintain a good view of all colon walls in the flexures when the unfolded-cube representation is not available. It has been validated that virtual colonoscopy is good for problem solving when doubt persists after viewing of source and MPR images (22,26).



View larger version (97K):
[in this window]
[in a new window]
  		Figure 5.  Virtual colonoscopy. (A) Retrograde and (B) antegrade views of polyp.

 


View larger version (127K):
[in this window]
[in a new window]
  		Figure 6.  Virtual colonoscopic view of polypoid lesion.

 
Tissue Transition Projection
In tissue transition projection (Fig 7, Movie 10), differences in attenuation value (CT) or signal intensity (MR imaging) of neighboring voxels (gradients) are calculated from the volume of interest. In homogeneous areas, such as the filled colon lumen, gradients are weak; in surface areas, such as the transition zone between colon lumen and colon wall, gradients are strong. First, a translucency value has to be determined for the whole data set. This value is modulated by the gradient strength. Areas with weak gradients become transparent, while areas with strong gradients make up most of the gray values of the calculated pixel (49). This technique was first described in 2000 by Rogalla et al (49). The result is a simulation of a DCBE (50). It takes about 5 minutes to perform a see-through view of the colon after CT or MR colonography; however, its diagnostic value is unknown.



View larger version (108K):
[in this window]
[in a new window]
  		Figure 7.  Tissue transition projection of inflammatory stenosis.

 
Shaded Surface Display
Shaded surface display (Fig 8, Movie 11) is a popular 3D rendering technique in which the voxels located at the edge of a structure are identified, usually by intensity thresholding, and displayed. The remaining voxels are invisible (44). The advantages of this technique include superior speed, flexibility, and clear depth cues. A large disadvantage is that only about 10% of the data set is used for rendering, and a great deal of information is lost (43). Performing a shaded surface display of the colon takes 5–10 minutes, mainly owing to segmentation of the colon from the rest of the data set. Because of the loss of information, it is not routinely used in CT and MR colonography.



View larger version (34K):
[in this window]
[in a new window]
  		Figure 8.  Shaded surface display of inflammatory stenosis.

 
MIP
MIP (Movie 12, 13) is a 3D rendering technique that examines each voxel along a line from the viewer's eye through the data set and selects the voxel with the highest intensity. This voxel is then used in the displayed volume. The rest of the voxels are ignored, resulting in an information loss of about 90% of the data set (43). The displayed pixel intensity will represent the voxel with the highest intensity along the projected ray (44). The advantage is that bright objects become more visible. Disadvantages include the lack of depth information, increase in mean background intensity, and high-intensity material such as calcifications at CT obscuring information (43,46). MIP might be useful in contrast-enhanced CT colonography and black lumen MR colonography, since enhancing polyps or inflammatory lesions may become more visible. Like MPR, MIP is easy to perform but is not routinely used in CT and MR colonography because of the loss of data.


   Discussion
Top
 Abstract
 Introduction
 CT and MR Colonography
 Postprocessing
 Discussion
Acknowledgement
 References
 
