Which magnetic resonance imaging findings accurately evaluate inflammation in small bowel Crohnʼs...

ORIGINAL ARTICLE

Which Magnetic Resonance Imaging Findings AccuratelyEvaluate Inflammation in Small Bowel Crohn’s Disease? ARetrospective Comparison with Surgical Pathologic Analysis

Magaly Zappa, MD,* Carmen Stefanescu, MD,† Dominique Cazals-Hatem, MD,‡

Frederic Bretagnol, MD,§ Lydia Deschamps, MD,‡ Alain Attar, MD,† Beatrice Larroque, MD,k

Xavier Treton, MD,† Yves Panis, MD,§ Valerie Vilgrain, MD,* and Yoram Bouhnik, MD†

Background: The aim was to evaluate the value of magnetic

resonance imaging (MRI) findings in Crohn’s disease (CD) in cor-

relation with pathological inflammatory score using surgical pa-

thology analysis as a reference method.

Methods: CD patients who were to undergo bowel resection sur-

gery underwent MR enterography before surgery. The CD patholog-

ical inflammatory score of the surgical specimens was classified

into three grades: mild or nonactive CD, moderately active CD, and

severely active CD; fibrosis was also classified into three grades:

mild, moderate, and severe. Mural and extramural MRI findings

were correlated with pathological inflammatory and fibrosis grades.

Results: Fifty-three consecutive patients were included retrospec-

tively. The mean delay between MRI and surgery was 24 days (range

1–90, median 14). The CD pathological inflammatory score was

graded as follows: grade 0 (11 patients, 21%), grade 1 (15 patients,

28%), and grade 2 (27 patients, 51%). MRI findings significantly asso-

ciated with pathological inflammatory grading were wall thickness (P< 0.0001), degree of wall enhancement on delayed phase (P <0.0001), pattern of enhancement on both parenchymatous (P ¼ 0.02),

and delayed phase, (P ¼ 0.008), T2 relative hypersignal wall (P <0.0001), blurred wall enhancement (P ¼ 0.018), comb sign (P ¼0.004), fistula (P < 0.0001), and abscess (P ¼ 0.049). The inflamma-

tion score correlated with the fibrosis score (r ¼ 0.63, P¼ 0.0001).

Conclusions: Our study identified MRI findings significantly

associated with surgical pathological inflammation. These lesions

are considered potentially reversible and may be efficiently treated

medically. We also showed that fibrosis was closely and posi-

tively related to inflammation.

(Inflamm Bowel Dis 2011;17:984–993)

Key Words: Crohn’s disease, MRI, pathology, inflammation

C rohn’s disease (CD) is a chronic inflammatory bowel

disorder characterized by remittent and relapsing epi-

sodes.1 Assessment of inflammatory activity is crucial to

decide optimal treatment: patients with severe inflammatory

lesions receive medical therapy, while patients without or

with moderate inflammatory lesions associated with ob-

structive symptoms undergo surgery.2,3

Numerous criteria are used to assess disease activity

including clinical symptoms, physical findings, laboratory

parameters, and endoscopic results.4 However, clinical and

biological criteria must be supported by results of other

diagnostic techniques because these scores are unreliable4,5;

moreover, clinical activity seems to be independent of the

severity of mucosal lesions and biological activity.5,6

Conventional and/or wireless capsule endoscopies

only provide information about the intestinal mucosa, while

conventional small bowel contrast studies only provide

indirect information about the intestinal wall. These limita-

tions are important because inflammation in CD is com-

monly transmural.1,7

New imaging tools such as computed tomography

(CT) and magnetic resonance imaging (MRI) provide infor-

mation about the mucosa, the intestinal wall, and adjacent

structure abnormalities, including complications such as fis-

tula or abscesses.8,9

The accuracy of CT enteroclysis seems to be good

for evaluating CD. Chiorean et al10 correlated CT findings

with surgical pathology to assess the accuracy of CT enter-

oclysis in differentiating inflammatory from fibrostenotic

lesions. A scoring system was defined to reliably differenti-

ate these lesions. However, because of the radiation doses

of CT scan and the increased risk of radiation-induced can-

cers,11,12 especially in young people, reliable nonradiation

diagnostic techniques must be developed to follow these

patients who require regular evaluations.13

Because of its excellent soft-tissue contrast, multipla-

nar images, the absence of exposure to radiation, safe con-

trast agents, and ability to show mural and extramural

Received for publication May 16, 2010; Accepted June 4, 2010.

From the *Department of Radiology, †PMAD Department of

Gastroenterology, ‡Department of Pathology, §PMAD, Department of

Colorectal Surgery, ||Department of Epidemiology, Hopital Beaujon, AP-

HP, Universite Paris 7, INSERM CRB3, Clichy, France.

Reprints: Magaly Zappa, MD, Service de Radiologie, Hopital Beaujon,

100, boulevard du General Leclerc, 92110 Clichy, France (e-mail: magaly.

