British Journal of Radiology (2005) 78, 202-206
© 2005 British Institute of Radiology
doi: 10.1259/bjr/15237296
MR colonography: baseline appearance of the unprepared rectosigmoid
V L Jardine, MA
E Sala, MD, PhD, FRCR
and
D J Lomas, FRCP, FRCR
Department of Radiology, University of Cambridge & Addenbrooke's Hospital, Cambridge, UK
Correspondence: Prof. David J Lomas, University Department of Radiology, Box 219, Level 5, Addenbrooke's Hospital, Cambridge CB2 2QQ, UK
 |
Abstract
|
|---|
A retrospective review of 100 routine pelvic MR examinations was conducted to evaluate the appearance of the unprepared rectosigmoid colon, with the aim of informing future strategies for minimal preparation MR colonography. All examinations were reviewed by two observers in consensus, and included matched-location axial T1 weighted and T2 weighted fat suppressed fast spin echo (FSFSE) images. Analysis revealed that the overall appearance of the faecal material in the colon could simulate tumour in 80% of T1 weighted and 17% of T2 weighted images. By matching the images from the two sequences for each patient the faeces had an overall appearance that would mimic tumour in only 7% of cases. However, luminal tumour-mimicking foci of signal occurred frequently, present in 91% of T1 weighted and 85% of T2 weighted studies. The results indicate that if bowel-cleansing regimens are to be avoided for MR colonography, effective strategies such as dietary restriction and use of oral contrast agents will be required to reduce luminal signal on T2 weighting and eliminate polyp-mimicking foci. The results also suggest that T2 weighted strategies should be further investigated and that combination with T1 weighted imaging may improve discrimination of lesions from normal faecal material.
|
Introduction
|
|---|
Colorectal cancer is the second leading cause of death due to cancer in the UK [1] and USA [2, 3], estimated to cause nearly half a million deaths worldwide each year [4]. The adenoma to carcinoma progression is well understood [5], occurring over several years [6] and providing the opportunity to detect and remove pre-cancerous lesions at a potentially curative stage [4, 7]. As a result several western countries, including the UK, are currently evaluating screening strategies as part of their public health programs.
Population screening ideally requires a sensitive, specific, safe and inexpensive screening test acceptable to the population into which it is introduced. Several methods for detecting colonic polyps and tumours are currently available, including faecal occult blood testing (FOBT) and imaging methods such as colonoscopy, sigmoidoscopy, barium enema and CT colonography.
Current proposed strategies suggest a cheap, simple and relatively insensitive non-imaging method such as FOBT [8] (or in the future faecal deoxyribonucleic acid (DNA) [9] or protein [10] testing) for initial screening. Subsequently, anatomical imaging methods with higher costs and morbidity would be utilized in a selected subpopulation to confirm the presence and location of any lesion. A constraint of current colonic imaging methods is their reliance on cleansing regimens, which are often uncomfortable for patients and associated with their own morbidity [11], especially in a relatively elderly target population. This is a substantial limitation since it may reduce compliance and influence accuracy, in turn undermining the benefit of a population-screening program. However colonic cleansing remains routine since the presence of normal colonic faecal material can compromise diagnostic performance by mimicking or masking tumours and polyps on all imaging techniques [12].
MRI offers a method for imaging the colon theoretically comparable with CT colonography, but without the burden of ionizing radiation. Initial MR feasibility studies have used air or fluid for colonic distension after cleansing procedures [13–17] to remove faecal material. However a key advantage of MRI is the ability to obtain a wide range of soft tissue contrast and imaging based on different signal characteristics. This has stimulated the investigation of minimal preparation techniques for MR colonography that might avoid colonic cleansing, by utilizing oral agents and/or dietary manipulation to "tag" stool [18–21]. The aim is to consistently change the signal of faeces to prevent it from mimicking or masking tumour, the ideal result being homogeneous material of substantially different signal intensity to the bowel wall and potential lesions. Despite these early attempts to manipulate signal in the colon there has been little work on the "normal" or baseline appearances of the colon at MR, or CT examination.