Colorectal cancer seems suitable for screening because of the adenoma-to-carcinoma progression and the long interval before a cancer develops from an adenoma. Which method is the most accurate and which method is or will become accepted by the general public? Until the mid-1990s, two methods for polyp detection were available: colonoscopy and DCBE. New methods such as CT and MR colonography have emerged since then. All four methods have their advantages and disadvantages (Table 3). The literature shows a wide range of results regarding sensitivity and specificity of DCBE compared with colonoscopy. Some review articles on this subject have been published (51–56). The sensitivity of DCBE for the detection of colon cancer was 85%–95%, the sensitivity for the detection of polyps equal to or larger than 1 cm in diameter was 75%–95%, and the sensitivity for the detection of polyps <1 cm was 50%–80% (51–56). The specificity of DCBE for polyps and cancer is 90%–95% (51,54). Our own research indicated that DCBE missed 5.1% of premalignant and malignant lesions in a group of symptomatic patients (57). The sensitivity of colonoscopy for the detection of colon cancer was 90%–95%, the sensitivity for the detection of polyps equal to or larger than 1 cm was 90%, and the sensitivity for the detection of polyps <1 cm was 75% (54–56). One author concluded that the sensitivity of colonoscopy is probably no more than 10% greater than that of DCBE for cancer and 15% greater for polyps (55). Reported miss rates for colonoscopy were 16%–27% for polyps equal to or smaller than 5 mm, 12%–13% for polyps 6–9 mm, and 0%–6% for polyps equal to or larger than 1 cm (58). The specificity of colonoscopy was assumed to be 100% (53,55). In as many as 15% of patients who undergo colonoscopy, the endoscopist is unable to reach the cecum (59). A positive factor in colonoscopy is the ability to perform an immediate intervention, such as biopsy or polypectomy, during the same procedure. Sedation is a positive factor from the perspective of patient acceptance but a major objection to the use of colonoscopy on a widespread basis. It accounts for about half the costs and most of the risks of colonoscopy, including vasovagal reactions and cardiopulmonary events (58). The use of sedation is culturally determined, ranging from 6% of cases in Finland to almost universal use in the United States (58).


View this table:
[in this window]
[in a new window]
  		TABLE 3. Colonoscopy versus DCBE versus CT and MR Colonography (40,51-56,62-69)

 
For any screening program to be effective, compliance should be as high as possible. Unfortunately, compliance in a screening setting for colonoscopy is only about 40% (60). This is partly due to the amount of pain experienced during the procedure (61). Also, embarrassment and the bowel-cleansing regimen have a negative effect on compliance in a screening program. Furthermore, obtaining a high specificity is important for both CT and MR colonography, since reducing the number of unnecessary colonoscopies would be important for the success of a screening program based on either of these modalities. A survey among potential patients showed that more than 75% would prefer virtual colonoscopy to conventional colonoscopy (60). Several reports address the fact that bowel preparation is considered the most uncomfortable part of both CT colonography and colonoscopy (62–65). Most patients preferred CT colonography to colonoscopy (62–65). In another study, overall satisfaction (which includes discomfort, tolerance, confidence in the results, treatment by physicians and nurses) was greater with colonoscopy; however, CT colonography caused less discomfort, was better tolerated, and was the preferred follow-up examination (66). Patients preferred conventional colonoscopy to CT colonography in only one study (67). Patients in this study reported more pain and discomfort during CT colonography (67).

Reported sensitivities of single-detector CT colonography vary according to polyp size. For polyps equal to or smaller than 5 mm, reported sensitivities are 11%–59%. For polyps 6–9 mm in diameter, sensitivities vary from 16% to 60%, and for polyps equal to or larger than 10 mm, sensitivities vary from 73% to 100% (27). Two studies that compared multidetector CT colonography with conventional colonoscopy have been published. Reported sensitivities for polyps equal to or smaller than 5 mm are 3%–12%, for polyps 6–9 mm 33%–70%, and for polyps equal to or larger than 10 mm 82%–93% (68,69). Specificity in these studies was 90%–97.7% (68,69). Recently, the results of a large study of 703 patients with a history of colorectal cancer or a strong family history of colorectal cancer or an iron deficiency anemia were published (70). Eighty-three percent of the patients were scanned with a four-row multidetector scanner; the rest were scanned with a single-section scanner. Double reading was performed. Sensitivity for polyps equal to or larger than 1 cm with double reading was 63%, and for polyps 5–9 mm in diameter, sensitivity was 54% (70). The difference between this study and previous studies was the low prevalence of polyps in the patient group.

Attempts have been made to reduce radiation dose at CT by reducing the tube current. It has been shown that reducing the tube current to 30 mAs substantially decreases image quality, while leaving polyp detection unimpaired (28). More and larger studies are needed to address this issue. The median effective dose calculated from 41 studies published between 1996 and 2002 was 7.2 mSv. Multisection scanners were associated with a higher effective dose than were single-section scanners (7.8 mSv vs 3.1 mSv). The risk of inducing a fatal cancer with an effective dose of 7.2 mSv was estimated as 1:5,500, if applied to a population aged 50 years (71). Recently, CT colonography findings of diverticular disease were described (72).