[email protected])

Copyright VC 2010 Crohn’s & Colitis Foundation of America, Inc.

DOI 10.1002/ibd.21414

Published online 18 August 2010 in Wiley Online Library

(wileyonlinelibrary.com).

Inflamm Bowel Dis � Volume 17, Number 4, April 2011984

abnormalities,14 MRI is becoming the standard imaging

technique for CD.3 Several studies have shown that MRI

plays a role in the assessment of CD activity, using differ-

ent variables.9,15–17 These variables have only been corre-

lated with clinical and biological scores (e.g., the Crohn’s

Disease Activity Index [CDAI]) or with endoscopic find-

ings, which all have certain limits because CD is a mural

and transmural disease.

To our knowledge, only one study has compared

MRI findings with the pathological analysis of surgical

specimens, but with a small number of patients.18 The aim

of our study was to determine the diagnostic value of MRI

findings in CD in correlation with inflammatory activity

with surgical pathological analysis as a reference method.

PATIENTS AND METHODS

PatientsEligibility criteria for this retrospective monocentric

cohort study in Beaujon Hospital (Clichy, France) were:

consecutive adult patients with documented CD who were

to undergo bowel resection for small bowel CD; small

bowel MRI within 3 months prior to surgery; and stable

medical treatment from MRI until surgery. All consecutive

patients seen from June 2006 to February 2009 were

included.

Surgical treatment of patients was decided during our

weekly medical and surgical rounds, where all patients

potentially requiring surgery for CD are discussed. More-

over, our policy has been to systematically perform MRI

before any surgical procedure in CD patients.

Patients receiving tumor necrosis factor antibody

(anti-TNF) treatment within 2 months before MRI or

between MRI and surgery were excluded because this treat-

ment can induce mucosal healing in CD.19

The following information was collected for all

patients: gender, age at surgery, duration of disease at sur-

gery, disease activity at surgery estimated by the Harvey–

Bradshaw index (HBI),20 and biological parameters (C-re-

active protein [CRP], leukocyte count), surgical indication,

previous bowel surgeries, and type of surgery. The study

was approved by our Institutional Review Board.

Methods

MRI TechniqueAfter oral administration of 1000 mL of hyperos-

motic water solution (mannitol 5%), MRI was performed

with a 1.5 T system (Intera, Philips Medical Systems, Best,

The Netherlands) using a body coil. Breath-hold imaging

was first performed in the coronal plane using a T2-

weighted single shot turbo spin echo sequence with fat sup-

pression (TR/TE 20,364/100, matrix 243 � 512, turbo fac-

tor 71, slice thickness 5 mm, 25 slices per image set, 20-

second breath-hold duration), and in the coronal and axial

planes using a true Fast Imaging with Balanced Steady-

state (true FISP) sequence (TR/TE 3/1.5, matrix 256 �256, flip 60�, slice thickness 5–8 mm, 25–40 slices per

image set, 21-second breath-hold duration). After intrave-

nous administration of 1 mg of glucagon (GlucaGen, Novo

Nordisk, Bagsvaerd, Denmark), a 3D FLASH T1-weighted

sequence was performed before and 90 seconds (coronal

plane) and 8 minutes (axial and coronal plane) after an in-

travenous injection of gadolinium chelates (Dotarem, Guer-

bet, France) at a dose of 0.2 mL/kg of body weight, with

fat suppression (TR/TE 3.7/1.3, matrix 269 � 236, flip 10�,slice thickness 8 mm, 40 slices, 20-second breath-hold du-

ration). Ninety seconds was considered the parenchymatous

time and 8 minutes the delayed time.16

MRI Data AnalysisAll MRI studies were reviewed by two experienced

radiologists (M.Z., V.V.), who were unaware of clinical,

surgical, or laboratory data for the patients, and a consen-

sus was reached.

The following imaging findings were evaluated:

• Degree of wall enhancement on both parenchymatous

(90 seconds) and delayed (8 minutes) T1-weighted

images, scored as moderate or marked, compared with

enhancement of adjacent normal bowel.16

• Pattern of enhancement: homogeneous or layered (i.e.,

visualization of two or three layers of the bowel wall) on

parenchymatous and on delayed T1-weighted images.5

• Well-defined or blurred wall enhancement on delayed

T1-weighted images.

• Wall signal intensity on T2-weighted images,17 evaluated

as mild or high.

• Wall thickness in both T2 (true FISP) and T1-weighted

images.

• The presence of upstream intestinal loop dilatation (at

least 1.5 � normal loop).

• Extramural findings: the presence or absence of fatty

proliferation (i.e., space-occupying lesion, separating

and/or displacing adjacent small bowel loops),16 comb

sign (i.e., segmental dilatation or engorgement of the

vasa recta),10 lymph nodes (i.e., presence of any lymph

nodes in the mesentery), fistula, and/or abscesses.

The length and location of abnormal segments were

also recorded.