The aim of this study is to evaluate the appearance of the unprepared rectosigmoid colon and the ability of faeces to mimic tumour in a series of patients undergoing routine pelvic MRI examination. The findings will be used to inform future preparation strategies for an MR colonography technique.
|
Methods and materials
|
|---|
A retrospective review of consecutive routine pelvic MR examinations obtained between July 2001 and September 2002 (for conditions unrelated to the gastrointestinal tract) was undertaken to obtain 50 male and 50 female studies with faeces within the lumen of the colon. Since the aim of the study was to assess the appearance of the "normal" faecal material within the rectosigmoid colon the following exclusion criteria were used: empty rectosigmoid, previous rectosigmoid surgery, bowel pathology, pelvic masses distorting the sigmoid colon, previous pelvic radiotherapy or chemotherapy within a year of the examination. Also studies rendered technically inadequate for analysis by motion or prosthesis-related artefacts were excluded.
All examinations were performed using a 1.5 T whole body MR system (GE Medical Systems, Milwaukee, WI) with a pelvic phased array coil. Matched location axial T1 weighted (echo time (TE) 14 ms, repetition time (TR) 700–750 ms) and fat suppressed T2 weighted fast spin echo (FSE) (8 echo train length, TE effective 90 ms, TR 3500 ms) images with 256 x 256 matrix, 5 mm section thickness and a 2.5 mm interslice gap were obtained in each case, covering the pelvis inferior to the aortic bifurcation. The rectum and distal sigmoid colon within the pelvis were evaluated by two observers in consensus for the following features of the intraluminal faecal material: overall contrast relative to the bowel wall, heterogeneity, focal areas of increased or decreased signal that could mimic polypoid lesions, luminal distension of the whole rectosigmoid, and the presence of any free fluid. Statistical analysis was performed using a non-parametric Wilcoxon signed rank test.
Where the colonic wall could not be clearly identified, adjacent gluteal muscle was used as a reference to assess the relative signal intensity of the intraluminal contents since skeletal muscle is of similar signal intensity to the smooth muscle of the colon wall. Focal increases or decreases in signal intensity within the lumen were recorded for areas of bowel containing faeces, not gas. Free fluid was defined as the presence of a gravity dependent fluid level or high T2 signal collection (comparable with that of the bladder) within the rectosigmoid.
|
Results
|
|---|
100 examinations of 50 male and 50 female patients aged between 18 years and 87 years (mean age 58 years) were available for assessment. The rectosigmoid colon was fully distended in only 3% of the studies. Free fluid was not present in any of the examinations assessed.
The overall signal intensity of the faecal material relative to the bowel wall for T1 and T2 weighted imaging is shown in Figure 1a
. The majority of T1 weighted images showed luminal material of signal intensity similar to that of the bowel wall (70%), whereas most T2 weighted images demonstrated luminal signal less than that of the wall (83%). The signal intensity of tumour is similar to that of the bowel wall, typically slightly increased in T2 weighted and decreased in T1 weighted images [22, 23]. Interpreting the results with this in mind (Figure 1b): in 80% of T1 weighted studies the overall signal of faeces could mimic tumour, whereas this was the case in significantly fewer T2 weighted studies (z=6.833, p<0.0001): just 17%. Furthermore, matching the T1 and T2 weighted images for each examination leaves only 7% with faeces whose overall signal intensity could mimic tumour, significantly less than T2 weighted alone (z=3.317, p<0.001).
View larger version (13K):
[in this window]
[in a new window]
|
Figure 1. (a) The overall signal intensity of intraluminal material relative to the bowel wall for both T1 and T2 weighted studies. (b) The capacity of this luminal signal material to mimic tumour overall for T1 weighted and T2 weighted strategies individually, and for the combined analysis of T1 and T2 weighting. Wilcoxon signed rank test statistics were used to detect significant differences: intraluminal material of signal intensity similar to that of tumour occurred less frequently on T2 than T1 weighted imaging, with combined analysis providing a lower incidence of such tumour-mimicking material than either T1 weighted or T2 weighted alone (p<0.001 in each case).
|
|
Intraluminal faecal material of the rectosigmoid was considered heterogeneous in 97% of T1 and 89% of T2 weighted examinations (Figure 2). Foci of signal in the lumen (small areas of different signal intensity to the majority of faeces) occurred in most examinations, with the majority of T1 weighted images demonstrating luminal foci of decreased signal and most T2 weighted images foci of increased signal (Figure 3a). On the same basis as above, luminal foci of greater signal intensity than the colon wall on T1 weighting are unlikely to be tumour, and similarly relatively low signal foci on T2 weighting are unimportant. However 91% of T1 weighted studies and 85% of T2 weighted studies, contained foci whose signal was considered similar to that of tumour (Figure 3b). Even after matching T1 weighted and T2 weighted studies 80% of examinations contained luminal foci capable of mimicking polyps (Figure 4).