The overall sensitivity of bright-lumen MR colonoscopy is 6% for polyps equal to or smaller than 5 mm, 61% for polyps 6–9 mm in diameter, and 96% for polyps equal to or larger than 1 cm. If a cut-off value of 1 cm is used, overall sensitivity is 93% and specificity is 99% (40). Conventional colonoscopy was the standard of reference with which MR colonography was compared (40). Preliminary results of bright-lumen MR colonography in patients with inflammatory bowel disease show that this technique might be useful in patients with Crohn disease (73). Recently, the first results of dark-lumen MR colonography were described in symptomatic patients. The overall sensitivity was 90%, and the specificity was 100% (74). Sensitivity and specificity of 100% was attained for all polyps larger than 5 mm (74). MR colonography with fecal tagging showed a sensitivity of 89.3% for the detection of colonic lesions varying from 6 to 55 mm in diameter, compared with conventional colonoscopy (42). Further research should focus on the value of combinations of described sequences and further improvement in spatial resolution within one breath hold.

In conclusion, CT and MR colonography are new techniques for imaging of the colon. In symptomatic patients, these new techniques show promising results for the detection of polyps equal to or larger than 1 cm in diameter. It must be remembered that in all research protocols, colonoscopy was considered to be the standard of reference, which implies that other imaging modalities with which colonoscopy is compared will always perform worse. In most studies, patients preferred CT colonography to conventional colonoscopy. The bowel-cleansing regimen is considered to be cumbersome, so from the patient acceptance point of view, fecal tagging techniques are promising. Their value in polyp detection still needs to be determined in large studies. In medicine, there is a trend toward performing noninvasive or less invasive imaging techniques rather than older and more validated invasive techniques. (MR angiography or CT angiography vs digital subtraction angiography, MR cholangiopancreatography vs endoscopic retrograde cholangiopancreatography). The invasive techniques are used for problem solving and interventions. CT and MR colonography fit in this trend perfectly. Both techniques have shown promising initial results in symptomatic patients and are still in evolution. Before these techniques can be implemented in daily practice, they must show the same accuracy as colonoscopy and should be cost-effective in both high-risk and screening patients. The radiation-dose issue in CT colonography must be discussed, and a consensus on the maximum acceptable dose for a screening patient must be reached. MR colonography has the advantage of being a zero-dose examination, but at this point, CT colonography is faster and provides images with higher resolution.


   Acknowledgement
Top
 Abstract
 Introduction
 CT and MR Colonography
 Postprocessing
 Discussion
 Acknowledgement
References
 
The authors thank Andries W. Zwamborn for his help in preparing the electronic version of this manuscript.


   Footnotes
 
Abbreviations: DCBE = double-contrast barium enema, GRE = gradient-echo, MIP = maximum-intensity projection, MPR = multiplanar reformatting, SSD = shaded surface display, TE = echo time, TR = repetition time, 3D = three-dimensional.


   References
Top
 Abstract
 Introduction
 CT and MR Colonography
 Postprocessing
 Discussion
 Acknowledgement
 References
 