Pathological AnalysisAll surgical small-bowel specimens showed CD

lesions and were analyzed prospectively by macroscopic

examination of unfixed specimens and microscopic exami-

nation after fixation with formalin and standard hematoxy-

lin-eosin safran stain. One expert gastrointestinal patholo-

gist (D.C.H.) who was unaware of the clinical, biological,

Inflamm Bowel Dis � Volume 17, Number 4, April 2011 MRI in Small Bowel CD

985

and radiological findings examined all specimens and

stained histological sections. Two main features were eval-

uated during the macroscopic examination: total length

(centimeters) of the resected specimens and the presence of

significant luminal strictures; the specimen was then care-

fully examined to identify any fistula or abscesses as

described by Borley et al.21 The thickness of the intestinal

wall was measured in millimeters on whole-mount slides

of transverse sections of the most severely involved intesti-

nal segment.

At least 10 samples (blocks) were prepared for each

surgical specimen to include all lesions observed macro-

scopically (strictures, ulcerations, fissures, and fistula) and

make sure that there was a representative evaluation of all

pathological changes. The most severe abnormalities

observed microscopically were selected to score inflamma-

tion and fibrosis.

The CD pathological inflammatory score routinely

used at Beaujon’s university hospital was graded into three

categories (grades 0-1-2) as follows: grade 0 (mild or non-

active CD): minimal neutrophil infiltrate limited to muco-

sae; grade 1 (moderately active CD): neutrophil infiltrate

limited to the mucosa and submucosa without muscular

involvement; grade 2 (severely active CD): transmural neu-

trophil infiltrate through the muscularis propria and/or fis-

tula and/or abscesses in the subserosa (Fig. 1A–C).

We also graded inflammation in all patients using the

score described by Chiorean et al,10 which includes both

acute and chronic inflammation.

Fibrosis was then graded into three categories (grade

0-1-2) as follows: grade 0: minimal fibrosis limited to sub-

mucosa; grade 1: massive submucosal fibrosis with pre-

served layers; grade 2: massive transmural fibrosis with

effacement of normal layers. We were unable to use Chior-

ean et al’s fibrosis score because of difficulties to grade

some items such as mild versus severe strictures or level of

muscular hyperplasia.

Statistical AnalysisSurgical pathology was the reference method for

assessing the accuracy of MRI for the diagnosis of CD

inflammation and fibrosis.

The correlations between the two inflammation scores

(CD pathological inflammatory score and the one used by

Chiorean et al10) and between CD pathological inflamma-

tory score and fibrosis score (as defined by the scoring sys-

tem above), were estimated using the Mantel–Haenszel v2,Spearman rank correlation coefficient.

The CD pathological inflammatory score was corre-

lated with thickness of the bowel wall. Sensitivity and

specificity were calculated for different wall thicknesses to

differentiate inflammation grade 0 patients and upper

grades, and grade 2 patients and lower grades. Sensitivity,

specificity, and accuracy were calculated for the detection

of fistula and/or abscesses by MRI and pathology.


lated with MRI variables for mural findings (degree of

enhancement in parenchymatous phase and delayed phase,

pattern of enhancement in parenchymatous phase and

delayed phase, T2 wall hypersignal, blurred wall enhance-

ment, upstream dilatation, wall thickness) and extramural

findings (Comb sign, fistula and abscesses, lymph nodes,

fatty proliferation) using the v2 Fisher exact test and Krus-

kal–Wallis test.

The sign-by-sign correlation between MRI and patho-

logical variables was performed for wall thickness, length

FIGURE 1. Whole-mount slides from transversal sections of affected small bowels illustrating inflammatory grades (hemato-xylin-eosin safran stain). (A) Grade 0: no active inflammation was observed (no neutrophil infiltrate); marked transmural fibro-sis with obliteration of normal layers causing severe stricture. (B) Grade 1: neutrophil infiltrate limited to the mucosa andsubmucosa creating fissures, without alteration of the muscularis. (C) Grade 2: massive neutrophil infiltrate and with trans-mural penetration creating fistula and abscesses into the serosa and the extramural fat.

Inflamm Bowel Dis � Volume 17, Number 4, April 2011Zappa et al

986

of affected segments, and presence of fistula and/or

abscesses using v2 Fisher exact test, Spearman rank corre-

lation coefficient, and Kruskal–Wallis test.


lated with the HBI, CRP, and white cell count using Krus-

kal–Wallis test. No specific hypothesis was performed for

missing data.

RESULTS

Demographic and Clinical DataFifty-three patients were included. The demographic

and clinical features are presented in Table 1. Eleven

patients who had received anti-TNF treatment within 2

months before surgery were excluded. Fourteen patients

who had previously been treated with an anti-TNF but the

last infusion was administered more than 2 months before

surgery were included. The mean delay between small

bowel MRI and surgery was 24 days (range 1–90, median

14). MRI was not performed or was performed more than

3 months before surgery in 9 and 11 patients, respectively,

who were excluded. Figure 2 shows a flowchart of patients.

Indications for surgery in our 53 patients were bowel

obstruction in 23 (43%) patients, refractory nonobstructive

disease in three (6%), and perforating disease (refractory

fistula or abscess) in 15 (28%); 12 (23%) patients had

more than one indication for surgery.

Evaluation of clinical activity by the HBI was

assessed in 49 patients and showed that 40 (82%) patients

TABLE 1. Demographic Characteristics of 53 PatientsOperated on for Small Bowel Crohn’s Disease

FactorNumber ofPatients (%)

Age Mean yrs 35 (15–74)

Sex Men 28 (53%)

Women 25 (47%)

Duration of disease Mean yrs 10 (0–34)

Location of diseaseat entry

Small bowel 6 (11%)

Ileocolonic 47 (89%)

Disease behavior for smallbowel according toMontreal classification

B1 inflammatory 6 (11%)

B2 stenotic 32 (61%)

B3 penetrating 15 (28%)

Pharmacotherapy at thetime of surgery

No treatment 18 (34%)

Steroids 6 (11%)

Immunosuppressant 11 (21%)

Steroids andimmunosuppressant

4 (8 %)

History of surgerybefore inclusion

Segmental smallbowel resection

7 (13%)

Ileocecal resection 17 (32%)

Colonic resection 6 (11%)

Smoking history Never 26 (49%)

Current* 10 (19%)

Previous 17 (32%)

More than 1 cigarette/day.

FIGURE 2. Flowchart of patient inclusion in the study.


987

had active disease (HBI �4) and 9 (18%) had mild or non-

active disease. Disease activity could not be assessed in

four patients because of the presence of stoma. The inflam-

matory syndrome was evaluated by the level of CRP

(available at the time of MRI in 39 (74%) patients) and the

white cell count in all patients. The association between

these parameters and CD pathological inflammatory score

is summarized in Table 2.

Pathological FeaturesAll intestinal resection specimens showed the fibro-

inflammatory changes of CD with a uniform pattern in

samples (blocks) from the same patient. The distribution of

the 53 patients into the three grades according to the CD

pathological inflammatory score was as follows: grade 0

(11 patients, 21%), grade 1 (15 patients, 28%), and grade 2

(27 patients, 51%). Their distribution into the three grades

according to the fibrosis score was as follows: grade 0 (two

patients, 4%), grade 1 (15 patients, 28%), and grade 2 (36

patients, 68%). This is summarized in Table 3.

CD pathological inflammatory score was correlated

with the fibrosis score (r ¼ 0.63, P ¼ 0.0001) and with the

inflammatory score used by Chiorean et al10 (r ¼ 0.80, P¼ 0.0001).

There was no significant difference in the length of the

specimens according to CD pathological inflammatory score;

the length was 21.5 6 25 cm in grade 0 patients, 35.3 6 40

in grade 1 patients, and 24.8 6 9.6 cm in grade 2 patients.

Macroscopic examination showed 44/53 (83%)

patients with strictures and 25/53 (47%) patients with fis-

tula or abscesses. Among them, 20/25 (80%) had concomi-

tant strictures.

The thickness of the bowel wall on whole-mount

slides was strongly correlated with the CD pathological

inflammatory score (6.7 6 1.4 mm for grade 0, 11.2 6 3

mm for grade 1, and 17.3 6 3.8 mm for grade 2, P <0.0001).

Correlations Between MRI and PathologicalCharacteristics

The location of abnormal intestinal segments was the

same for resected surgical specimens and MRI findings.

The mean length of surgical specimens was 271 6 252

mm, the mean length of involved bowel measured by MRI

was 228 6 164 mm (r ¼ 0.52, P ¼ 0.0002). The mean

wall thickness measured by the pathologist on slide sec-

tions was 13.4 6 5.4 mm.

The mean wall thickness measured by MRI was 8.9 62.6 mm and 9.3 6 2.6 mm on T2- and T1-weighted images,

respectively, and was significantly correlated in each patient

(r ¼ 0.9, P < 0.0001). The correlation with pathological

wall thickness measurements was also significant (r ¼ 0.68,

P < 0.0001 for T2 measurements, r ¼ 0.70, P < 0.0001 for

T1 measurements). To differentiate patients with grade 0

from others (grade 1 or 2), a threshold of 6 mm had a sensi-

tivity and specificity of 63.6% and 92.9%, and a threshold of

7 mm had a sensitivity and specificity of 90.9% and 78.6%,

respectively (on T2 sequences). To differentiate patients with

grade 2 from the others (grade 0 or 1), a threshold of 10 mm

had a sensitivity and specificity of 55.6% and 84.6%, respec-

tively, and a threshold of 11 mm had a sensitivity and speci-

ficity of 40.7% and 92.3%, respectively.

Detection of fistula and/or abscesses by MRI had a

sensitivity, specificity, and accuracy of 88%, 93%, and

91%, respectively (P < 0.0001).

Upstream dilatation of intestinal loops on MRI was

not significantly associated (P ¼ 0.71) with the pathologi-

cal finding of stenosis.

TABLE 2. Association Between CD PathologicalInflammatory Score in Surgical Specimens, ClinicalCrohn’s Disease Activity (HBI), and Biologic Inflammationin 53 Patients Operated on for Crohn’s Disease (CD)

CDPathologicalInflammatoryScore*

Harvey-BradshawIndex (n ¼ 49)Mean 6 SE (n)

C-reactiveProtein

mg/L (n ¼ 39)Mean 6 SE (n)

White CellCount G/L(n ¼ 53)

Mean 6 SE (n)

Grade 0 5.8 6 2.3 (9) 4.2 6 3.2 (9) 7.6 6 3.9 (11)

Grade 1 7.1 6 3.3 (5) 12.0 6 16.1 (8) 8 6 4.8(15)

Grade 2 6.3 6 2.9 (25) 39.5 6 48.0 (22) 10 6 4.3 (27)

P 0.71 0.0014 0.051

*Grade 0: mild or nonactive CD; grade 1: moderately active CD; grade 2:severely active CD.

TABLE 3. Distribution of the 53 Surgical Specimens ofPatients Operated on for Crohn’s Disease According tothe CD Pathological Inflammatory and the FibrosisScores

Fibrosis Scorea

Grade 0 1 2 Total

CD pathologicalinflammatory scoreb

0 2 4 5 11

1 0 11 4 15

2 0 0 27 27

Total 2 15 36 53

aGrade 0: no or minimal fibrosis limited to submucosa; grade 1: massivesubmucosal fibrosis with preserved layers; grade 3: massive transmuralfibrosis with effacement of normal layers.bGrade 0: mild or nonactive CD; grade 1: moderately active CD; grade 2:severely active CD.


988

Association Between MRI signs, CD PathologicalInflammatory, and Fibrosis Scores

Table 4 shows the results of associations for each

imaging sign within the three groups according to CD path-

ological inflammatory score (Figures 3A-C). When compar-

ing the association between MRI signs and CD pathologi-

cal inflammatory score or between MRI signs and the

inflammatory score described by Chiorean et al, we found

three signs significantly associated with our score but not

with the one of Chiorean et al: Pattern of enhancement

delayed phase (P ¼ 0.008 versus 0.5), blurred wall

enhancement (P < 0.001 versus 0.14), and abscess (P ¼0.049 versus 0.59).

Table 5 shows the results of associations for each

imaging sign within the three groups according to fibrosis

score.

DISCUSSIONOur study permitted us to identify MRI signs signifi-

cantly associated with pathological inflammation. The abil-

ity to recognize inflammation is of major interest in the

management of patients in the clinical setting, as this con-

dition is potentially reversible using medical therapy, and

its absence usually leads to surgical therapy. Two analyti-

cal methods (MRI and pathological analysis of resected

specimens) were selected that evaluate both mural and

extramural components. Indeed, histological analysis of in-

testinal resection is essential to determine CD activity

because active lesions are often transmural, enclosing fis-

sures or fistula mixed with edema and neutrophil infiltra-

tion if the disease is still active.

Certain mural MRI findings were found to be associ-

ated with our CD pathological inflammatory score of the

TABLE 4. MRI Findings and CD Pathological Inflammatory Score in 53 Surgical Specimens of Patients Operated on forCrohn’s Disease

CD Pathological Inflammatory Score

Grade 0*(n ¼ 11) % (n)

Grade 1*(n ¼ 15) % (n)

Grade 2*(n ¼ 27) % (n) P

Wall thickness (mm; m6SD)

T2 sequences 6.1 6 1 8.4 6 1.8 10.3 6 2.4 <0.0001

T1 sequences 6.2 6 1.3 9.2 6 1.8 10.6 6 2.2 <0.0001

Mural findings

Degree of enhancement

Parenchymatous phase 0.57

Mild 55% (6) 33% (5) 37% (10)

Marked 45% (5) 67% (10) 63% (17)

Delayed phase <0.0001

emsp;Mild 64% (7) 47% (7) 7% (2)

Marked 36% (4) 53% (8) 93% (25)

Pattern of enhancement


Layered 36% (4) 80% (12) 81% (22)

Homogeneous 64% (7) 20% (3) 19% (5)

Delayed phase 0.008

Layered 18% (2) 73% (11) 30% (8)

Homogeneous 82% (9) 27 (4) 70% (19)

T2 wall hypersignal 0% (0) 0% (0) 56% (15) <0.0001

Blurred wall enhancement 0% (0) 0% (0) 37% (10) 0.018

Upstream dilatation 38% (4) 47% (7) 44% (12) 0.87

Extramural findings

Comb sign 36% (4) 67% (10) 89% (24) 0.004

Fistula 0% (0) 13% (2) 81% (22) <0.0001

Abscess 0% (0) 7% (1) 30% (8) 0.049

Lymph nodes 36% (4) 73% (11) 56% (15) 0.19

Fatty proliferation 64% (7) 47% (7) 70% (19) 0.48

*Grade 0: mild or non active CD; grade 1: moderately active CD; grade 2: severely active CD.


989

disease. Wall thickness on either T1- or T2-weighted

sequences as well as wall thickness on whole-mount slides

were shown to be significantly increased with disease activ-

ity. Our results show that a threshold of 6 mm can be used

to distinguish between patients with inactive disease (grade

0) and active disease (grade 1 or 2) (sensitivity and speci-

ficity, 63.6% and 92.9%, respectively). This value agrees

with other studies which suggest that a normal bowel wall

is between 1 and 3 mm thick, and the minimal value for

wall thickness in CD is 4 mm22 or 5 mm.16,23 Furthermore,

a threshold of 11 mm provides a distinction between severe

active CD patients (grade 2) and others (grade 0 or 1) (sen-

sitivity and specificity, 40.7% and 92.3%, respectively).

Although this measurement may be obtained on T1- or T2-

weighted sequences, it was easier on T2 sequences. This

finding was found to be significantly associated in most CT

or MRI studies.10,15–17,24

Bowel wall enhancement has also been extensively

investigated in CD patients. However, results from the lit-

erature are confusing because various acquisition times,

different enhancement patterns (homogeneous, layered, mu-

cosal), and various estimations of the degree of wall

enhancement10,15–17 have been used. Interestingly, all the

enhancement criteria evaluated in this study were signifi-

cantly related to the CD pathological inflammatory score in

the different grades, except the degree of wall enhancement

at the parenchymatous phase. The most striking finding to

differentiate inactive CD (grade 0) from moderately active

(grade 1) and severely active (grade 2) CD patients was

the degree of wall enhancement on delayed T1-weighted

sequences. Therefore, delayed T1-weighted sequences seem

to play an important role in assessing disease activity.

As reported by others,17 wall signal intensity in T2-

weighted sequences was associated with disease activity in

this series. We also found that a blurred enhancement pat-

tern in delayed phase imaging was well correlated with ac-

tivity grading, and was essentially observed in severely

active CD patients. This probably corresponds to

FIGURE 3. MRI findings: delayed T1-weighted images illustrating CD inflammatory grades. (A) Grade 0: coronal image in a24-year-old male; the pathological wall is 5 mm thick, with mild and homogeneous enhancement; there is no comb sign. (B)Grade 1: coronal image in a 59-year-old female; the pathological wall is 9 mm thick, with marked and layered enhancement.Note the presence of fatty proliferation with displaced adjacent loops. (C) Grade 2: axial image in a 20-year-old female; thepathological wall is 11 mm thick, with marked, homogeneous, and blurred enhancement. Note the presence of a cutaneousfistula.


990

transmural inflammation extending to adjacent mesentery

fat, described by Colombel et al5 as perienteric inflamma-

tion, which correlated well with increased CRP, as in our

study.

Certain extramural MRI findings were significantly

associated with disease activity: the presence of the comb

sign, fistula, and abscesses, findings which were mostly

observed in active CD patients. Despite one negative

study,16 the comb sign has mostly been reported in active

CD patients.10,15,17 Diagnosis of abscesses is a major issue

in clinical practice because it changes the course and man-

agement of the disease and usually requires interventional

treatment.17

Conversely, no relationship was found between the

presence of lymph nodes or fatty proliferation and disease

activity. Chiorean et al and Gourtsoyiannis et al observed a

significant number of lymph nodes in active CD

patients.10,15–17,24 This discrepancy could be explained

because we considered all lymph nodes whatever the size,

number, or enhancement.

To our knowledge, only two studies compared MRI

and surgical pathological findings, including 18 and 17

patients, respectively.18,25 One of these studies did not

mention pathological classification reference and used

MRI-determined categories.25

Most of the findings that were significant in our study

were also described in the study of Punwani et al.18 Wall

thickness, T2 hypersignal, and pattern of enhancement are

associated with histologic markers of acute inflammation,

whereas degree of enhancement at parenchymatous time

(70 seconds for this study) and lymph nodes are not associ-

ated.18 Moreover, we think our study brings additional

TABLE 5. MRI Findings and Fibrosis Score in 53 Surgical Specimens of Patients Operated on for Crohn’s Disease

Fibrosis Score

Grade 0* (n ¼ 2) % (n) Grade 1* (n ¼ 15) % (n) Grade 2* (n ¼ 36) % (n) P

Wall thickness (mm; m6SD)

T2 sequences 5 6 0 7.5 6 1.6 9.7 6 2.6 0.0018

T1 sequences 5 6 10 8.3 6 1.8 9.9 6 2.5 0.0039

Mural findings

Degree of enhancement


Mild 0% (0) 53% (8) 36% (13)

Marked 100% (2) 47% (7) 64% (23)

Delayed phase 0.15

Mild 50% (1) 47% (7) 22% (8)

Marked 50% (1) 53% (8) 78% (28)

Pattern of enhancement


Layered 50%(1) 80% (12) 69% (25)

Homogeneous 50% (1) 20% (3) 31% (11)

Delayed phase 0.1

Layered 0% (0) 60% (9) 33% (12)

Homogeneous 100% (2) 40 (6) 67% (24)

T2 wall hypersignal 0% (0) 0% (0) 42% (15) 0.026

Blurred wall enhancement 0% (0) 0% (0) 28% (10) 0.062

Upstream dilatation 50% (1) 40% (6) 44% (16) 1

Extramural findings

Comb sign 100% (2) 47% (7) 81% (29) 0.03

Fistula 0% (0) 13% (2) 61% (22) 0.0011

Abscess 0% (0) 7% (1) 22% (8) 0.49

Lymph nodes 100% (2) 73% (11) 47% (17) 0.11

Fatty proliferation 50% (1) 53% (8) 67% (24) 0.68

*Grade 0: no or minimal fibrosis limited to submucosa; grade 1: massive submucosal fibrosis with preserved layers; grade 3: massive transmural fibrosiswith effacement of normal layers.


991

information. We used a pathological score for acute inflam-

mation and fibrosis which was, as in their study, adapted

from the classification of Borley et al.21 However, we sim-

plified this score, which makes it easier for daily practice.

Our data were qualitative and not quantitative, contrary to

their study, as it has been suggested that quantitative meas-

urements are less reliable than visual estimates for identify-

ing active CD.5 Last, we also studied all the findings at a

very delayed time (8 minutes), and we have shown that

this timing is very important in distinguishing moderately

active CD patients (grade 1) from severely active CD

patients (grade 2).

Compared with the study which evaluated CT scan

imaging and surgical pathological findings,10 our pathologi-

cal score has the two major advantages of not mixing

chronic (lymphoid infiltrates) and acute (neutrophil infil-

trates) inflammation and distinguishing perforating and

nonperforating phenotypes.

In our study, fibrosis was evaluated and found to be

well correlated with inflammation. The relationship

between transmural fibrosis and the severity of acute

inflammation (deep fissures and fistula) is consistent with

previous pathological studies, showing the correlation

between fistula formation and strictures with massive mural

fibrosis.26,27 This correlation between acute inflammation

and fibrosis is an important and innovative point, suggest-

ing that it may not be relevant to make an exclusive dis-

tinction, as is usually done, between inflammatory patients

and fibrotic patients. Moreover, we did not find any corre-

lation between the fibrosis score and MRI pattern or degree

of enhancement wall.

The present study has certain limitations. First, it is a

retrospective study. Second, we deliberately chose oral

administration of contrast medium and therefore did not

investigate certain mural criteria such as superficial ulcers,

fold thickening, and nodularity. As shown previously by

Gourtsoyiannis et al,16 MRI enteroclysis provides better

results for these subtle abnormalities than MRI without

enteroclysis. Nevertheless, although these are early CD

findings, the relationship between their presence and dis-

ease activity is questionable.16 Moreover, naso-jejunal tube

placement is uncomfortable, making movement artifacts

more frequent, and it has clearly been shown that small

bowel MRI with oral administration is much more accepta-

ble to patients than MRI enteroclysis.28

In conclusion, this study identified several MRI find-

ings that were significantly associated with inflammatory

activity in CD using surgical pathological examination as

the reference method. Demonstration of the high accuracy

of MRI for detecting the presence of inflammatory activity

is of significant clinical relevance because of its therapeutic

implications as patients with severe inflammatory lesions

will receive medical therapy, while those without or with

only moderate inflammatory lesions will preferably

undergo surgery. Altogether, these results could help us to

select patients for medical or surgical therapy and also be

used to monitor responses to therapeutic interventions both

in the clinical setting and in therapeutic trials.

REFERENCES1. Podolsky DK.Inflammatory bowel disease. N Engl J Med. 2002;347:

417–429.2. Treton X, Stefanescu C, Attar A, et al. [Ileal stenosis.] Gastroenterol

Clin Biol. 2007;31:387–392.3. Assche GV, Dignass A, Panes J, et al. The second European evi-

dence-based consensus on the diagnosis and management of Crohn’sdisease: definitions and diagnosis. J Crohn’s Colitis. 2010;4:7–27.

4. Sandborn WJ, Feagan BG, Hanauer SB, et al. A review of activityindices and efficacy endpoints for clinical trials of medical therapy inadults with Crohn’s disease. Gastroenterology. 2002;122:512–530.

5. Colombel JF, Solem CA, Sandborn WJ, et al. Quantitative measure-ment and visual assessment of ileal Crohn’s disease activity by com-puted tomography enterography: correlation with endoscopic severityand C reactive protein. Gut. 2006;55:1561–1567.

6. Cellier C, Sahmoud T, Froguel E, et al. Correlations between clinicalactivity, endoscopic severity, and biological parameters in colonic orileocolonic Crohn’s disease. A prospective multicentre study of 121cases. The Groupe d’Etudes Therapeutiques des Affections Inflamma-toires Digestives. Gut. 1994;35:231–235.

7. Schreiber S, Rutgeerts P, Fedorak RN, et al. A randomized, placebo-controlled trial of certolizumab pegol (CDP870) for treatment ofCrohn’s disease. Gastroenterology. 2005;129:807–818.

8. Siddiki HA, Fidler JL, Fletcher JG, et al. Prospective comparison ofstate-of-the-art MR enterography and CT enterography in small-bowelCrohn’s disease. AJR Am J Roentgenol. 2009;193:113–121.

9. Parisinos CA, McIntyre VE, Heron T, et al. Magnetic resonance fol-low-through imaging for evaluation of disease activity in ileal Crohn’sdisease: an observational, retrospective cohort study. Inflamm BowelDis. 2009;16:1219–1226.

10. Chiorean MV, Sandrasegaran K, Saxena R, et al. Correlation of CTenteroclysis with surgical pathology in Crohn’s disease. Am J Gastro-enterol. 2007;102:2541–2550.

11. Brenner DJ, Hall EJ.Computed tomography—an increasing source ofradiation exposure. N Engl J Med. 2007;357:2277–84.

12. Desmond AN, O’Regan K, Curran C, McWilliams S, Fitzgerald T,Maher MM, Shanahan F. Crohn’s disease: factors associated with ex-posure to high levels of diagnostic radiation. Gut. 2008;57:1524–1529.

13. Martin DR, Lauenstein T, Sitaraman SV. Utility of magnetic reso-nance imaging in small bowel Crohn’s disease. Gastroenterology.2007;133:385–390.

14. Masselli G, Vecchioli A, Gualdi GF. Crohn disease of the smallbowel: MR enteroclysis versus conventional enteroclysis. AbdomImaging. 2006;31:400–409.

15. Koh DM, Miao Y, Chinn RJ, et al. MR imaging evaluation of the ac-tivity of Crohn’s disease. AJR Am J Roentgenol. 2001;177:1325–1332.

16. Gourtsoyiannis NC, Grammatikakis J, Papamastorakis G, et al. Imag-ing of small intestinal Crohn’s disease: comparison between MRenteroclysis and conventional enteroclysis. Eur Radiol. 2006;16:1915–1925.

17. Maccioni F, Viscido A, Broglia L, et al. Evaluation of Crohn diseaseactivity with magnetic resonance imaging. Abdom Imaging. 2000;25:219–228.

18. Punwani S, Rodriguez-Justo M, Bainbridge A, et al. Mural inflamma-tion in Crohn disease: location-matched histologic validation of MRimaging features. Radiology. 2009;252:712–720.

19. Colombel JF, Sandborn WJ, Reinisch W, et al. Infliximab, azathio-prine, or combination therapy for Crohn’s disease. N Engl J. Med;362:1383–1395.

20. Harvey RF, Bradshaw JM.A simple index of Crohn’s disease activity.Lancet. 1980;1:514.


992

21. Borley NR, Mortensen NJ, Jewell DP, et al. The relationship betweeninflammatory and serosal connective tissue changes in ileal Crohn’sdisease: evidence for a possible causative link. J Pathol. 2000;190:196–202.

22. Horsthuis K, Stokkers PC, Stoker J.Detection of inflammatory boweldisease: diagnostic performance of cross-sectional imaging modalities.Abdom Imaging. 2008;33:407–316.

23. Furukawa A, Saotome T, Yamasaki M, et al. Cross-sectional imagingin Crohn disease. Radiographics. 2004;24:689–702.

24. Florie J, Horsthuis K, Hommes DW, et al. Magnetic resonanceimaging compared with ileocolonoscopy in evaluating diseaseseverity in Crohn’s disease. Clin Gastroenterol Hepatol. 2005;3:1221–1228.

25. Lawrance IC, Welman CJ, Shipman P, et al. Correlation of MRI-determinedsmall bowel Crohn’s disease categories with medical response and surgicalpathology.World J Gastroenterol. 2009;15:3367–3375.

26. Oberhuber G, Stangl P, Schober E, et al. Significant association ofstrictures and internal fistula formation in Crohn’s disease. VirchowsArch. 2000;437:293–297.

27. Sheehan A, Gear M, Shepherd N.Fat wrapping in Crohn’s disease:pathological basis and relevance to surgical practice. Br J Surg. 1992;79:955–958.

28. Negaard A, Paulsen V, Sandvik L, et al. A prospective randomizedcomparison between two MRI studies of the small bowel in Crohn’sdisease, the oral contrast method and MR enteroclysis. Eur Radiol.2007;17:2294–2301.


993

Which magnetic resonance imaging findings accurately evaluate inflammation in small bowel Crohnʼs...

Documents

Transcript of Which magnetic resonance imaging findings accurately evaluate inflammation in small bowel Crohnʼs...