Pediatr Radiol (2010) 40:1615–1624
DOI 10.1007/s00247-010-1781-9
ORIGINAL ARTICLE
Feasibility of 3-T MRI for the evaluation of Crohn disease
in children
Charuta Dagia & Michael Ditchfield & Michael Kean &
Anthony Catto-Smith
Received: 11 May 2009 / Revised: 7 December 2009 / Accepted: 8 January 2010 / Published online: 6 August 2010
# Springer-Verlag 2010
Abstract
Background Crohn disease (CD) is a chronic inflammatory
bowel disease that can affect any part of the gastrointestinal
tract from the oral cavity to the anal canal. It occurs in all
ages and is a significant cause for morbidity in children.
Interest in MRI evaluation of CD has increased because of
the concern regarding cumulative radiation dose from
contrast fluoroscopic studies and CT. Several reports have
demonstrated MRI to be a useful technique for CD. Most of
these studies were performed at 1.5-T field strength.
Imaging at a higher field strength, with a greater signal-
to-noise ratio, has the potential of reducing scan times and
increasing the resolution. However, there is a concurrent
increase in artefacts, and these can be pronounced with
abdominal imaging at 3 T.
Objective To determine the feasibility of 3-T MRI for CD
in children and to assess the value of different sequences
and the effect of artefacts that could potentially limit the
role of bowel MR imaging at higher field strengths.
Materials and methods A retrospective study of 46 children
with biopsy-proven CD (ages 8–19 years, 53% boys) was
performed. Sixty-eight consecutive MRI studies were
performed on a 3-T scanner between 2005 and 2007; 42
of the abdomen (62%) and 26 of the pelvis/perineum
(38%). Sorbitol was administered for the abdominal
C. Dagia : M. Ditchfield (*) : M. Kean
Department of Medical Imaging and Murdoch Childrens Research
Institute, The Royal Children’s Hospital,
Flemington Road, Parkville,
Victoria 3052, Australia
e-mail: Michael.Ditchfield@southernhealth.org.au
A. Catto-Smith
Department of Gastroenterology, The Royal Children’s Hospital,
Victoria, Australia
studies; orally for 36/42 (86%) studies and via a naso-
jejunal (NJ) tube for 6/42 (14%) studies. For the abdomen,
T2-W half-fourier acquisition single-shot turbo spin-echo
(T2-W HASTE), true steady-state free precession (true
FISP), pre-contrast and contrast-enhanced (CE) T1-volume
interpolated gradient-echo (T1-W VIBE) and CE T1-W fast
low-angle shot (T1-W FLASH) sequences were performed.
For the perianal and pelvic assessment, fat-saturated T2-W
turbo spin-echo (TSE), pre-contrast and CE T1-W FLASH
or VIBE sequences were performed. The sequences were
scored for diagnostic quality by two paediatric radiologists
for visualisation of the bowel wall, whether normal or
pathological and the visualization of extra intestinal
manifestations. The effects of distension, susceptibility
artefact and motion were assessed.
Results Six (14%) abdominal MRI studies were normal.
Thirty-six (86%) were abnormal with good correlation with
endoscopic findings. The pelvic and perianal MRI studies
were all abnormal (26/26, 100%) with good correlation
with proctoscopy and examination under anaesthesia. All
the sequences had high average scores (greater than or close
to 3), except true FISP with a score of 2.4. The score was
greatest in those who had NJ administration of sorbitol;
however, satisfactory distension was also possible with oral
administration of contrast. True FISP was the sequence
most affected by a combination of suboptimal distension
and artefact from colonic contents. With adequate disten-
sion, true FISP image quality improved remarkably. The
overall score of this sequence was satisfactory in the
absence of susceptibility and movement artefact.
Conclusion With appropriate attention to technique, with
optimal distension and control of movement, high-quality,
3-T assessment of the abdomen, pelvis and perineum is
possible. All sequences used at 1.5 T can be used at 3 T,
however true FISP was the most prone to artefact.
1616
Keywords Crohn disease . 3 T . MRI . Artefact . Child
Introduction
Crohn disease (CD) is a chronic inflammatory bowel
disease that can affect any part of the gastrointestinal tract,
from the oral cavity to the anal canal. It occurs in all ages
and is a significant cause of morbidity in children [1–3].
The disease is characterised by segmental bowel wall
inflammation with complications such as abscesses and
sinus and fistulous tracts [3, 4].
Endoscopy and biopsy are important in making the initial
diagnosis of CD. Imaging is used to evaluate the small bowel
at presentation, to detect the extent of small and large bowel
involvement, to detect extra intestinal manifestations or
complications and to monitor response to treatment [5, 6].