  1. Potter JD, Stattery ML, Bostick RM, Gapstur SM. Colon cancer: a review of epidemiology. Epidemiologic Rev 1993; 15:499-545.[Free Full Text]
  2. Visser O, Coebergh JWW, Schouten LJ, et al. Incidence of cancer in the Netherlands 1996 Utrecht, The Netherlands: Vereniging van integrale kankercentra, 2000.
  3. Seidman H, Mushinski MH, Gelb SK, et al. Probability of eventually developing and dying of cancer: United States, 1985. CA Cancer J Clin 1985; 35:36-56.[Abstract/Free Full Text]
  4. Muto T, Bussey HJR, Morson BC. The evolution of cancer of the colon and rectum. Cancer 1975; 36:2251-2270.[Medline]
  5. Vogelstein B, Fearon ER, Hamilton SR, et al. Genetic alterations during colorectal-tumor development. N Engl J Med 1988; 319:525-532.[Abstract]
  6. Morson BC. The evolution of colorectal carcinoma. Clin Radiol 1984; 35:425-431.[CrossRef][Medline]
  7. Toribara NW, Sleisinger MH. Screening for colorectal cancer. N Engl J Med 1995; 332:861-867.[Free Full Text]
  8. Winawer SJ, Zauber AG, Diaz B. The national polyp study: temporal sequence of evolving colorectal cancer from the normal colon. Gastrointest Endosc 1987; 33:167.
  9. Levine MS, Rubesin SE, Laufer I, Herlinger H. Diagnosis of colorectal neoplasms at double-contrast barium enema examination. Radiology 2000; 216:11-18.[Abstract/Free Full Text]
  10. O'Brien MJ, Winawer SJ, Zauber AG, et al. The national polyp study: patient and polyp characteristics associated with high-grade dysplasia in colorectal adenomas. Gastroenterology 1990; 98:371-379.[Medline]
  11. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J Med 1993; 329:1977-1981.[Abstract/Free Full Text]
  12. Ferrucci JT. Colon cancer screening with virtual colonoscopy. AJR 2001; 177:975-988.[Free Full Text]
  13. Fischer AW. Ueber eine neue roentgenologische untersuchungsmethode des dickdarms: kombination von kontrasteinlauf und luftaufblaehung. Klin Wochenschr 1923; 2:1595-1598.[CrossRef]
  14. Welin S. Results of the Malmö technique of colon examination. JAMA 1967; 199:369-371.[CrossRef][Medline]
  15. Figiel SJ.. Colon examination technique: detection of colon lesions. From the First Standardization Conference-1969 Chicago: American College of Radiology, 1973; 132-145.
  16. Rubesin SE, Levine MS, Laufer I, Herlinger H. Double-contrast barium enema examination technique. Radiology 2000; 215:642-650.[Abstract/Free Full Text]
  17. Kanazawa T, Tanaka M. Endoscopy of the colon. Gastroenterol Endosc Tokyo 1965; 7:398-400.
  18. Oshiba S, Wanatabe A. Endoscopy of the colon. Gastroenterol Endosc Tokyo 1965; 7:400-402.
  19. Niwa H. Endoscopy of the colon. Gastroenterol Endosc Tokyo 1965; 7:402-408.
  20. Vining DJ, Shifrin RY, Grishaw EK, Liu K, Gelfand DW. Virtual colonoscopy (abstr). Radiology 1994; 193(P):446.
  21. Luboldt W, Bauerfeind P, Steiner P, Fried M, Krestin GP, Debatin JF. Preliminary assessment of three-dimensional magnetic resonance imaging for various colonic disorders. Lancet 1997; 349:1288-1291.[CrossRef][Medline]
  22. MacFarland EG. Reader strategies for CT colonography. Abdom Imaging 2002; 27:275-283.[Medline]
  23. Lauenstein TC, Debatin JF. Magnetic resonance colonography for colorectal cancer screening. Semin Ultrasound CT MR 2001; 22:443-453.[CrossRef][Medline]
  24. Macari M, Lavell M, Pedrosa I. Effect of different preparations on residual fluid at CT colonography. Radiology 2001; 218:274-277.[Abstract/Free Full Text]
  25. Yee J, Hung RK, Akerkar GA, Wall SD. The usefulness of glucagon hydrochloride for colonic distention in CT colonography. AJR Am J Roentgenol 1999; 173:169-172.[Abstract/Free Full Text]
  26. Luboldt W, Fletcher JG, Vogl TJ. Colonography: current status, research directions and challenges: update 2002. Eur Radiol 2002; 12:502-524.[Medline]
  27. Fletcher JG, Luboldt W. CT colonography and MR colonography: current status, research directions and comparison. Eur Radiol 2000; 10:786-801.[CrossRef][Medline]
  28. Van Gelder RE, Venema HW, Serlie IWO, et al. CT colonography at different radiation dose levels: feasibility of dose reduction. Radiology 2002; 224:25-33.[Abstract/Free Full Text]
  29. Hara AK, Johnson CD, McCarty RL, Welch TJ, McCullough CH, Harsens WS. CT colonography: single- versus multi-detector row imaging. Radiology 2001; 219:461-465.[Abstract/Free Full Text]