View larger version (51K):
[in this window]
[in a new window]
|
Figure 2. Location matched (a) T1 weighted and (b) T2 weighted images of the rectum and pelvic sigmoid colon with heterogeneous faecal material. These images also demonstrate the potential value of comparing multiple sequences to aid differentiation between lesions and intraluminal faeces. On T1 weighting (a) the area indicated (thick arrow) is of similar signal to the bowel wall and therefore could be tumour. T2 weighted (b) however demonstrates good negative luminal contrast in this region: no tumour is present. However this strategy is clearly limited: it would be impossible to exclude a lesion in either image at the area indicated by the thin arrow.
|
|
View larger version (13K):
[in this window]
[in a new window]
|
Figure 3. (a) The presence of focal areas of luminal increased or decreased signal intensity compared with the remainder of intraluminal material. The majority of examinations demonstrated heterogeneous material, with foci of increased, decreased, or both increased and decreased signal present in T1 and T2 weighted studies. (b) The capacity of luminal foci to mimic colonic lesions in T1 and T2 weighted strategies.
|
|
View larger version (47K):
[in this window]
[in a new window]
|
Figure 4. Location matched (a) T1 and (b) T2 weighted images of the rectosigmoid colon. A focus of slightly increased signal intensity relative to the bowel wall is demonstrated within the lumen on T2 weighting, whereas on T1 weighting this area is of similar signal to the wall. Consequently differentiation between faeces and a lesion is impossible.
|
|
|
Discussion
|
|---|
This study was performed to inform which strategies for manipulating intraluminal signal are most likely to succeed in the development of a MR colonography technique that avoids colonic cleansing. Ideally any oral contrast agents and dietary restriction strategies must be safe, cheap, palatable and influence luminal signal consistently and evenly throughout the entire colon, providing maximal contrast between the material and potential lesions. The use of a minimal preparation strategy should also allow easy identification of those who have failed to follow the preparation requirements to reduce the risk of misinterpretation. Furthermore it is clearly fundamental that any preparation strategy must be better tolerated than the standard colonic cleansing methods or there would be little benefit.
Lauenstein et al [18, 20] have investigated the use of barium sulphate as a negative oral contrast agent, utilizing a T1 weighted three-dimensional (3D) gradient-echo imaging strategy with colonic distension obtained by means of a water enema. This T1 weighted "dark lumen" technique requires intravenous gadobenate dimeglumine to provide adequate contrast between the lumen and enhancing lesions which would otherwise appear relatively low signal. The same group have also tried oral gadoterate meglumine with a T1 weighted imaging strategy in conjunction with a dilute gadolinium based water enema [19]. Using this technique tumour appears low signal against a high signal lumen.
Our work indicates that the overall signal of faecal material in the majority of T1 weighted images is of similar signal intensity to the bowel wall. This is highly problematic since tumour signal is likely to be the same or slightly less than that of the wall on T1 weighted imaging, so material would commonly mimic lesions. For T1 weighted imaging without intravenous enhancement the maximum contrast between material and lesions will occur when the luminal signal is substantially greater than that of the bowel wall, but does not obscure haustral folds or small polyps. T1 weighted imaging sequences are currently relatively attractive in terms of their performance, providing good volume coverage within acceptable breath-holding times. This is mainly as a result of the development of angiographic techniques designed to suppress signal from longer T1 weighted tissues and detect the shortened T1 in blood resulting from intravenous gadolinium enhancement. Lauenstein et al [18] have pursued the same strategy by enhancing polyps and tumours using intravenous gadolinium media against a dark lumen background with moderate success using barium as an oral contrast agent combined with a low fat, low residue diet. However this strategy requires a pre and post intravenous comparison study typically limiting the examination to one orientation (i.e. prone or supine), which may in turn limit evaluation.