Interest in MRI evaluation of CD has increased because of
the concern regarding cumulative radiation dose from
contrast fluoroscopic studies and CT. In addition, MRI has
better soft-tissue contrast than these examinations. Several
reports have demonstrated MRI to be a useful technique for
CD [4, 6–9]. Most of these studies were performed at 1.5-T
field strength. Imaging at higher field strength, with a greater
signal-to-noise ratio, has the potential of reducing scan times
and increasing the resolution [10–12]. However, there is a
concurrent increase in artefacts, and these can be pronounced
with abdominal imaging at 3 T [12].
An earlier study compared the MRI findings on different
sequences with other modalities (barium studies, CT and
US) in a smaller patient group [3]. This study was
performed to determine the feasibility of 3-T MRI for CD
in a larger group of children and to more comprehensively
assess the value of different sequences and the effect of
artefacts that could potentially limit the role of bowel MR
imaging at higher field strengths.
Materials and methods
After obtaining approval from the institution’s research and
ethics board, 46 children with biopsy-proven CD (age range
8-19 years [mean 13.5], 53% male) were identified on the
radiology information system of a tertiary paediatric
hospital. Sixty-eight consecutive MRI studies, consisting
of either abdominal (42/68, 62%) or pelvic and perianal
(26/68, 38%) scans, performed over a 2-year period
between 2005 and 2007 were reviewed. The imaging was
performed on a 3-T scanner (Siemens TIM TRIO, Erlangen,
Germany) with a phased-array body coil.
The studies were performed without sedation or general
anaesthesia. Children <10 years old were offered a practice
session on a mock MRI unit prior to the examination.
Pediatr Radiol (2010) 40:1615–1624
The abdominal MRI studies were performed after 6 h
fasting and no bowel preparation. Sorbitol (Sorbilax oral
solution [70%], Pfizer) was the enteral contrast agent. The
solution was prepared by adding 15 ml of sorbitol to each
500 ml of water, with 1–1.5 l administered orally for 36/42
(86%) studies. Sorbitol was administered via a naso-jejunal
(NJ) tube for 6/42 (14%) studies, when the child was
unable to tolerate oral contrast due to vomiting or general
illness.
In those given sorbitol orally, ingestion commenced
60 min prior to the scan, at a rate of 250–300 ml/15 min. If
distension of the proximal small bowel was considered
inadequate on the initial acquisition, another 250–300 ml of
sorbitol was administered. The adequacy of distension was
assessed by measuring the luminal diameter for older
children, with 3 cm considered good distension, and by
visually comparing the luminal diameter of each segment
with that of the best distended portion of bowel in children
<50 kg. Intravenous metoclopromide (Maxalon, Pfizer, NY,
USA; 0.15 mg/kg) was given in the event of nausea. In
those given sorbitol by NJ tube, the intubation was
performed under fluoroscopic guidance (fluoroscopy time,
1-3 min). The sorbitol solution was instilled manually on
the MRI table, using 50 ml syringes. If bowel loops were
collapsed on the initial sequence, good distension was
ensured by instilling more sorbitol. Intravenous buscopan
(hyoscine butyl-bromide, Boehringer Ingelheim, Germany;
0.5 mg/kg, up to 20 mg) was administered immediately
before the start of image acquisition for all abdominal
studies. Apart from the complaint of minor diarrhoea (8/42,
19% studies), no other side effect was encountered.
The children were placed supine on the MRI scan table
as this position was better tolerated, especially by the
younger children, and enabled distraction with audiovisual
material. A dedicated 12-channel body array coil was used
and all sequences, unless stated otherwise in Table 1, were
obtained with breath-hold.
For the abdomen, T2-W half-fourier acquisition single-
shot turbo-spin-echo (HASTE) imaging was performed at
two echo times (TE). The long TE images (TR 1470 ms,
TE 248 ms, flip angle [FA] 109°, PAT factor 2, breath-hold)
were obtained with and without fat suppression. The short
TE images (TR 1300 ms, TE 88 ms, FA 109°, PAT factor 2,
respiratory control off) were all obtained without fat
suppression. True steady-state free precession (FISP)
images were obtained in the coronal plane (TR 3.6 ms,
TE 1.55 ms, FA 31°, PAT factor 3, respiratory control off).
Two types of T1 sequences were used: T1-volume
interpolated gradient-echo (VIBE) (TR 6.03 ms, TE
2.71 ms or 1.46 ms for delayed, FA 11°, PAT factor 3,
breath-hold) and T1 fast low-angle shot (FLASH) (TR
181 ms, TE 2.46 ms, FA 90°, PAT factor 2, TA 1.05 min,
breath-hold). A pre-contrast coronal T1-W VIBE sequence
Pediatr Radiol (2010) 40:1615–1624 1617