  30. Morrin MM, Farrel RJ, Kruskal JB, Reynolds K, McGee JB, Raptopoulos V. Utility of intravenously administered contrast material at CT colonography. Radiology 2000; 217:765-771.[Abstract/Free Full Text]
  31. Oto A, Gelebek V, Oguz BS, et al. CT attenuation of colorectal polypoid lesions: evaluation of contrast enhancement in CT colonography. Eur Radiol 2003; 13:1657-1663.[CrossRef][Medline]
  32. Lefere PA, Gryspeerdt SS, Dewyspelaere J, Baekelandt M, Holsbeeck BG. Dietary fecal tagging as a cleansing method before CT colonography: initial results—polyp detection and patient acceptance. Radiology 2002; 224:393-403.[Abstract/Free Full Text]
  33. Thomeer M, Carbone I, Bosmans H, et al. Stool tagging in thin-slice multidetector computed tomography colonography. J Comput Assist Tomogr 2003; 27:132-139.[CrossRef][Medline]
  34. Pineau BC, Paskett ED, Chen GJ, et al. Virtual colonoscopy using oral contrast compared with colonoscopy for the detection of patients with colorectal polyps. Gastroenterology 2003; 125:304-310.[CrossRef][Medline]
  35. Saar B, Heverhagen JT, Obst T, et al. Magnetic resonance colonography and virtual magnetic resonance colonoscopy with the 1.0-T system: a feasibility study. Invest Radiol 2000; 35:521-526.[CrossRef][Medline]
  36. Pappalardo G, Polettini E, Frattaroli FM, et al. Magnetic resonance colonography versus conventional colonoscopy for the detection of colonic endoluminal masses. Gastroenterology 2000; 119:300-304.[CrossRef][Medline]
  37. Morrin MM, Hochman MG, Farrell RJ, Marquesuzaa H, Rosenberg S, Edelman RR. MR colonography using colonic distention with air as the contrast material: work in progress. AJR 2001; 176:144-146.[Free Full Text]
  38. Lomas DJ, Sood RR, Graves MJ, Miller R, Hall NR, Dixon AK. Colon carcinoma: MR imaging with CO2 enema—pilot study. Radiology 2001; 219:558-562.[Abstract/Free Full Text]
  39. So NM, Lam WW, Mann D, Leung KL, Metreweli C. Feasability study of using air as a contrast medium in MR colonography. Clin Radiol 2003; 58:555-559.[CrossRef][Medline]
  40. Luboldt W, Bauerfeind P, Wildermuth S, Marincek B, Fried M, Debatin JF. Colonic masses: detection with MR colonography. Radiology 2000; 216:383-388.[Abstract/Free Full Text]
  41. Lauenstein TC, Herborn CU, Vogt FM, Gohde SC, Debatin JF, Ruehm SG. Dark lumen MR-colonography: initial experience. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 2001; 173:785-789.[Medline]
  42. Lauenstein TC, Goehde SC, Ruehm SG, Holtmann G, Debatin JF. MR colonography with barium based fecal tagging: initial clinical experience. Radiology 2002; 223:248-254.[Abstract/Free Full Text]
  43. Calhoun PS, Kuszyk BS, Heath DG, Carley JC, Fishman EK. Three-dimensional volume rendering of spiral CT data: theory and method. RadioGraphics 1999; 19:745-764.[Abstract/Free Full Text]
  44. Cody DD. AAPM/RSNA physics tutorial for residents: topics in CT. Image processing in CT. RadioGraphics 2002; 22:1255-1268.
  45. Tomandl BF, Hastreiter P, Rezk-Salama C, et al. Local and remote visualization techniques for interactive direct volume rendering in neuroradiology. RadioGraphics 2001; 21:1561-1572.[Abstract/Free Full Text]
  46. Hernandez-Hoyos M, Orkisz M, Puech P, Mansard-Desbleds C, Douek P, Magnin IE. Computer-assisted analysis of three-dimensional MR angiograms. RadioGraphics 2002; 22:421-436.[Abstract/Free Full Text]
  47. Rubin GD, Beaulieu CF, Argiro V, et al. Perspective volume rendering of CT and MR images: applications for endoscopic imaging. Radiology 1996; 199:321-330.[Abstract/Free Full Text]
  48. Vos F, Serlie I, Van Gelder R, Stoker J, Vrooman H, Post F. A review of technical advances in virtual colonoscopy. Medinfo 2001; 10:938-942.[Medline]
  49. Rogalla P, Bender A, Bick U, Huitema A, Terwisscha van Scheltinga J, Hamm B. Tissue transition projection (TTP) of the intestines. Eur Radiol 2000; 10:806-810.[CrossRef][Medline]
  50. Luboldt W, Luz O, Vonthein R, et al. Three-dimensional double-contrast MR colonography: a display method simulating double-contrast barium enema. AJR 2001; 176:930-932.[Free Full Text]
  51. Ott DJ. Accuracy of double-contrast barium enema in diagnosing colorectal polyps and cancer. Semin Roentgenol 2000; 4:333-341.
  52. Ott DJ. Role of the barium enema in colorectal carcinoma. Radiol Clin N Am 1993; 31:1293-1313.[Medline]
  53. Smith C. Colorectal cancer. Radiol Clin N Am 1997; 35:439-456.[Medline]