In this study the overall signal of faeces on T2 weighted imaging was lower than that of the bowel wall in the majority of cases, and since the signal of tumour is likely to be the same or slightly greater than the wall a preparation strategy to reduce luminal signal should have the least difficulty in maximizing contrast between material and lesions. Thus a T2 weighted approach may also prove feasible for minimal preparation MR colonography, provided that the preparation successfully eliminates all the foci of increased signal. These are most likely related to water-retaining high residue fruit and fibrous dietary components that are not fully digested in the proximal intestine. T2 weighted sequences are generally of lower signal to noise ratio than T1 weighted sequences and so obtaining adequate coverage within a single breath-hold is likely to be more difficult, although improvements in coil design and system efficiency are gradually overcoming these limitations. Additionally the longer TE employed with T2 weighted imaging could increase the likelihood of susceptibility artefacts in relation to distension with gases. However, this could be counterbalanced by the use of a T2 weighted spin echo technique compared with the gradient echo techniques use for T1 weighting.
A further strategy for MR colonography was implicit from the significant decrease in the incidence of examination in which overall luminal signal was consistent with that of tumour when matched-location T1 and T2 weighted images were compared. This suggests that using multiple contrast imaging could be of value, especially if luminal signal manipulation is successful for all the sequences examined. Clearly this raises issues on the effective review of examinations concerning the avoidance of image misregistration and the correlation of two or more sets of images.
This retrospective study does have limitations, primarily that only the rectosigmoid colon was studied. Although the majority of tumours occur in the rectosigmoid it is possible that the appearances in the remainder of the colon may differ. For instance, the hydration of the intraluminal content would be expected to decrease along the length of the colon, which may have biased the results in favour of the T2 weighted approach described above. This could be enhanced by fasting patients prior to MR colonography and thereby promoting the dehydration of faeces. We did not investigate the diets of those taking part assuming that, as a cross-section of our referral population for pelvic examination, they would reflect a similar range of dietary habits encountered in a population referred for examination of the colon. Similarly patient drug history was not evaluated, although this could also have influenced intraluminal signal. A referred population for screening would inevitably contain individuals taking a range of drugs although this would be likely to be influenced by how the screening population is selected. Finally, although all patients with known colonic pathology were excluded from the study, it is possible that some of the cases reviewed had asymptomatic lesions of the rectosigmoid. Therefore some of the foci of signal noted could have been "false positives", with intraluminal faeces being adequate for image evaluation. The incidence of such lesions would be low, and given the overall results the impact on the findings in this study is likely to be negligible.
The results above suggest that dietary control will be essential as part of any strategy to avoid traditional colonic cleansing. Different food constituents are known to have particular MR characteristics, fats for instance typically giving high signal on T1 weighting, and water retaining materials such as fruit often producing high signal on T2 weighting, low on T1 weighting. Future work will require optimization of suitable diets that can increase the ability to discriminate faeces from tumour and the bowel wall. Ideally the diet will promote homogeneity within the faecal material, avoiding undigested residues that may retain water and produce tumour-mimicking foci. Both diet and oral contrast agents will also be influenced by the individual variation in intestinal transit time and the optimal combinations of diet and oral contrast agent need to be identified.
In conclusion, this study indicates that the unprepared rectosigmoid colon typically contains faeces that is heterogeneous on MRI, potentially mimicking tumour in the majority of cases on T1 and T2 weighted imaging either by overall signal intensity or by the presence of foci within intraluminal material. The results suggest that both dietary control and oral contrast agent preparation are essential if colonic cleansing is to be avoided, and that strategies using T2 weighted and combinations of T1 weighted and T2 weighted imaging require further investigation and optimization.
|
Acknowledgments
|
|---|
The authors would like to acknowledge the assistance of the Fund for Addenbrooke's, Elmore Fund, James Baird Fund and the Radiographers of the MRIS Unit.
Received for publication January 5, 2004.
Revision received August 18, 2004.
Accepted for publication October 13, 2004.
|
References
|
|---|
- Office for National Statistics. Mortality Statistics 1998. ONS, 1998. Series DH2 no 25.
- American Cancer Society, Cancer, facts and figures. GA:ACS, 2001. Atlanta GA: American Cancer Society, 2001:5.
- Reis LAG, Eisner MP, Kosary CL, et al. SEER cancer statistics review. 1973–2000. Bethesda MD: National Cancer Institute, 2000.
- Boyle P, Leon ME. Epidemiology of colorectal cancer. Br Med Bull 2002;64:1–25.[Abstract/Free Full Text]
- Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990;61:759–67.[CrossRef][Medline]
- O'Brein MJ, Winawer SJ, Zauber AG. The National Polyp Study. Patient and polyp characteristics associated with high-grade dysplasia in colorectal adenomas. Gasteroenterology 1990;98:371–9.[Medline]
- Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. New Engl J Med 1993;329:1977–81.[Abstract/Free Full Text]
- Lowenfels AB. Fecal occult blood testing as a screening procedure for colorectal cancer. Ann Oncol 2002;13:40–3.[Free Full Text]
- Traverso G, Shuber A, Levin B. Detection of APC mutations in fecal DNA from patients with colorectal tumors. New Engl J Med 2002;346:311–20.[Abstract/Free Full Text]
- Davies RJ, Freeman A, Morris LS, et al. Analysis of minichromosome maintenance protiens as a novel method for detection of colorectal cancer in stool. Lancet 2002;359:1917–9.[CrossRef][Medline]
- Toledo TK, Dipalma JA. Review article: colon cleansing preparation for gastrointestinal procedures. Aliment Pharmacol Ther 2001;15:605–11.[CrossRef][Medline]
- Lefere PA, Gryspeerdt SS, Dewyspelaere J, et al. 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]
- Lomas DJ, Sood RR, Graves MJ, et al. Colon carcinoma: MR imaging with CO2 enema - pilot study. Radiology 2001;219:558–62.[Abstract/Free Full Text]
- Luboldt W, Steiner P, Bauerfeind P, et al. Detection of mass lesions with MR colonography: preliminary report. Radiology 1998;207:59–65.[Abstract/Free Full Text]
- Luboldt W, Bauerfeind P, Wildermuth S, et al. Colonic masses: detection with MR colonography. Radiology 2000;216:383–8.[Abstract/Free Full Text]
- Morrin MM, Hochman MG, Farrell RJ, et al. MR colonography using colonic distension with air as the contrast material: work in progress. AJR Am J Roentgenol 2001;176:144–6.[Free Full Text]
- Lauenstein TC, Herborn CU, Vogt FM, et al. Dark lumen MR-colonography: initial experience. Rofo 2001;173:785–9.[Medline]
- Lauenstein TC, Goehde SC, Ruehm SG, et al. MR colography with barium-based fecal tagging: initial clinical experience. Radiology 2002;223:248–54.[Abstract/Free Full Text]
- Weishaupt D, Patak MA, Froehlich J, et al. Faecal tagging to avoid colonic cleansing before MRI colography. Lancet 1999;354:835–6.[Medline]
- Lauenstein TC, Holtmann G, Schoenfelder D, et al. MR colonography without colonic cleansing: a new strategy to improve patient acceptance. AJR Am J Roentgenol 2001;177:823–7.[Abstract/Free Full Text]
- Ajaj W, Pelster G, Treichel U, et al. Dark lumen magnetic resonance colonography: comparison with conventional colonoscopy for the detection of colorectal pathology. Gut 2003;52:1738–43.[Abstract/Free Full Text]
- Bottomley PA, Hardy CJ, Argersinger RE, et al. A review of 1H nuclear magnetic resonance relaxation in pathology: are T1 and T2 diagnostic? Med Phy 1987;14:1–37.
- Bottomley PA, Foster TH, Argersinger RE, et al. A review of normal tissue hydrogen NMR relaxation times and relaxation mechanisms from 1-100 MHz: dependence on tissue type, NMR frequency, temperature, species, excision and age. Med Phys 1984;11:425–48.[CrossRef][Medline]
This article has been cited by other articles:
|
|
|
|
 
C. Torres, S. Szomstein, and S. D. Wexner
Virtual Colonoscopy in Colorectal Cancer Screening
Surgical Innovation,
March 1, 2007;
14(1):
27 - 34.
[Abstract]
[PDF]
|
|
|
|
|
|
|
BJR Review of the Year - 2005.
Br. J. Radiol.,
March 1, 2006;
79(939):
183 - 187.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