Table 1 Imaging sequence parameters. WE* water excitation; post-contrast VIBE parameters given out of brackets are for early acquisition,
within brackets are for the delayed phase; ** Differed depending on plane of acquisition

T2-W True T2-W T2-W T1-W Pre-contrast T1 WE* VIBE T1-W FLASH T1-W VIBE
HASTE FISP HASTE TSE VIBE (abdo and (abdo and pelvis) (abdo and FS (pelvis)
(abdo) (abdo) (abdo) (pelvis) pelvis) pelvis)
Short TE Long TE

Plane Coronal Coronal Coronal 3 planes Coronal Coronal Axial Axial

TR (ms) 1300 3.6 1470 3200 6.03 6.03 181 3.36
TE (ms) 88 1.55 248 250 1.23 2.71 (1.46 for 2.46 1.4
delayed)
TA 0:13 0:20 1:28 3:31 0:21 0:21 1:05 0:19
FA 109° 31° 109° 11° 11° 90° 13°
PAT factor 2 3 2 2 3 3 2 2
Voxel size 1.8×1.3× 1.2× 1.8×1.3× 1.4×1.1× 1.5×1.4×2.5 1.5×1.4×2.0 1.2×0.9×3.5 1.5×1.2×
(mm) 6.0 1.1× 2.0 2.5 2.5
1.3
SNR 1:00 1:00 1:00 1:00 1:00 1:00 1:00 1:00
Slice 6 1.3 2 2.5 2.5 2.0 3.5 2.5
thickness
(mm)
Distance 75% 20% 10% 5% 20% 20% 5% 20%
Factor
Averages 1 1 1 1 1 1 1 1
Fat sat Strong None Weak and Strong Fat Sat WE, Normal Strong Fat Sat
Mode none (Fast for
delayed)
Respiratory Off Off Breath- Breath- Breath-hold Breath-hold Breath-hold Breath-hold
Control hold hold

was obtained. A gadolinium-based agent at a dose of
0.2 mmol/kg (gadopentetate dimeglumine [Magnevist,
Bayer Schering Pharma, Levekusen, Germany] or gadoter-
idol [Prohance, Bracco Diagnostics, N.J. USA]) was
intravenously administered and the contrast-enhanced
(CE) T1-W VIBE coronal sequence acquired up to three
times; immediately after contrast administration, at 1- to 2-
min and then 5-min intervals. A CE T1 FLASH axial
sequence was also obtained.
The MRI studies for perianal and pelvic assessment were
performed without fasting, bowel preparation, oral contrast
or buscopan.
Imaging of the perianal region was obtained with fat-
saturated T2-W TSE (TR 3200 ms, TE 250 ms, FA 103°,
PAT factor 2, breath-hold) in all three planes. This was
followed by pre-contrast coronal T1-W VIBE and CE T1-
W FLASH or VIBE images in the axial and coronal or
sagittal planes.

Table 2 The average scores of

different sequences performed Sequence Number of studies with each sequence Score (0–4)
for abdominal and perineal
evaluation Abdomen
T2-W HASTE short TE (coronal) 42/42 (100%) 3.0
T2-W HASTE long TE (+/- fat-sat) (coronal) 42/42 (100%) 2.9
True FISP (coronal) 24/42 (57%) 2.4
CE T1-W VIBE (coronal) 39/42 (93%) 3.0
CE T1-W FLASH (axial) 24/42 (57%) 3.3
Perineum and pelvis
T2-W FSE (axial and coronal) 26/26 (100%) 3.6
T2-W FSE (sagittal) 26/26 (100%) 3.0
CE T1-W VIBE (axial and coronal) 14/26 (54%) 3.8
CE T1-W FLASH (axial and sagittal) 16/26 (62%) 3.5
1618 Pediatr Radiol (2010) 40:1615–1624

Table 3 A comparison of
abdominal sequence scores Sequence Oral NJ
obtained with and without NJ
tube n Score n Score