  54. Glick S, Wagner JL, Johnson CD. Cost-effectiveness of double-contrast barium enema in screening for colorectal cancer. AJR 1998; 170:629-636.[Free Full Text]
  55. Glick S. Double-contrast barium enema for colorectal cancer screening: a review of the issues and a comparison with other screening alternatives. AJR 2000; 174:1529-1537.[Free Full Text]
  56. Gazelle GS, McMahon PM, Scholz FJ. Screening for colorectal cancer. Radiology 2000; 215:327-335.[Abstract/Free Full Text]
  57. Geenen RWF, Hussain SM, De Graaf M, et al. The impact of double contrast barium enema (DCBE) with double reading in patients with suspected colon disease: a retrolective cohort study with long term follow up—preliminary results (abstr). Radiology 2002; 225(P):243.
  58. Rex DK. Colonoscopy. Gastrointest Endosc Clin N Am 2000; 10:135-160.[Medline]
  59. Anderson ML, Heigh RI, McCoy GA, et al. Accuracy of assessment of the extent of examination by experienced colonoscopists. Gastrointest Endosc 1992; 38:560-563.[Medline]
  60. Angtuaco TL, Banaad-Omniotek GD, Howden CW. Differing attitudes toward virtual and conventional colonoscopy for colorectal cancer screening: surveys among primary care physicians and potential patients. Am J Gastroenterol 2001; 96:887-893.[CrossRef][Medline]
  61. Steine S. Which hurts the most? A comparison of pain rating during double-contrast barium enema examination and colonoscopy. Radiology 1994; 191:99-101.[Abstract/Free Full Text]
  62. Svensson MH, Svensson E, Lasson A, Hellstrom M. Patient acceptance of CT colonography and conventional colonoscopy: prospective comparative study in patients suspected of having colorectal disease. Radiology 2002; 222:337-345.[Abstract/Free Full Text]
  63. Thomeer M, Bielen D, Vanbeckevoort D, et al. Patient acceptance for CT colonography: what is the real issue? Eur Radiol 2002; 12:1410-1415.[CrossRef][Medline]
  64. Ristvedt SL, McFarland EG, Weinstock LB, Thyssen EP. Patient preferences for CT colonography, conventional colonoscopy and bowel preparation. Am J Gastroenterol 2003; 98:578-585.[CrossRef][Medline]
  65. Gluecker TM, Johnson CD, Harmsen WS. Colorectal cancer screening with CT colonography, colonoscopy, and double-contrast barium enema examination: prospective assessment of patient perceptions and preferences. Radiology 2003; 227:378-384.[Abstract/Free Full Text]
  66. Taylor SA, Halligan S, Saunders BP, Vance M, Bartram CI. The psychological acceptability of multiditector-row CT colonograophy compared to barium enema, flexible sigmoidoscopy and colonoscopy (abstr). Eur Radiol 2003; 13(suppl):S115-S116.
  67. Akerkar GA, Yee J, Hung R, McQuaid K. Patient experience and preferences toward colon cancer screening: a comparison of virtual colonoscopy and conventional colonoscopy. Gastrointest Endosc 2001; 54:310-315.[CrossRef][Medline]
  68. Macari M, Bini EJ, Xue X, et al. Colorectal neoplasms: prospective comparison of thin-section low-dose multi-detector row CT colonography and conventional colonoscopy for detection. Radiology 2002; 224:383-392.[Abstract/Free Full Text]
  69. Gluecker T, Dorta G, Keller W, Jornod P, Meuli R, Schnyder P. Performance of multidetector computed tomography colonography compared with conventional colonoscopy. Gut 2002; 51:207-211.[Abstract/Free Full Text]
  70. Johnson CD, Harmsen WS, Wilson LA, et al. Prospective blinded evaluation of computed tomographic colonography for screen detection of colorectal polyps. Gastroenterology 2003; 125:311-319.[CrossRef][Medline]
  71. Stoker J, Van Gelder RE, Glas AS, Venema HW. Radiation in CT colonography: an overview of effective doses calculated from scan parameters reported in the literature (abstr). Eur Radiol 2003; 13(suppl):S94.
  72. Lefere P, Gryspeerdt S, Baekelandt M, Dewyspelaere J, Van Holsbeeck B. Diverticular disease in CT colonography. Eur Radiol 2003. Available at: http://www.springerlink.com/app/home/issue.asp. Accessed July 17, 2003.
  73. Schreyer AG, Seitz J, Herfarth H, Feuerbach S. A comparison of MRI based colonography with conventional colonoscopy in patients with inflammatory bowel disease (abstr). Eur Radiol 2003; 13(suppl):S104.
  74. Ajaj W, Lauenstein T, Papanikolaou N, Debatin J, Ruehm SG. Dark lumen MR colonography: comparison to standard colonoscopy (abstr). Eur Radiol 2003; 13(suppl):S104-S105.



