| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Academic Medical Center, G1-211, PO Box 22700,
Amsterdam 1100 DE, The Netherlands.
2 Department of Radiology, Leiden University Medical Center, Leiden, The
Netherlands.
3 Department of Radiology, Erasmus MC-University Medical Center, Rotterdam, The
Netherlands.
4 Department of Gastroenterology and Hepatology, Erasmus MC-University Medical
Center, Rotterdam, The Netherlands.
Received September 14, 2004;
accepted after revision March 3, 2005.
The contrast agents used in this study were provided free of charge by
formerly Nycomed Amersham, now GE Healthcare, Oslo, Norway.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Forty-eight patients in two hospitals who had clinically suspected exacerbation of Crohn's disease were included in this study. In three levels of thickened small-bowel wall, axial dynamic T1-weighted sequences were performed every 4-6 sec for a total duration of 2-3 min after contrast administration; static T1-weighted turbo spin-echo sequences were acquired both before and after contrast administration. The slope of enhancement, enhancement ratio, time to enhancement, enhancement time, and thickness of the small-bowel wall were determined. These MRI results were compared with overall clinical grade, Crohn's disease activity index (CDAI), and Van Hees activity index. Clinical grade was based on clinical information, physical findings, laboratory studies, endoscopy, surgery, and other imaging studies. Spearman's correlation coefficient and p values were determined per hospital. Fisher's z-transformation was applied before pooling the correlation coefficients from both hospitals.
RESULTS. The enhancement ratio based on the static series showed significant correlation with the clinical grade (r = 0.29, p = 0.045), CDAI (r =0.31, p = 0.033), and Van Hees activity index (r = 0.36, p = 0.016). The enhancement ratio based on the dynamic series correlated significantly with the CDAI (r = 0.38, p = 0.016). Wall thickness correlated significantly with clinical grade (r = 0.47, p = 0.003) and Van Hees activity index (r = 0.41, p = 0.007).
CONCLUSION. These data suggest that the enhancement ratio of bowel wall after IV administration of gadodiamide and bowel wall thickness are weak to moderate indicators of the severity of Crohn's disease.
Keywords: Crohn's disease • dynamic MRI • gastrointestinal imaging • inflammatory bowel disease • MRI • small bowel
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A simple technique that meets these requirements is not presently available. Ileocolonoscopy (including biopsies) and, to a lesser extent, enteroclysis are used for establishing disease activity. Drawbacks of ileocolonoscopy are the small proportion of small bowel that can be visualized, the risk of procedure-related complications, and the low patient acceptance because of discomfort during the procedure and the extensive bowel preparation that is necessary [4]. MRI has been advocated as a noninvasive tool in the diagnostic workup of Crohn's disease. Studies have shown that MRI can determine local activity of Crohn's disease [5, 6], and even the degree of activity [7-14], on the basis of an increased enhancement in the more severely inflamed bowel wall on T1-weighted sequences after gadodiamide injection.
Because local vascularization increases with the severity of the disease [15], it is likely that not only the degree of enhancement—the enhancement ratio (ER)—but also the steepness of the enhancement curve—or slope of enhancement (SoE)—as measured on dynamic scanning, is higher in more severely inflamed tissues. In arthritis, which, like Crohn's disease, is a chronic inflammatory disease, dynamic MR enhancement curves have been associated with laboratory and clinical indicators of inflammation and information provided on patient questionnaires [16]. A limited number of studies have shown that in Crohn's disease a correlation exists between the severity of inflammation and the ER [7, 9, 11], but to our knowledge no study has evaluated the SoE as an indicator of disease activity. Measuring the signal increase of the bowel wall using quantitative parameters after the administration of IV gadodiamide possibly has the additional advantage of providing a reviewer-independent interpretation of the MR images, making this technique ideal for the follow-up of patients with chronic inflammatory diseases.
The purpose of this study, therefore, is to evaluate the role of dynamic contrast-enhanced MRI in determining disease activity in patients with Crohn's disease.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The aim was to enroll 62 patients in a 15-month period; however, because the inclusion rate was slower than expected, the inclusion was halted. Therefore, 52 patients were included in this study during a 26-month period: 29 patients in hospital 1 and 23 in hospital 2. Four of the 52 patients who entered the study were excluded during the study. Three patients ingested less than 400 mL of the oral contrast medium and one patient did not receive gadodiamide. The total study population therefore constituted 48 individuals (14 men and 34 women) having a mean age of 37.2 ± 13.0 (SD) years (range, 18-68 years).
Patient Preparation
Before imaging, 400-800 mL of the oral contrast medium ferristene
(Abdoscan, GE Healthcare) was administered in four volumes of 200 mL each at
30-min intervals. Forty-three patients ingested a volume of 800 mL of oral
contrast medium. The remaining five ingested at least 400 mL. The oral
contrast medium was administrated to facilitate identification of thickened
bowel wall segments and subsequent measurements. To prevent motion artifacts,
20 mg of butylscopolamine bromide (Buscopan, Boehringer Ingelheim) was
injected intramuscularly in 19 of 26 patients studied at hospital 1 before the
start of the first sequence and in 24 of 26 patients directly before the
dynamic sequence (19/24 received the spasmolytic agent twice). The decision to
administer bowel relaxants was based on the presence of bowel motion artifacts
on the scout views. Because all patients in hospital 2 were scanned on a 0.5-T
scanner that is less sensitive to motion artifacts, none of the patients
studied there received butylscopolamine bromide. Immediately before the start
of the dynamic series, IV gadodiamide (Omniscan, GE Healthcare), 0.1 mmol/kg
of body weight, was injected manually as a fast bolus.
MRI Protocol
MRI was performed at 1.0 T (Gyroscan NT 10, Philips Medical Systems) using
a body quadrature coil in 26 patients at hospital 1 and at 0.5 T (Philips
Gyroscan T5II) using a body wraparound surface coil in 22 patients at hospital
2.
Before the administration of gadodiamide, an axial respiratory-triggered T2-weighted turbo spin-echo sequence (TR range/TE range, 1,800-2,000/90-120; number of excitations, 4-6; field of view, 375 mm; matrix, 205 x 256; slice thickness, 10 mm; slices, 24; acquisition time, 180 sec) and axial and coronal breath-hold T1-weighted turbo spin-echo sequences (TR range/TE, 240-249/10; number of excitations, 2-4; field of view, 350 mm; matrix, 205 x 256; slice thickness, 10 mm; slice gap, 1 mm; 3 slices per breath-hold of 13-20 sec; 16-24 slices; acquisition time, 104-107 sec) were performed. The field of view of the sequences encompassed the complete small bowel.
During gadodiamide injection, an axial dynamic T1-weighted turbo spin-echo sequence (TR/TE, 240/10; number of excitations, 2; field of view, 270 mm; matrix size, 128-140 x 256; slice thickness, 10 mm; minimal slice gap, 1 mm per time point of 4-6 sec; 3 slices acquired at 3 levels [1 slice at each of the 3 levels]; scanning time, at least 2 min with a maximum of 3 min; 20-45 time points) was performed at three levels of thickened bowel wall in free breathing. When only a short segment of bowel was thickened, the three slices were targeted at this segment. When a larger segment or multiple segments were involved, the three dynamic slices were planned to include up to three bowel segments with the thickest bowel wall. The planning was performed by an abdominal radiologist who was not informed of the clinical findings.
After the dynamic sequence, axial and coronal T1-weighted turbo spin-echo sequences were performed identically to the sequences before gadodiamide injection.
Data Postprocessing
The images were evaluated by an experienced abdominal radiologist (3 years
of pertinent clinical experience as an abdominal radiologist) who was blinded
to the clinical findings and the findings of other imaging or endoscopic
examinations. This radiologist analyzed all sequences for the presence of
thickened small-bowel wall and enhancing small-bowel segments.
"Thickened small-bowel wall" was defined as thickening of the
bowel wall of more than 3 mm. Bowel wall signal intensity measurements were
performed at the most prominent lesion using the static T1-weighted sequences
(axial or coronal) and the dynamic sequence. The measurement encompassed
placing a region of interest at the thickened small-bowel wall with the
fastest and most intense enhancement. The region of interest was changed
(range, 3-48 mm2) according to the thickness and length of the
abnormal bowel wall.
The signal intensities of the dynamic scan were measured and displayed in a
graph. The graphs showed a typical enhancement pattern
(Fig. 1), starting with
baseline intensity (SIbase), increasing with the passing of the
bolus, and ending with an elevated and stable but slowly decreasing intensity
(SIend). In some cases, a short intermediate higher maximum
intensity could be observed before the stable elevated intensity was reached;
this intermediate maximum was not used in the evaluations. The time point of
the start of the administration of the contrast agent (tinject =
t0), the time point of the start of the enhancement
(tstart), and the time point of the maximum enhancement
(tend) were assessed. The time interval tend -
tstart was called 
t.
|
 |
 |
To calculate the ERstat for the static series—that is, the
unenhanced and contrast-enhanced T1-weighted turbo spin-echo
sequences—one has to correct for differences in amplifier gain and image
scaling between the two series. Therefore, a reference region of interest
containing mesenteric fatty tissue (that should not enhance) was used, and
ERstat was calculated as follows:
 |
The SoE, ERdyn, and ERstat were calculated by a research fellow who was not involved in evaluating the imaging sequences or the reference standard and who was not aware of any of the other findings.
Further parameters included in the analysis were tstart,
t, and the bowel wall thickness. The bowel wall thickness of the most
thickened lesion was measured on both the dynamic and the static series. The
largest of these two measurements was recorded.
Reference Indexes
To determine the clinical value of the MRI features, three parameters were
used as reference index: clinical grade, Crohn's disease activity index (CDAI)
[17], and Van Hees activity
index [18]. The first
reference index is a clinician's subjective opinion; the latter two are
validated indexes.
Clinical grade was rated on a 4-point Likert scale (1, remission or inactive; 2, mild; 3, moderate; and 4, severe disease activity) by one gastroenterologist per hospital who was not aware of the MRI findings. The scale was subjective, based on patient symptoms, physical findings, activity indexes, surgery, and all examinations that the patients underwent (endoscopy, laboratory, and other imaging studies) within 2 weeks before or after the MRI.
CDAI and Van Hees activity index use different parameters for the assessment of the activity of Crohn's disease, as is shown in Appendix 1 (CDAI) and Appendix 2 (Van Hees activity index). Although no upper and lower limits exist, the CDAI ranges approximately from 0 to 600 and the Van Hees activity index from 30 to 400, with higher scores indicating more severe illness. Cutoff values for CDAI are less than 150, inactive disease; 150-450, inflammatory activity; and greater than 450, very severe inflammation. Cutoff values for the Van Hees activity index are less than 100, inactive disease; 100-150, slight inflammatory activity; 150-210, moderate inflammatory activity; and greater than 210, severe to very severe inflammatory activity. When the required laboratory tests (hematocrit, serum albumin, and erythrocyte sedimentation rate) had not been performed within 2 weeks before the MRI examination, blood was drawn at the examination from the IV contrast administration line before scanning. Questionnaires for CDAI and Van Hees activity index were given by a gastroenterologist (one per hospital) just before the MRI for determination.
In all 48 patients, the clinical indexes—clinical grade, CDAI, and Van Hees activity index—were determined (Table 1). Clinical grade was based on a variety of examinations performed in these patients, mostly barium enteroclysis (16 patients), colonoscopy (15 patients), or both barium enteroclysis and colonoscopy (10 patients). Besides these two techniques, surgery was performed in 15 patients (all with a histologic diagnosis), scintigraphy in 14, sonography in 11, and CT in one patient. The patient groups of the two hospitals were similar with respect to age, clinical grade, and Van Hees activity index, but not with respect to CDAI. The average CDAI for hospital 1 was 276 (± 80) and for hospital 2 was 140 (± 53), whereas the average Van Hees activity index was 139 (± 56) for hospital 1 and 121 (± 42) for hospital 2. We suspect that the subjective elements in the CDAI questionnaire were interpreted differently in the two hospitals.
|
Statistical Analysis
The averages of the SoE, ER, and time intervals were calculated per
hospital for patients without and those with disease activity on the basis of
the clinical grade. Because the number of patients with inactive disease was
small (n = 3 for hospital 1, n = 2 for hospital 2), no
statistical analysis was performed on these data. The measurement parameters
(SoE, ER, time intervals, and bowel wall thickness) were correlated with the
reference indexes (clinical grade, CDAI, and Van Hees activity index) by
calculating Spearman's correlation coefficient with one-tailed p
values.
Systematic differences in reference indexes (e.g., CDAI scores) or in MRI
parameters between hospitals (or scanners) could introduce an artificial
correlation if data from different hospitals are analyzed in a single step,
but such a shift does not affect correlations calculated per hospital.
Therefore, we first calculated standard correlation coefficients per hospital.
In the next step, correlation coefficients from both hospitals were pooled
using the standard approach for pooling correlations
[19]. This approach uses the
Fisher's z-transformation to obtain more normally distributed values of the
correlation coefficients based on the following formula:
 |
 |
The pooled estimate is the weighted average of the transformed estimates weighted by the inverse of their variance (1 / SE2), the standard approach in meta-analysis. The pooled correlation coefficient was then transformed back to the original scale. Correlation coefficient values were interpreted as follows: 0.0, no association; 0.2, weakly correlated; 0.5, moderately correlated; 0.8, strongly correlated; and 1.0, perfectly correlated [20]. Values for p of less than 0.05 were considered significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
t, and in 38 patients the
ERdyn, small-bowel wall thickness, and tstart, could be
evaluated. The main reason that prevented measurements in the dynamic series
(n = 10) and in the static series (n =7) was the absence of
a thickened small-bowel wall or bowel wall that was too thin for accurate
measurements. In one patient, a measurement error prevented the calculation of
the SoE and t. The absence of representative homogeneous adipose tissue
on the static series prevented the calculation of ERstat in one
patient. The average values of the SoE, ERdyn, and
ERstat per hospital for patients without and those with disease
activity are shown in Table
2.
|
|
|
|
|
|
|
The fact that the average ER for fat after administration of gadodiamide in all patients was 1.01 (± 0.02; range, -0.01 to 0.06; n = 20) showed that adipose tissue did not enhance on the dynamic series and was rightfully used as a reference region in the static measurements.
On the basis of the dynamic and static series, in, respectively, 10 and seven patients, bowel wall was not clearly identified or was too thin for accurate measurement, and wall thickness and enhancement measurements could not be determined. In five of these patients, the clinical grade was scored as inactive disease (Table 3). In one patient, however, the clinical grade was scored as severe inflammation. This score was mainly based on surgery, endoscopy, and scintigraphy that showed inflammation of the descending colon, sigmoid, and rectum and a rectovaginal fistula, but normal small bowel.
|
Of the 41 patients in whom the bowel wall was measured
(Table 3), 37 showed thickened
bowel wall on MRI (mean, 6.4 mm; range, 4-13 mm). Of the four patients without
bowel wall thickening (wall thickness 
3 mm), one patient had inactive,
one patient had mild, and two patients had moderate disease on the basis of
clinical grade. In the two patients in whom clinical grade was scored as
moderate, the discrepancy was partly caused by a fistula with involvement of
the surrounding tissue found at sonography in one patient, and by inflammation
of the terminal ileum found at surgery (serositis, fatty overgrowth) in the
other patient. In five patients with a clinical grade of inactive disease,
bowel wall thickness, SoE, and ER could be determined.
The average values for tstart for patients with active disease
and those with inactive disease based on the clinical grade were,
respectively, 27 (± 20) and 21 (± 16) sec, and the average
values for t were 39 (± 19) and 34 (± 14) sec.
Correlations
The results of the correlation between the measurement parameters (SoE,
ERdyn, ERstat, bowel wall thickness, and time
measurements) and reference indexes are shown in
Table 4. Between any two of the
reference indexes, significant correlation was seen (clinical grade with CDAI:
r =0.44, p = 0.002; clinical grade correlated with Van Hees
activity index: r = 0.63, p < 0.001; CDAI with Van Hees
activity index: r = 0.56, p < 0.001). No significant
correlation was seen between the SoE and any of the index parameters. The
ERdyn correlated significantly with the CDAI (r =0.38,
p = 0.016) but not with clinical grade and Van Hees activity index.
Significant correlation was seen between the ERstat and all three
reference indexes: clinical grade (r = 0.29, p = 0.045),
CDAI (r = 0.31, p = 0.033), and Van Hees activity index
(r =0.36, p = 0.016). Bowel wall thickness (in mm)
correlated significantly with clinical grade (r =0.47, p =
0.003) and Van Hees activity index (r =0.41, p = 0.007). No
correlation was seen between any of the time measurements and any of the
reference indexes.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The best correlation was seen between bowel wall thickness and clinical grade and between bowel wall thickness and Van Hees activity index, with correlation coefficients of, respectively, 0.47 (p = 0.003) and 0.41 (p = 0.007). Although significant, these correlation coefficients were classified as weak to moderate and can only partly predict the disease activity. Particularly for the ERstat, its clinical use is probably limited, because the correlation coefficients were 0.29-0.36.
Several limitations must be considered: the imperfect reference standard, the limited value of signal intensity measurements on MRI, the use of two different scanners, and the amount of fluid taken for bowel distention. These will be discussed in the next paragraphs.
Reference Indexes
The main problem in this study, as in previous studies on Crohn's disease,
is that no ideal reference standard exists. Clinical grade was determined by
the clinician; it was based on his opinion as to the disease activity in the
patient. Because the decision of whether— and, if so, how—to
change treatment is made by a clinician, this index should be clinically
relevant. In general, however, the presence of stenoses, abscesses, and
fistulas will lead to a higher clinical grade of disease activity that does
not reflect the actual disease activity in the small-bowel wall, and that we
expected to diminish the usefulness of this index in this study. Therefore, we
included more objective indexes that reflect disease activity. The CDAI is the
most commonly used index in clinical trials and is currently the gold standard
for evaluation of disease activity
[21]. The Van Hees activity
index has been prospectively validated but correlated poorly with the CDAI in
other studies [21].
In our study we observed a significant difference in CDAI values between the two hospitals (Table 1), whereas this was not the case for the clinical grade and the Van Hees activity index. As an explanation, we hypothesize that subjective questions in the CDAI questionnaire were interpreted differently in the two hospitals because it is unlikely that the patient populations of the two university hospitals differ. Both hospitals are tertiary referral centers in the same area of the country with comparable patient populations regarding Crohn's disease. Such a systematic difference due to interpretation differences could introduce a spurious correlation. To prevent confounding, we performed a stratified analysis by calculating the correlation coefficient per hospital. These two coefficients were then pooled using a fixed effect approach after Fisher's z-transformation.
Ileocolonoscopy could be a good reference index to assess the degree of inflammatory activity of the large bowel wall and terminal ileum. However, because only patients with severe inflammation undergo ileocolonoscopy and because of the frequent impossibility of introducing the scope in the terminal ileum and more proximal parts of the small bowel, the use of this technique as a reference index in this study is limited.
MRI Measurements
Evaluation of enhancement was feasible in only 40 of 48 patients based on
the static series and in 38 of 48 patients based on the dynamic series. In
most of the patients (n = 5) in whom bowel wall measurements were not
possible, the clinical grade was scored as inactive disease. The absence of
thickened bowel wall or the absence of bowel wall enhancement renders
identification of the bowel wall more difficult, thus limiting the
determination of bowel wall thickness, SoE, and ER. Despite this seemingly
inherent problem of the technique, we were able to determine the SoE and ER in
five patients with the clinical grade of inactive disease.
In contrast to values from CT images, which are given in Hounsfield units and represent exact reproducible values, values from MR images depend on a large number of variables and have no absolute meaning. Therefore, one must take care to use the same sequences and sequence parameters for different patients, and to use ratios of values from within the same sequence to correct for variations in amplifier values. When results from MR images with different field strengths are to be combined, field dependency of T1 of tissues—and gadolinium-based contrast-enhanced images are T1-weighted—should be taken into account. The two-step approach (first calculating separate correlation coefficients and then pooling) has avoided bias that might be caused by simple pooling of the MRI parameters in the same way as for the CDAI values.
Other Studies
For accurate measurements, good delineation of the bowel wall is necessary.
It can be expected that lesions will be depicted better in adequately
distended bowel, which leads to more accurate measurements. In our study, 800
mL of oral contrast medium was administered because this volume was deemed
sufficient for identification of thickened bowel wall. In the literature,
volumes up to 1.5 L [6,
11] of oral contrast medium
have been used for distention. Although those studies show good results, the
study by Shoenut et al. [7], in
which no oral contrast agent was used, did not show inferior results. Future
studies should weigh the diagnostic advantage versus the patient burden of
using large volumes of oral contrast agent.
Most of the limitations mentioned do not apply only to this study but also in general hamper evaluation of disease activity using MRI in patients with Crohn's disease. This is probably the reason that in many studies the severity of Crohn's disease has been based on the presence and extent of abnormalities typical for Crohn's disease [5, 6] (e.g., length of wall thickening, presence of fistulas, abscesses, and stenoses) rather than a parameter reflecting local severity (e.g., ER).
The degree of bowel wall enhancement is probably the parameter that is most closely related to the degree of inflammation. To quantify the actual degree of inflammation, either a subjective categoric scale [12-14, 22] or an ER [7-11] is used. Most of the studies that use categoric scales show that inflamed bowel tissue shows more enhancement than normal bowel wall and that there is a correlation between the degree of enhancement and disease activity. Studies that try to quantify the degree of enhancement by calculation of an ER (or enhancement increase), however, show varying results ranging from good correlation to no correlation with the CDAI. The study by Pauls et al. [11] showed significant correlation between the maximum contrast uptake and the CDAI (r = 0.591, p = 0.033), but the number of patients was small (n = 13). A larger study (n = 82) by Schunk et al. [9] showed a correlation coefficient between the increase in bowel wall enhancement and the CDAI (r = 0.25, p = 0.02) that was comparable to that in our study.
Wall thickness is correlated with disease activity; however, previous inflammation might have caused fibrosis and wall thickening in a noninflamed segment. Nevertheless, in our study, bowel wall thickness, irrespective of enhancement, showed a significant correlation with clinical grade and Van Hees activity index (r = 0.47 and r = 0.41, respectively) and showed the strongest correlation coefficient of all measurement parameters.
A higher ER, as is seen in more severely inflamed tissue, can be
hypothesized to be related to a steeper enhancement curve (higher SoE), a
longer time interval of enhancement (t), or a combination of both.
Because no significant correlation was seen in this study between reference
parameters and steepness of the curve (SoE) or the time interval (t),
the higher ER is not significantly related to any of these two, meaning that
in one patient it could be caused by a longer time interval and in another
patient by a steeper enhancement curve. This fact makes it difficult to
provide a simple explanation for the higher ER values in more severely
inflamed bowel wall. Probably the higher ER values are caused by a combination
of factors such as increased blood flow (due to vasodilatation,
neovascularization, or increased cardiac output) and increased vascular
permeability. Because of these assumed multifactorial causes, correlations
will be more difficult to prove.
Conclusion
In conclusion, the results of this study are in line with those in the
current literature, in that the ER and bowel wall thickening correlate weakly
to moderately with indicators of the severity of Crohn's disease. However,
this study also shows that the SoE does not seem to correlate in patients with
a suspicion of exacerbation of Crohn's disease. Therefore, ERs based on static
imaging and bowel wall thickness measurements are the best parameters on which
to base the severity of disease activity in Crohn's disease. Dynamic
measurements do not add any information.
APPENDIX 1: Assessment of Activity of Crohn's Disease Using Crohn's Disease Activity
Index (CDAI)
[17]
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
APPENDIX 2: Van Hees Activity Index
[18]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
a 1 = None, 2 = dubious, 3 = diameter < 6 cm, 4 = diameter 6–12 cm, 5 = diameter > 12 cm
b 1 = Well formed; 2 = soft, variable; 3 = watery
Acknowledgments
We thank Hans Reitsma for his valuable help with the statistical
analysis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  T. A. Jaffe, A. M. Gaca, S. Delaney, T. T. Yoshizumi, G. Toncheva, G. Nguyen, and D. P. Frush Radiation Doses from Small-Bowel Follow-Through and Abdominopelvic MDCT in Crohn's Disease Am. J. Roentgenol., November 1, 2007; 189(5): 1015 - 1022. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