T2-W HASTE short TE 36/36 (100%) 2.8 6/6 (100%) 3.9

T2-W HASTE Long TE 36/36 (100%) 2.7 6/6 (100%) 3.9
True FISP 21/36 (58%) 2.2 3/6 (50%) 3.8
CE T1-W VIBE 33/36 (92%) 2.9 6/6 (100%) 3.8
CE T1-W FLASH 18/36 (50%) 3.1 6/6 (100%) 4

The studies were retrospectively reviewed by two
paediatric radiologists blinded to the findings on previous
imaging and endoscopy.
Each sequence was assessed for visualisation of the bowel
wall, whether normal or pathological, and the visualization of
extra intestinal manifestations. For the abdomen the images
were scored by dividing the bowel into three segments:
jejunum and proximal ileum, terminal ileum and caecum and
colon. The bowel wall was evaluated for the ability to
visualise the wall and if diseased, the extent of involvement,
wall thickness (normal <4 mm with adequate distension),
degree of enhancement relative to surrounding equivalent,
well-distended loops on visual assessment and complications
such as abscesses and sinus and fistulous tracts [9]. Extra-
intestinal (mesenteric) manifestations of fibro-fatty pro-
liferation, mesenteric oedema, hypervascularity and
lymphadenopathy were also recorded. Each of the three
regions and the extra-intestinal manifestations were
assigned a score of 1 (0–0.25, not possible to visualise;
0.26–0.5, poorly visualized; 0.51–0.75, reasonable visu-
alization; 0.76–1, excellent visualization). The abdominal
sequences were therefore assigned a score between 0 and 4.
For assessment, each of the three bowel segments was
divided into four subsegments and individually scored by
each radiologist. If the scores differed by one grade, these
were averaged to reach the final score for that subsegment.
A difference of more than one grade was resolved by
consensus. The scores of the four subsegments thus
obtained were added to reach the final score of that
particular segment.
The children underwent endoscopy and biopsy at initial
presentation and surgery was performed in 8/68 (12%) to treat
abnormalities identified on MRI—these were correlated with
the MRI findings (mean time difference between MRI and
biopsy 17.4 days, minimum 0 days, maximum 66 days).
Table 4 The effect of disten-
sion on different abdominal Sequence NJ
sequences obtained after oral
contrast administration
For the pelvic and perineal examinations, visualisation of
the rectosigmoid and anal canal were assessed in a similar
manner to the bowel in the abdominal studies as were the
extra-intestinal manifestations, visualisation of collections,
sinus and fistulous tracts and were scored and assigned a
score between 0 and 4.
These findings were correlated with prior proctoscopy or
examination under anaesthesia.
The effect of distension was assessed by comparing
those who had sorbitol orally with those who had sorbitol
through NJ tubes. In addition, the studies were divided into
those with and without adequate distension and the two
groups compared. The effect of susceptibility artefact from
colonic contents such as bowel gas and faecal matter was
evaluated by comparing those with significant artefact to
those without. The effect of bowel and respiratory motion
artefact was evaluated by comparing those with significant
artefact to those without.
Results
Six (6/42, 14%) abdominal MRI studies were normal; all of
these had mild endoscopic disease. Thirty-six (36/42, 86%)
were abnormal with good correlation of endoscopic findings
in the colon, caecum and terminal ileum with regard to
disease extent, type of involvement (oedema or fibrotic
change) and complications. All of those who had surgery
(8/8, 100%) had abnormalities on MRI, with good correla-
tion of the MRI findings.
The pelvic and perianal MRI studies were all abnormal
(26/26, 100%) with good correlation with proctoscopy
and examination under anaesthesia, although these did
not give significant detail of the presence and nature of
tracts.
Oral
Overall score Good distension Inadequate distension

T2-W HASTE 3.9 2.8 3.4 2.7

True FISP 3.8 2.2 3.4 2.4
CE T1-W 3.9 2.9 3.4 2.6
Pediatr Radiol (2010) 40:1615–1624 1619

Fig. 1 True FISP images.

(a) Poor bowel distension
considerably limits image
quality. (b) with good bowel
distension, diagnostic quality is
much improved

The mean scores of the different abdominal and pelvic/
perineal sequences are summarised in Table 2. All the
sequences had high average scores (at or close to 3), except
true FISP with a score of 2.4.
Distension
The effect of distension was assessed by comparing studies
performed after oral administration of sorbitol with studies
in which sorbitol was administered via NJ tube (Table 3).
Of the studies obtained after oral contrast, image quality of
studies with good distension was compared with image
quality of studies demonstrating inadequate bowel disten-
sion (Table 4). The score was greatest in those patients who
had NJ administration of sorbitol. However, with satisfac-
tory distension, oral contrast could also achieve good
results (Fig. 1) [9].
on the longer TE images (Fig. 5). Mild or early bowel wall
and mesenteric oedema and increased signal in lymph nodes
were easier to identify on the fat-suppressed T2 images,
though they were also visible without fat suppression. Vasa
recta changes were well seen with and without fat
suppression, but these, and the reactive mesenteric lymph-
adenopathy, were best assessed on the CE sequence (Table 4).
In this study, true FISP was most affected by a
combination of suboptimal distension and artefacts from
colonic contents. With adequate distension, true FISP
image quality improved remarkably (Fig. 6). Overall scores
for this sequence also improved in the absence of
susceptibility and movement artefact.

Susceptibility artefact

All sequences were affected by susceptibility artefact, but

the effect was more pronounced on the gradient-echo
images—true FISP, CE T1-W VIBE (Fig. 2) and FLASH.
The T2-W HASTE sequence was less affected (Table 5).

Bowel and respiratory movement artefact

Movement artefact was also an important cause of image-

quality degradation (Fig. 3). T2-W HASTE was the least
affected; true FISP was most affected, followed by CE T1-
W VIBE and FLASH (Table 6).

T2-W sequences

T2-W HASTE images acquired at shorter TE without fat

suppression provided good morphological assessment of the
bowel (Fig. 4) as well as mesentery (Table 4). Bowel wall Fig. 2 This CE T1-W VIBE sequence demonstrates susceptibility
oedema was more apparent with increased T2 weighting, i.e. artefact from colonic air obscuring the bowel wall
1620 Pediatr Radiol (2010) 40:1615–1624

Table 5 The effect of susceptibility artefact on image quality of different sequences

Sequence Overall Not limited by susceptibility artefact Limited by susceptibility artefact

n Score n Score n Score

True FISP 24 2.4 13/24 (54%) 2.8 11/24 (46%) 1.9

T2-W HASTE long TE 42 2.9 35/42 (83%) 3.0 7/42 (17%) 2.6
T2 HASTE short TE 42 3.0 34/42 (81%) 3.1 8/42 (19%) 2.6
CE T1-W FLASH 24 3.2 14/24 (58%) 3.8 10/24 (42%) 2.2
CE T1-W VIBE 39 2.9 19/39 (49%) 3.7 20/39 (51%) 2.3

CE studies demonstrated uniform transmural enhancement on the

Contrast-enhancement patterns were identified as striated,
superficial or transmural. Some segments with a superficial
pattern of enhancement observed on the early acquisitions
(within 2 min of intravenous contrast administration),
delayed images obtained 3–5 min after contrast adminis-
tration (Fig. 7). No change in enhancement pattern was
evident on further delayed acquisitions (10 min after
contrast). CE images were very useful in the assessment
of the pelvis and perineum (Fig. 8).

Discussion

The relapsing-remitting course of CD necessitates serial

Fig. 3 T2-W HASTE sequence. (a) movement artefact considerably
limits assessment. (b) reduced peristalsis after administering hyoscine
butyl-bromide with marked improvement in image quality
assessment. With intrinsic high soft-tissue contrast, multi-
planar imaging capability, lack of ionising radiation and the
ability to simultaneously assess the small bowel, colon and
extra-intestinal manifestations, MRI is potentially an ideal
technique for imaging children with IBD. However,
paediatric MRI poses unique challenges in view of patient
size and cooperation for obtaining optimal images.
MR imaging at 3T has the advantage of increasing the
signal-to-noise ratio (SNR) approximately 1.7-1.8 times
compared to imaging at 1.5T. This can be useful for
improving the spatial or temporal resolution, or both. The
higher spatial resolution at 3T can potentially improve the
detection of early, superficial or subtle abnormalities. How-
ever, attention to technique is very important at 3T to reduce
artefacts that can be a greater problem at high field strengths.
Adequate distension is a prerequisite for good image
quality; it improves assessment of the bowel wall and
replaces the colonic gases which thereby reduces suscepti-
bility artefact. This study demonstrates that this can be
reliably obtained with NJ administration of sorbitol, but can
also be achieved less invasively and without fluoroscopy
with oral administration of an adequate volume of sorbitol
over an appropriate time period. In the authors’ opinion, the
oral route should be the preferred mode of sorbitol
administration; NJ tube may be reserved only for those
unable to drink the required volume of contrast due to
nausea, systemic illness or other such cause.
Three types of movement can affect MR image quality
in evaluating the bowel: peristalsis, respiratory and patient
movement. Motion artefact can be reduced by faster
imaging, the use of breath-hold sequences and buscopan
Pediatr Radiol (2010) 40:1615–1624 1621

Table 6 The effect of respirato-

ry and bowel movement on Sequence Overall Not limited by movement Limited by movement
image quality of different
abdominal sequences n Score n Score n Score

True FISP 24 2.4 20/24 (83%) 2.2 4/24 (17%) 1.8

T2-W HASTE long TE 42 2.9 33/42 (79%) 3.0 9/42 (21%) 2.6
T2-W HASTE short TE 42 3.0 33/42 (79%) 3.1 9/42 (21%) 2.6
CE T1-W FLASH 24 3.2 14/24 (58%) 3.8 10/24 (42%) 2.2
CE T1-W VIBE 39 2.9 16/39 (41%) 3.7 23/39 (59%) 2.3

[13, 14]. Performing these MRI studies at 3T enables the use
of parallel imaging techniques to reduce the scan time, while
maintaining adequate SNR [15]. Breath-hold techniques
require good patient cooperation and we have shown this
can be successfully performed in children over the age of
8 years. The level of cooperation can be improved by prior
explanation of what is required and a practice session that we
routinely offer to children under the age of 10 years.
Buscopan is short acting. Hence, the timing of its adminis-
tration just before the start of image acquisition is important
as with previous studies performed at 1.5T [16].
Abdominal MRI is prone to artefacts such as magnetic
susceptibility, peripheral image distortion and dielectric or
B1 inhomogeneity effects. These increase at higher field
strengths and, if not addressed, can markedly impede
diagnostic image quality when assessing the bowel at 3T.
Susceptibility artefact from air and faecal material was an
important factor degrading image quality [12]. All sequen-
ces were affected; the effect was more marked on the
gradient-echo images (true FISP, T1-W VIBE, T1-W
FLASH). To reduce this we now attempt to perform the
MRI study within 48 h of endoscopy and believe that when
this is not possible, bowel preparation should be performed.
The T2-W HASTE sequence performed well at 3T. As a
single-shot technique it is much less sensitive to motion
artefacts, with considerably less degradation by peristalsis
and other forms of movement when compared to the
gradient-echo sequences. These findings have been
reported with bowel imaging at 1.5T and hold true at 3T,
as evident in this study [6]. It was also less prone to
magnetic susceptibility artefact. The reported sensitivities
of this sequence to intraluminal flow voids and K-space
filtering effects obscuring mesenteric changes were not
found to be significant limiting factors in this study [6]. T2-
W HASTE images acquired at shorter TE, without fat
suppression, provided good morphological assessment with
contrast of the normal low-signal bowel wall against the
high signal of mesenteric fat and intraluminal fluid, and
thus, were useful for the initial overview. Complications
and mesenteric changes were also well demonstrated.
True FISP imaging was more difficult at 3T. Signal
characteristics and contrast on true FISP are a function of
T1/T2 ratio for each tissue. However this sequence is
usually T2-weighted because the TE and TR are so short
that the T1 is almost constant [17]. It is relatively
insensitive to intraluminal flow voids. At lower field

Fig. 4 Short TE (88 ms) T2-W HASTE sequence without fat Fig. 5 Long TE (251 ms) T2-W HASTE sequence demonstrates
saturation demonstrates bowel wall and mesenteric structures bowel wall oedema (arrows) in a loop of bowel with CD
1622
Fig. 6 True FISP sequence of a segment of bowel with CD with good
bowel distension demonstrates a good overview of the bowel with
minor degradation of the image where there is air within the bowel
strengths, this sequence is useful for delineating overall
anatomy [18]. However, in this study true FISP was,
overall, the sequence most severely limited by susceptibil-
ity artefact from colonic contents and by inadequate
distension. This is because at 3T, true FISP images can
undergo marked degradation by off-resonance effects from
magnetic field inhomogeniety and susceptibility effects
from air and faecal residue in the bowel [12, 19, 20].
However, with good preparation and optimal bowel
distension, this sequence provided good overall anatomical
detail with markedly improved scores. Bowel wall oedema
was not as well depicted as on T2-W HASTE sequence.
This is explained by the reduced tissue contrast at 3T, due
to the decrease in flip angle necessitated by SAR
constraints, and the increase of minimum TR and TE due
to the lower radio-frequency pulse used [20]. In addition,
the black boundary artefact at fat-water interfaces (thin dark
Fig. 7 CE images of a segment
of bowel with CD demonstrates
(a) superficial mucosal
enhancement at 2 min and
(b) transmural enhancement
at 5 min
Pediatr Radiol (2010) 40:1615–1624
rim, India ink or chemical shift artefact) can affect image
quality and potentially impede the perception of subtle
bowel wall thickening on true FISP images (Fig. 9) [19,
20]. This study suggests that true FISP does not add
additional diagnostic information and is often degraded by
artefact and the authors no longer perform it as part of the
routine protocol.
Post-gadolinium T1 imaging was very useful in demon-
strating disease activity. Abnormal bowel wall enhancement
correlates well with active inflammation identified clinical-
ly, with biochemical markers, endoscopy and biopsy [9,
21]. In this study, it was very helpful for assessing
superficial bowel wall inflammation with minimal wall
thickening. This may be difficult to identify on other
sequences.
Some series have attempted correlation of enhancement
pattern with disease stage and pathological changes in the
bowel wall [23]. Reports mention striated wall enhance-
ment to be highly specific for active disease and transmural
enhancement to be associated with chronic, fibrotic change
[6, 22]. In this study, active inflammation was evident as
striated, superficial and transmural patterns of enhance-
ment. Some segments with a superficial pattern of
enhancement observed on the early acquisitions (within
2 min of intravenous contrast administration) demonstrated
uniform transmural enhancement on the delayed images
obtained 4–5 min after contrast (Fig. 7). No change in
enhancement pattern was evident on further delayed
acquisitions (10 min after contrast). To the authors’
knowledge, this finding has not been previously reported.
Studies have demonstrated MRI to be superior to both
CT and fistulography for accurate depiction of compli-
cations of CD (abscesses and sinus and fistulous tracts)
in a non invasive manner [4, 23]. In this study, these
complications were identifiable on both T2-W as well as
CE T1-W imaging, though, at times, better characterised
on the latter [24].
Pediatr Radiol (2010) 40:1615–1624 1623

Fig. 9 True FISP sequence shows black boundary artefact at fat-water

interfaces that can impede the perception of bowel wall detail

This study is limited by the lack of a practical

satisfactory gold standard in the examination of the small
bowel. It is also retrospective and the low incidence of CD
reduces the volume of children available for study.
Nevertheless, the study does demonstrate that 3T evaluation
of the bowel is possible in children with CD.

Conclusion

With high soft-tissue contrast, 3T MRI assessment for

abdominal and perineal CD activity, extent and complica-

Fig. 8 Axial T2-W FSE of the pelvis shows (a) a horseshoe collection
around the anal canal and (b) shows the wall and confirms the fluid
content. Sinus tract is demonstrated on both images (arrow)

B1 inhomogeneity effect was evident as relatively dark

areas, especially on the gradient-echo images (Fig. 10) [12].
In this study, this was controlled by appropriate shimming
and changing the acquisition plane. It was never found to
limit the diagnostic image quality [19]. As these studies
were all performed for targeted assessment of the bowel
rather than the solid abdominal organs (with appropriate
adjustment of the isocentre), peripheral image distortion
was also not a factor limiting evaluation. The increased
chemical shift between water and fat allows for relatively
easier fat suppression at 3T. However, susceptibility artefact
can result in inhomogeneous fat suppression. We did not
find significant interference to fat suppression from
susceptibility or off resonance effects, with uniform fat
Fig. 10 True FISP image demonstrates B1 inhomogeneity effect with
suppression and good contrast enhancement obtained in all a dark band over the liver, left side of the image brighter than right and
studies. multiple parallel lines over the bowel
1624
tions can be performed without sedation or ionising
radiation and in a noninvasive manner.
This study demonstrates that with appropriate attention
to technique, optimal distension and control of movement,
high-quality 3T assessment of the bowel for childhood CD
can be performed. All sequences used at 1.5T can be used
at 3T, though in this study, true FISP was the most prone to
artefact.
References
1. Hyams JS et al (2007) Inflammatory bowel disease. In: Kliegman
RM, Behrman RE, Jenson HB (eds) Nelson textbook of
Pediatrics, 18th edn. Saunders, Detroit, pp 1580–1585
2. Mamula P, Markowitz JE, Baldassano RN (2003) Inflammatory
bowel disease in early childhood and adolescence: special
considerations. Gastroenterol Clin North Am 32:967–995
3. Dagia C, Ditchfield M, Kean M et al (2008) Imaging for Crohn
disease: use of 3-T MRI in a paediatric setting. J Med Imaging
Radiat Onc 52:480–488
4. Essary B, Kim J, Anupindi S et al (2007) Pelvic MRI in children
with Crohn disease and suspected perianal involvement. Pediatr
Radiol 37:201–208
5. IBD Working Group of the European Society for Paediatric
Gastroenterology, Hepatology and Nutrition (2005) Inflamma-
tory bowel disease in children and adolescents: recommenda-
tions for diagnosis—the Porto criteria. J Pediatr Gastroenterol
Nutr 41:1–7
6. Toma P, Granata C, Magnano G et al (2007) CT & MRI of
paediatric Crohns disease. Pediatr Radiol 37:1083–1092
7. Magnano G, Granata C, Barabino A et al (2003) Polyethylene
glycol and contrast-enhanced MRI of Crohn’s disease in children:
preliminary experience. Pediatr Radiol 33:385–391
8. Hugot JP, Bellaiche M (2007) Inflammatory bowel diseases: the
paediatric gastroenterologist’s perspective. Pediatr Radiol
37:1065–1070
9. Alexopoulou E, Roma E, Loggitsi D et al (2009) Magnetic
resonance imaging of the small bowel in children with idiopathic
Pediatr Radiol (2010) 40:1615–1624
inflammatory bowel disease: evaluation of disease activity. Pediatr
Radiol 39:791–797
10. van Gemert-Horsthuis K, Florie J, Hommes DW et al (2006)
Feasibility of evaluating Crohn’s disease activity at 3.0 Tesla. J
Magn Reson Imaging 24:340–348
11. Schick F (2005) Whole-body MRI at high field: technical
limitations and clinical potential. Eur Radiol 15:946–959
12. Rimola J, Rodríguez S, García-Bosch O et al (2009) Role of 3.0-T
MR Colonography in the evaluation of inflammatory bowel
disease. Radiographics 29:701–719
13. Debatin JF, Patak MA (1999) MRI of the small and large bowel.
Eur Radiol 9:1523–1534
14. Prassopoulos P, Papanikolau N, Grammatikakis J et al (2001) MR
enteroclysis imaging of Crohn disease. Radiographics 21:S161–
S172
15. Schindera ST, Merkle EM, Dale BM et al (2006) Abdominal
magnetic resonance imaging at 3.0 T what is the ultimate gain in
signal-to-noise ratio? Acad Radiol 13:1236–1243
16. Magnano G, Granata C, Barabino A et al (2003) Polyethylene
glycol and contrast-enhanced MRI of Crohn’s disease in children:
preliminary experience. Pediatr Radiol 33:385–391
17. Horsthuis K, Lavini C, Stoker J (2005) MRI in Crohn’s Disease. J
Magn Reson Imaging 22:1–12
18. Martin DR, Danrad R, Herrmann K et al (2005) Magnetic
resonance imaging of the gastrointestinal tract. Top Magn Reson
Imag 16:77–98
19. Patak MA, von Weymarn C, Froehlich JM (2007) Small bowel
MR imaging: 1.5 T versus 3 T. Magn Reson Imaging Clin N Am
15:383–393
20. Lauenstein TC, Saar B, Martin DR (2007) MR colonography:
1.5 T versus 3 T. Magn Reson Imaging Clin N Am 15:395–402
21. Malago R, Manfredi R, Benini L et al (2008) Assessment of
Crohn’s disease activity in the small bowel with MR-enteroclysis:
clinico-radiological correlations. Abdom Imaging 33:669–675
22. Koh DM, Miao Y, Chinn RJ et al (2001) MR imaging evaluation
of the activity of Crohn’s disease. AJR 177:1325–1332
23. Koelbel G, Schmiedl U, Majer MC et al (1989) Diagnosis of
fistulae and sinus tracts in patients with Crohn disease: value of
MR imaging. AJR 152:999–1003
24. Clinical Practice Committee (2003) American Gastroenterological
Association: medical position statement: perianal Crohn’s disease.
Gastroenterology 125:1503–1507