This article has been cited by other articles:


Home page
 GutHome page
C. A Kuehle, J. Langhorst, S. C Ladd, T. Zoepf, M. Nuefer, F. Grabellus, J. Barkhausen, G. Gerken, and T. C Lauenstein
Magnetic resonance colonography without bowel cleansing: a prospective cross sectional study in a screening population
Gut, August 1, 2007; 56(8): 1079 - 1085.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
C. A. Kuehle, P. Veit, G. Antoch, F. Grabellus, P. Robert, T. Beyer, and C. U. Herborn
Contrast-Enhanced Dark Lumen PET/CT and MR Colonography in a Rodent Polyp Model: Initial Results with Histopathologic Correlation
Am. J. Roentgenol., October 1, 2005; 185(4): 1045 - 1047.
[Abstract] [Full Text] [PDF]

Home page
 ANN INTERN MEDHome page
B. P. Mulhall, G. R. Veerappan, and J. L. Jackson
Meta-Analysis: Computed Tomographic Colonography
Ann Intern Med, April 19, 2005; 142(8): 635 - 650.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
M. A. Barish, J. A. Soto, and J. T. Ferrucci
Consensus on Current Clinical Practice of Virtual Colonoscopy
Am. J. Roentgenol., March 1, 2005; 184(3): 786 - 792.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow MPEG movies
Right arrow All Versions of this Article:
e18v1
24/1/e18    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal