Diffusion-weighted magnetic resonance imaging is effective to
detect ileocolonic ulcerations in Crohn
’s disease
A. Buisson*,†, C. Hordonneau‡, M. Goutte*,†, L. Boyer‡,§, B. Pereira¶& G. Bommelaer*,†
*Gastroenterology department, University Hospital Estaing, Clermont-Ferrand, France.
†Microbes, Intestine, Inflammation
and Susceptibility of the Host, UMR 1071 Inserm/Universite d’Auvergne; USCINRA 2018, Clermont-Ferrand, France.
‡Radiology Department, University
Hospital Estaing, Clermont-Ferrand, France.
§ISIT UMR 6284 CNRS UdA,
Clermont-Ferrand, France.
¶Biostatistics Unit- DRCI, GM
Clermont-Ferrand University and Medical Center, Clermont-Ferrand, France.
Correspondence to:
Dr A. Buisson, Gastroenterology Department, University Hospital Estaing, 1 place Aubrac, Clermont-Ferrand 63100, France.
E-mail: [email protected]
Publication data Submitted 17 March 2015 First decision 13 April 2015 Resubmitted 12 May 2015 Accepted 1 June 2015 EV Pub Online 18 June 2015
This article was accepted for publication after full peer-review.
SUMMARY
Background
Magnetic resonance entero-colonography enables accurate assessment of ileoco-lonic Crohn’s disease, but the need for bowel cleansing and rectal enema limits considerably its use in daily practice.
Aim
We evaluated the accuracy of diffusion-weighted magnetic resonance entero-colonography with neither bowel cleansing nor rectal enema to assess endo-scopic activity.
Methods
Forty-four Crohn’s disease patients underwent prospectively and consecutively diffusion-weighted magnetic resonance entero-colonography [with apparent dif-fusion coefficient (ADC) and Clermont score calculation] and ileocolonoscopy [with Crohn’s Disease Endoscopic Index of Severity (CDEIS) and Simplified
Endoscopic score for Crohn’s Disease (SES-CD) calculation].
Results
Mean ADC was inversely correlated with total CDEIS (q = 0.40; P = 0.0067) and total SES-CD (q = 0.33; P = 0.032). Considering the 194 segments, ADC
was inversely correlated with segmental CDEIS ( 0.48; P< 0.001) and segmental
SES-CD ( 0.44; P < 0.001). ADC values were lower in segments with deep ulcers
(1.30 0.23) or superficial ulcerations (1.75 0.64) than in non-ulcerated
seg-ments (2.15 0.5) (P = 0.001). Using a receiver operating curve, we determined
that segmental ADC <1.42 detected endoscopic deep ulcerations with
sensitiv-ity = 0.91 and specificity = 0.83 (Area under the curve = 0.84; P < 0.001).
Seg-mental ADC <1.88 detected endoscopic superficial ulcerations with
sensitivity = 0.64 and specificity = 0.75. The segmental ADC values decreased
when the ulcerations size increased (P = 0.0001). Clermont score correlated with
ileal CDEIS (0.63; P < 0.05) and ileal SES-CD (0.58; P < 0.05). Clermont score
was higher in ulcerated segments (23.3 8.4) than in non-ulcerated segments
(12.4 10.0) (P = 0.006) and increased with ulcers size (P = 0.012). Clermont
score>18.9 detected ulcerations with sensitivity = 0.79 and specificity = 0.73.
Conclusion
Diffusion-weighted magnetic resonance entero-colonography using apparent diffusion coefficient and Clermont score was effective to indirectly detect endoscopic ulcerations in ileocolonic Crohn’s disease.
INTRODUCTION
Crohn’s disease (CD) is a chronic relapsing and remitting disorder that can involve the entire length
of the digestive tract.1 Even not clearly defined, the
concept of mucosal healing should be the therapeutic goal in CD in the era of biologics, as it is associated
with more favourable outcomes.2–4 Ileocolonoscopy is,
to date, the gold standard to evaluate this endpoint. However, repeating this procedure to monitor the dis-ease is not acceptable by the patients in daily practice and has led physicians to look for alternative non-invasive approaches. Thus, magnetic resonance imag-ing (MRI) has been increasimag-ingly used for the diagno-sis and the monitoring of CD because of its lack of ionizing radiation, and its ability to accurately assess
inflammation in CD.5 The magnetic resonance index
of activity (MaRIA), is highly correlated with the
Cro-hn’s Disease Endoscopic Index of Severity (CDEIS)6, 7
and was able to monitor therapeutic response, and could represent an alternative in the evaluation of
ile-ocolonic CD.8 Unfortunately, the MaRIA was
devel-oped using an MRI protocol including bowel
cleansing [4 L of polyethylen glycol (PEG)], intestinal distension (1.5 L of PEG) and rectal enema during the examination that does not make this procedure any more acceptable than colonoscopy in daily
prac-tice.9 Diffusion-weighted magnetic resonance
entero-colonography (DW-MREC) is a well-tolerated and a non-time consuming tool, performed with no bowel
cleansing and no rectal enema.10–15 We previously
reported a good accuracy for detecting and assessing inflammation in ileal CD using the Clermont score as it is highly correlated with the MaRIA and the
simpli-fied endoscopic score for Crohn’s disease (SES-CD).14–16
We also showed the high performances of apparent
diffusion coefficient (ADC), the quantitative parameter
of DWI, in detecting and assessing colonic
inflamma-tion in CD compared to the MaRIA.15 However, the
performances of ADC in detecting and assessing
colo-nic inflammation in CD compared with endoscopy
remains unknown.
The aims of our study were (i) to evaluate the correla-tion between DW-MREC parameters (i.e. segmental ADC, mean ADC and Clermont score), and endoscopic scores (i.e. CDEIS and SES-CD) in CD; (ii) to assess the performances of DW-MREC parameters in detecting
superficial or deep ulcerations in CD; (iii) to define the
concept of MRI remission based on endoscopic mucosal healing in CD.
METHODS
Ethical considerations
The study was performed in accordance with the Decla-ration of Helsinki, Good Clinical Practice and applicable regulatory requirements. The study was approved by local Ethics Committee (IRB number 00008526 Ref 2015/CE06).
Population studied
We performed an observational study of a single-centre cohort in which standardised evaluation was used by experienced clinicians for all patients. A total of 44 patients from the Clermont-Ferrand IBD Unit with an established diagnosis of CD according to Lennard-Jones
criteria17 were prospectively and consecutively included
between December 2012 and May 2014. All the patients consecutively underwent DW-MREC with no bowel cleansing, with no rectal enema and colonoscopy within
4 weeks (mean interval= 17 11 days), without any
therapeutic intervention during this interval.
Colonos-copy and DW-MREC were performed during a flare or
to rule out active disease. In our centre, the diffusion-weighted sequences are performed routinely during MRI.
Radiologists were blinded from endoscopic findings and
endoscopists were blinded from radiologic findings.
Exclusion criteria were: patients with claustrophobia or other common MRI contraindications such as implanted cardiac device, metallic intraocular foreign body, allergy
to gadolinium or severe renal failure [modification of
diet in renal disease (MDRD)<30 mL/min] and patients
with contraindications to colonoscopy. Endoscopy
The severity and extent of endoscopic lesions were
assessed by CDEIS18 and SES-CD19 as routinely used in
the department. The CDEIS and the SES-CD were applied to each segment (distal ileum, caecum/right colon, transverse colon, left/sigmoid colon and rectum) to obtain a segmental CDEIS and SES-CD. Colonoscopy
was considered the ‘gold standard’ for assessing
ileoco-lonic lesions. Patients followed a bowel cleansing proto-col via oral ingestion of 2000 mL of PEG (Fortrans, Ipsen Pharma, Paris, France) on the evening before examination and 2000 mL the morning of the examina-tion. Endoscopy was performed under anaesthesia with
propofol (PROPOFOL DAKOTA PHARM; Sano
two experienced endoscopists (AB, GB) using column video colonoscopy (QFC L 140; Olympus, Tokyo, Japan). Magnetic resonance imaging examinations
MREC with diffusion-weighted and injected sequences were performed in all patients per standard protocol in our Radiology department. Each examination was inter-preted independently by an experienced radiologist (CH)14, 15 who was blinded from clinical data. On the day of DW-MREC, patients had to have been fasting for at least 4 h before the examination. The MRI imaging examinations with no bowel cleansing and with no rectal enema were performed with a 1.5 Tesla GE Optima MR
450w (General Electric HealthCare, Fairfield, CT, USA)
as previously described.14, 15 To achieve an adequate
dis-tension of the distal ileum, 1000 mL of PEG solution (Fortrans, Ipsen Pharma) was given to all patients, 25 min before MRI in patients with previous small bowel resection or 40 min before MRI in non-operated patients. To reduce bowel peristalsis, 1 mg of glucagon
was administrated intravenously before contrast
sequences. Injected sequences were performed after intravenous administration of 0.2 mL/kg body weight of gadobemate dimeglumine (Multi-Hance, Bracco Imaging, Evry, France). MR protocol was given in supplementary files. The DW sequence used a diffusion factor b fixed at
800 s/mm2in the axial plane as this value represents the
best compromise between signal-to-noise ratio and lesion
detection sensitivity.14, 15, 20 Image analysis was
per-formed using a dedicated post-processing workstation (Advantage Window Work Station, General Electric HealthCare). For segmental analysis, we used a division
into five segments as for endoscopic scores, i.e. CDEIS
and SES-CD. For quantitative assessment, the ADC was calculated for each segment in the area of highest signal intensity in the bowel wall. As previously published, the
definition of this area was based on the judgment of the
radiologist.14–16 Mean ADC was defined as the sum of
each segmental ADC divided by the number of
seg-ments. The Clermont score14–16 was calculated in the
terminal ileum using the following formula:
1.6469 bowel thickness 1.3219 ADC + 5.613 9
oedema + 8.306 9 ulcers + 5.039.
Data managing and statistical analysis
Study data were collected and managed using REDCap electronic data capture tools hosted at Clermont-Ferrand
University Hospital.21 REDCap (Research Electronic
Data Capture) is a secure, web-based application designed to support data capture for research studies,
providing (i) an intuitive interface for validated data entry; (ii) audit trails for tracking data manipulation and export procedures; (iii) automated export procedures for seamless data downloads to common statistical packages and (iv) procedures for importing data from external sources.
Statistical analysis was performed usingSTATAsoftware
(version 13; StataCorp, College Station, TX, USA). The
tests were two-sided, with a Type I error set at a = 0.05.
Baseline characteristics were presented as mean (s.d.) or median [interquartile range] according to statistical distribution (assumption of normality assessed using the
Shapiro–Wilk test) for continuous data and as the
num-ber of patients and associated percentages for categorical parameters. Comparisons of patient’s characteristics between the independent groups were performed using
the v2 or Fisher’s exact tests for categorical variables,
and using Student’s t-test or the Mann–Whitney test for
quantitative parameters (homoscedasticity verified using
Fisher–Snedecor test). The correlation between
quantita-tive parameters was explored using coefficient correla-tions (Pearson or Spearman according to statistical distribution). To study repeated measures for a single patient (due to several segments), random-effects models were performed to take into account between and within subject variability. Diagnosis parameters such as sensitiv-ity, specificsensitiv-ity, negative predictive value and positive pre-dictive value were expressed as values and associated 95% confident interval (95% CI). As proposed by some statisticians, we chose to report all the individual p-val-ues without doing any mathematical correction for
dis-tinct tests comparing groups.22 A particular focus was
given to the magnitude of improvement and to the
clini-cal relevance.23
As no data on the correlation between ADC and endoscopic scores in ileal and colonic segments existed, the sample size calculation was performed based on the work from Oussalah et al. which included 44 CD
patients.13
RESULTS
Baseline characteristics of the patients
A total of 44 CD patients undergoing DW-MREC and ileocolonoscopy were analysed in this study. Baseline characteristics of the patients are given in Table 1. Per patients analysis
Among the 44 CD patients, mean ADC (sum of each segmental ADC divided by the number of segments) was
inversely correlated with total CDEIS (q = 0.40;
P = 0.0067) and total SES-CD (q = 0.33; P = 0.032).
Total MaRIA was correlated with total CDEIS (q = 0.47;
P < 0.05) and total SES-CD (q = 0.45; P < 0.05). Mean
ADC correlated with total MaRIA (q = 0.65;
P < 0.05).
Per segments analysis
Overall 194 segments, including 36 ileal segments and 158 colorectal segments were analysed in both
DW-MREC and endoscopy. Distribution of endoscopic lesions is given in Table 2.
Correlation between MRI parameters and endoscopic scores. Considering these 194 segments, segmental ADC
was inversely correlated with segmental CDEIS
(q = 0.48; P < 0.001) and segmental SES-CD
(q = 0.44; P < 0.001). Segmental MaRIA was
corre-lated with segmental CDEIS (q = 0.48; P < 0.05) and segmental SES-CD (q = 0.45; P < 0.05).
Regarding colorectal segments (n = 158), segmental
ADC was inversely correlated with segmental CDEIS
(q = 0.41; P < 0.05) and segmental SES-CD
(q = 0.36; P < 0.05). Segmental MaRIA was correlated
with segmental CDEIS (q = 0.40; P < 0.05) and segmen-tal SES-CD (q = 0.37; P < 0.05).
Regarding ileal segments (n= 36), segmental ADC
was inversely correlated with segmental CDEIS
(q = 0.56; P < 0.05) and segmental SES-CD
(q = 0.55; P < 0.05). Clermont score was correlated
with ileal CDEIS (q = 0.63; P < 0.05) and ileal SES-CD (q = 0.58; P < 0.05). Segmental MaRIA was correlated with segmental CDEIS (q = 0.67; P < 0.05) and
segmen-tal SES-CD (q = 0.61; P < 0.05).
MRI accuracy for the detection of endoscopic ulcerations in CD. Using endoscopy as reference, MRI directly
visu-alised ulcerations with a poor sensitivity (Se= 0.29) but
high specificity (Spe = 0.92).
We assessed the correlation between the ADC values and the ulcerations depth. We found that ADC values
were lower in the segments with deep ulcers
(1.30 0.23) or superficial ulcerations (1.75 0.64)
than in segments with no ulcers (2.15 0.5)
(P= 0.001) (Figure 1). Using a ROC curve, we
deter-mined that segmental ADC< 1.42 detected endoscopic
deep ulcers with Se = 0.91, and Spe = 0.83(AUC = 0.84;
P < 0.001) (Figure 2a and Table 3). In addition,
seg-mental ADC<1.88 detected endoscopic superficial
ulcer-ations with Se = 0.64, Spe = 0.75 (Figure 2b and
Tables 2 and 3). We also analysed the relationship between segmental ADC and the ulcerations size. We showed that the segmental ADC values decreased when the ulcerations size increased. We found that the ADC
values were 2.17 0.5 for segments with no ulcers,
2.09 0.04 for segments with ulcers smaller than
5 mm, 1.79 0.65 for segments ranging from 6 to
20 mm and 1.50 0.53 for segments with ulcerations
greater than 20 mm (1.50 0.53) (P = 0.0001)
(Fig-ure 3). Table 1 | Characteristics of the Crohn’s disease
patients (n = 44) included in the study
Age at diagnosis, mean (s.d.) 27.9 (15.0)
Disease duration at inclusion, median [IQR] 5.5 [0.4–14.2] Male gender,n (%) 21 (47.7) Tobacco use,n (%) Nonsmokers 25 (56.8) Former smokers 7 (15.9) Active smokers 12 (27.3)
Previous intestinal resection,n (%) 9 (20.5)
Montreal classification
Disease location L1,n (%) 7 (15.9) L2,n (%) 10 (22.7) L3,n (%) 27 (61.4) L4,n (%) 4 (9.1) Behaviour B1,n (%) 22 (50.0) B2,n (%) 12 (27.3) B3,n (%) 9 (20.5) Perianal lesions,n (%) 10 (22.7) Concomittant therapies 5-ASA,n (%) 2 (4.6) Corticosteroids,n (%) 13 (29.6) Budesonide,n (%) 4 (9.1) Thiopurines,n (%) 5 (11.4) Infliximab, n (%) 4 (9.1) Adalimumab,n (%) 4 (9.1) Ustekinumab,n (%) 1 (2.3)
CDAI, median [IQR] 188 [106–266]
Total CDEIS, median [IQR] 4.8 [1.8–9.3]
Total SES-CD, median [IQR] 7 [4–13]
CRP, median [IQR] 11.4 [3.5–38.8]
Faecal Calprotectin, median [IQR] 1555 [384–1800]
Mean ADC, median [IQR] 2.04 [1.78–2.25]
Total MaRIA, median [IQR] 39.3 [23.4–44.2]
s.d., standard deviation; IQR, interquartile; n, number; CDAI,
Crohn’s disease activity index; CDEIS, Crohn’s disease
endo-scopic index of severity; SES-CD, simplified endoscopic score
for Crohn’s disease; CRP, C-reactive protein; ADC, apparent
diffusion coefficient; MaRIA, magnetic resonance index of
Considering ileal segments (n= 37), the Clermont
score was higher in the ulcerated segments (23.3 8.4)
than in the segments with no ulcers (12.4 10.0)
(P= 0.006). Clermont score >18.9 detected ulcerations
with Se= 0.79 and Spe = 0.73 (Table 3). The Clermont
Score increased with the ulcers size. We found that the
Clermont score was 11.3 10 for segments with no
ulcers or ulcerations smaller than 5 mm, 20.6 9.2 for
Table 2 | Distribution of endoscopic ulcerations and accuracy of the quantitative parameters of diffusion-weighted magnetic resonance entero-colonography (i.e. Clermont score for the ileum and ADC for the colon/rectum) to detect endoscopic ulcerations in Crohn’s disease
Deep ulcerations
Superficial
ulcerations Any ulcerations
MRI accuracy to detect endoscopic
ulcerations (Clermont score>18.9
for ileum and ADC<1.88 for
colon/rectum) Ileum,n 4/36, 11.1% 13/36, 36.1% 14/36, 38.9% 27/36, 75.0% Caecum/right colon,n 3/38, 7.9% 13/38, 34.2% 14/38, 36.8% 24/38, 63.2% Transverse colon,n 2/39, 5.1% 10/39, 25.6% 10/39, 25.6% 32/39, 82.1% Left/Sigmoid colon,n 2/39, 5.1% 9/39, 23.1% 9/39, 23.1% 33/39, 84.6% Rectum,n 0/42, 0.0% 10/42, 23.8% 10/42, 23.8% 34/42, 81.0%
All ileocolonic segments,n 8/44, 18.2% 28/44, 63.6% 28/44, 63.6% 150/194, 77.3%
n, number of patients; ADC, apparent diffusion coefficient; endoscopic remission is defined as the absence of superficial or deep
ulcerations, and the absence of endoscopic stenosis; Clermont score= 1.646 9 bowel thickness 1.3219 ADC + 5.613 9 1 (if
oedema)+ 8.306 9 1 (if ulcers) + 5.039).
4.5 Median Mean 4 3.5 3 2.5
Apparent diffusion coefficient (ADC) en mm
2/8 2 1.5 1 0.5 0 No ulceration Superficial ulcerations Deep ulcerations Endoscopic ulcerations depth
Figure 1 | Relationship between the values of apparent diffusion coefficient (ADC) retrieved from diffusion-weighted resonance magnetic entero-colonography and endoscopic ulcerations depth in Crohn’s disease.
0.00 0.00 0.25 0.50 Sensitivity 0.75 1.00 0.25 0.50 1 - Specificity Area under ROC curve = 0.8810
0.75 1.00 0.00 0.00 0.25 0.50 Sensitivity 0.75 1.00 0.25 0.50 1 - Specificity Area under ROC curve = 0.6812
0.75 1.00
(a)
(b)
Figure 2 | Receiver operating curve showing the performances of apparent diffusion coefficient (ADC) retrieved from diffusion-weighted resonance magnetic entero-colonography, to detect deep (a) or superficial (b) ileocolonic endoscopic ulcerations in Crohn’s disease.
segments ranging from 6 to 20 mm and 27.5 1.5 for
segments with ulcerations greater than 20 mm
(1.50 0.53) (P = 0.012; Figure 4).
MRI accuracy for the definition of endoscopic remission
in CD: the diffusion-weighted MRI remission. Among the 44 CD patients, while 10 (22.7%) were in endoscopic remission (defined as the absence of superficial or deep ulcerations, and the absence of endoscopic stenosis), eight (18.2%) were in DW-MRI remission defined as no
segmental ADC<1.88, and no Clermont score ≥18.9 and
no stenosis on MRI. The DW-MRI remission detected
endoscopic remission with Se, Spe, NPV and PPV of 50.0%, 90.9%, 85.7% and 62.5%.
Among the three false positives, i.e. the patients con-sidered in MRI remission with persistent endoscopic ulcerations, no patients presented with deep ulcerations. The ulcerations size was smaller than 2 cm in each patient. The total CDEIS for these three patients was 1.2, 8.0 and 11.8.
DISCUSSION
DW Imaging is a method, deriving its image contrast from differences in the motion of water molecules Table 3 | Performances of apparent coefficient diffusion (ADC) and Clermont score calculated from
diffusion-weighted magnetic resonance entero-colonography for the detection superficial or deep ulcerations in Crohn’s disease
Sensitivity Specificity NPV PPV
ADC < 1.42 for the detection
of deep ulcerations
90.9 [58.7–99.8] 82.7 [76.6–87.7] 99.4 [96.6–100.0] 22.7 [11.5–37.8]
ADC < 1.88 for the detection
of superficial ulcerations
63.6 [49.6–76.2] 75.0 [67.3–81.7] 85.1 [77.9–90.6] 47.9 [36.1–60.0]
Clermont score> 18.9 for the
detection of ulcerations
76.9 [46.2–95.0] 71.0 [52.0–85.8] 88.0 [68.8–97.5] 52.6 [28.9–75.6]
NPV, negative predictive value; PPV, positive predictive value; ADC, apparent diffusion coefficient.
ADC is expressed in910 3mm2/s sensitivity, specificity, NPV and PPV are expressed in value and associated 95% confidence
interval. Clermont score (=1.646 9 bowel thickness 1.3219 ADC + 5.613 9 oedema + 8.306 9 ulcers + 5.039) is dedicated
to the terminal ileum.
4.5 Median Mean 4 3.5 3 2.5 2 1.5 1 0.5 0 No ulceration <0.5 cm 0.5–2 cm >2 cm Endoscopic ulcerations size
Apparent diffusion coefficient (ADC) en mm
2/8
Figure 3 | Relationship between the values of apparent diffusion coefficient (ADC) retrieved from diffusion-weighted resonance magnetic entero-colonography and endoscopic ulcerations size in Crohn’s disease.
40 35 30
Clermont score
(= 1.646×bowel thickness–1.321×ADC+ 5.613×edema +8.306×ulceration+5.039)
25 20 15 10 5 0 <0.5 cm 0.5–2 cm >2 cm
Endoscopic ulcerations size
Figure 4 | Relationship between the values of Clermont score dedicated to the terminal ileum (=1.646 9 bowel thickness 1.3219 ADC + 5.613 9 1 (if
oedema)+ 8.306 9 1 (if ulcers) + 5.039) and endoscopic ulcerations size in Crohn’s disease.
between tissues, used in different topics especially neuro-vascular diseases and oncology. Recently, our team and
others confirmed that DWI-MREC with no bowel
cleansing and no rectal enema was a well-tolerated, nontime-consuming, accurate and reproducible tool for detecting and assessing inflammation in ileal and colonic CD.11–16, 20, 24, 25 The Clermont score (= 1.646 9 bowel
thickness – 1.321 9 ADC + 5.613 9 oedema + 8.306 9
ulcers+ 5.039) dedicated to the ileum is an accurate tool
to assess inflammatory activity in the terminal ileum,
and is highly correlated with the MaRIA14, 15 and the
Simplified Endoscopic Score for Crohn’s Disease (SES-CD).16, 19 The Clermont score built and validated
espe-cially for the ileum14–16 offers several advantages
com-pared to the MaRIA coming from the different MRI protocols used to calculate these two scores. The MRI protocol used to calculate Clermont score did not need bowel cleansing and rectal enema and could avoid peripheral vein installation and its potential complica-tions. In addition, avoiding gadolinium injection could
prevent nephrogenic systemic fibrosis, a rare but serious
complication, and could also shorten the MRI duration. All of these former points lead to highly improve the MRI tolerability in limiting the burden felt by the patients. Regarding the performances of these two scores, we previously reported that ADC calculation was more reproducible than related contrast enhancement (Lin’s
concordance coefficient 0.96 vs. 0.83).15We reported also
a high accuracy of the ADC, the quantitative parameter
of DW Imaging, to detect and assess inflammatory
activ-ity in colonic CD compared to the MaRIA.15Few studies
previously compared MRI with DWI sequences to
ileo-colonoscopy.12, 15, 19, 25 The authors focused on DWI
hyperintensity as marker of inflammation and found similar results with gadolinium contrast enhancement in
detecting inflammation. To our knowledge, the present
study is the first comparing ADC, the quantitative
parameter of DWI and endoscopy in both ileal and colo-nic CD.
In our study, we observed a moderate correlation
between total MaRIA and the CDEIS (q = 0.47;
P < 0.05) or the SES-CD (q = 0.45; P < 0.05). Data
from the tertiary centre of Barcelona, Spain, reported
higher correlation ratio for the CDEIS reaching 0.83.7, 8
This discrepancy could be explained by the addition of bowel cleansing and colonic distension (rectal enema)
which could improve the detection of colonic lesions.26
In addition, a recent study reported that 3T MRI could have higher sensitivity than 1.5T MRI to directly
visual-ize ulcers in CD.27 We reported that mean ADC (=sum
of each segmental ADC divided by the number of seg-ments) was correlated with total MaRIA and moderately
correlated with CDEIS (q = 0.40; P = 0.0067) and total
SES-CD (q = 0.33; P = 0.032).
Considering the per segment analysis, ADC, the quan-titative parameter of DW-MREC was inversely correlated
with segmental CDEIS (q = 0.48; P < 0.001) and
seg-mental SES-CD (q = 0.44; P < 0.001). This correlation
was moderate in colonic segments and slightly higher in ileal segments. We observed the same results with MaRIA which moderately correlated with segmental CDEIS (q = 0.48; P < 0.05) and segmental SES-CD
(q = 0.45; P < 0.05). This discrepancy could also be
explained by the use of bowel cleansing and colonic
dis-tension (rectal enema) and the use of 3T MRI.27 The
impact of switching from 1.5T to 3T on diffusion-weighted parameters remains unexplored and should be investigated in the future. In ileal segments, the
Cler-mont score was correlated with CDEIS (q = 0.63;
P < 0.05) and ileal SES-CD (q = 0.58; P < 0.05). It was
in line with a previous work which showed a good corre-lation between SES-CD and Clermont score in ileal CD
(r= 0.78; P < 0.0001).16
These correlation studies, reporting disappointing
results, confirmed that MR enterocolonography including
DW-MREC did not provide the same information as endoscopy. While endoscopy detects mucosal lesions (erythema, oedema, apthoid erosions, superficial or deep ulcerations), MRI detects transmural lesions (oedema,
bowel wall thickening, bowel wall inflammation using
DWI or contrast enhancement) including digestive
dam-age (stenosis and fistula). The two available endoscopic
scores focused on lesions severity and lesions extent (ulcerated or not). In contrast, MRI parameters focused only on the severity of the lesions. The IBD physicians should be aware that MRI is not as accurate as endos-copy to detect mucosal abnormalities and should be interpreted differently. The disappearance or at least a clear decrease in ulcerations is probably to date the most
useful definition of mucosal healing.28 Owing to these
remarks, we investigated the accuracy of DW-MREC parameters, i.e. ADC and Clermont score to detect ulcer-ations which are the most important lesions to target in the era of biologics. As expected we found that DW-MREC with no bowel cleansing and no bowel enema is not a good tool to directly visualize ulcers especially in colonic segments. The sensitivity was poor (0.29) while
we reported a very high specificity (0.92). We previously
reported that the inter-observer agreement to calculate ADC was very high in ileal and colonic CD (Lin’s
con-cordance coefficient = 0.96).15 The additional value of ADC calculation is clearly demonstrated in our cohort.
ADC was highly effective in detecting deep (Se= 0.91
and Spe= 0.83) or superficial ulcerations (Se = 0.64 and
Spe = 0.75). We reported also a very high NPV (0.99
and 0.85, respectively), confirming the potential role of DW-MREC parameters in detecting healed segments
(defined as not ulcerated). In contrast, we found a high
number of false positives. Some authors reported false
positives rate as a weakness of DWI hyperintensity.29
However, we considered that the high false positives ratio could be partly explained by the shorter experience in the evaluation of these DWI sequences. However, even using ADC calculation by an experienced
radiolo-gist in this field, IBD physicians could be kept in mind
the rate of false positives when they used DW-MREC in daily practice. In addition, the ADC is linked to the depth and the size of the ulcerations. Considering ileal CD, the Clermont score was also highly effective in
detecting the presence of ulcerations (Se= 0.79 and
Spe = 0.73). These results underlined the very promising
potential for DW-MREC parameters, i.e. ADC and Cler-mont score to indirectly detect ulcerations in ileal and colonic segments.
We also attempt to define, for the first time, the
con-cept of DW-MRI remission using DW-MREC parame-ters. We chose a definition based on the absence of the main ulcerations in each digestive segment. This concept is original as endoscopic remission or response was
mostly defined using global scores of severity.18, 19, 28
However, we considered that these scores failed to accu-rately assess CD involving only one segment especially in pure ileal disease. We reported a very high positive predictive value and specificity for MRI remission in determining patients with no ulcerations. It means that in using MRI remission we did not overtreat any patients. Although this concept of DW-MRI remission seems very promising, this preliminary definition should be taken with caution and needs further validation in an independent larger cohort.
The main strengths of this study were the prospective design, including patients from the Clermont-Ferrand IBD centre with experienced radiologists in the DW-MREC use, and the choice of endoscopy as gold stan-dard. In addition, the analysis included almost 200 diges-tive segments which is approximately the same as the largest cohort reported so far on 40 CD patients and 211
segments.13 Regarding the study limitations, the addition
of the MRI protocol previously described by Rimola
et al.6, 7 in this present study could have been
interest-ing. The addition of a questionnaire to compare endos-copy and MRI acceptability will be of great value. A
previous study9 reported that endoscopy was more
acceptable than MRI requiring rectal enema. The place-ment of the rectal tube was rated as the most unpleasant
part (7.3 2.1) on a 10-point scale (1, good, 10, poor
acceptance) compared with endoscope placement
(3.8 2.3) (P < 0.001).9
In conclusion, we reported for the first time that
DW-MREC, with no bowel cleansing and no rectal enema, is effective to indirectly detect endoscopic ulcer-ations both in ileal and colonic CD using ADC and Clermont score. DW-MREC is then a well-tolerated and nontime consuming tool, which could be used and repeated both in daily practice and clinical trials to assess inflammatory activity in ileal and colonic CD. The sensitivity of DW-MREC parameters to medical therapy remained to be explored. Our group is cur-rently conducting a multicentre study to address this question.
AUTHORSHIP
Guarantor of the article: Anthony Buisson.
Author contributions: A Buisson: study concept and design; acquisition of data; analysis and interpretation of data; drafting of the manuscript. C Hordonneau: study concept and design; acquisition of data; analysis and interpretation of data. M Goutte: study concept and design; acquisition of data; analysis and interpretation of data; critical revision of the manuscript for important intellectual content. L Boyer: critical revision of the man-uscript for important intellectual content. B Pereira: sta-tistical analysis. G Bommelaer: study concept and design; critical revision of the manuscript for important intellec-tual content.
All authors approved the final version of the article,
including the authorship list.
ACKNOWLEDGEMENTS
Declaration of personal interests: Anthony Buisson has served as a speaker or a consultant for Abbvie, MSD, Takeda and Ferring. The other authors declare no perso-nal interests.
Declaration of funding interests: None.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online version of this article:
Table S1. Characteristics description of MRI sequences used in this study protocol.
REFERENCES
1. Peyrin-Biroulet L, Loftus EV Jr, Colombel J-F, et al. The natural history of adult Crohn’s disease in population-based cohorts. Am J Gastroenterol 2010; 105: 289–97.
2. Baert F, Moortgat L, Van Assche G, et al. Mucosal healing predicts sustained clinical remission in patients with early-stage Crohn’s disease. Gastroenterology 2010;138: 463–8. 3. Schnitzler F, Fidder H, Ferrante M,
et al. Mucosal healing predicts long-term outcome of maintenance therapy with infliximab in Crohn’s disease 2009; 15: 1295–301.
4. Frøslie KF, Jahnsen J, Moum BA, et al. Mucosal healing in inflammatory bowel disease: results from a Norwegian population-based cohort.
Gastroenterology 2007;133: 412–22. 5. Panes J, Bouhnik Y, Reinisch W, et al.
Imaging techniques for assessment of inflammatory bowel disease: joint ECCO and ESGAR evidence-based consensus guidelines. J Crohns Colitis 2013;7: 556–85.
6. Rimola J, Rodriguez S, Garcıa-Bosch O, et al. Magnetic resonance for
assessment of disease activity and severity in ileocolonic Crohn’s disease. Gut 2009;58: 1113–20.
7. Rimola J, Ordas I, Rodriguez S, et al. Magnetic resonance imaging for evaluation of Crohn’s disease: validation of parameters of severity and
quantitative index of activity. Inflamm Bowel Dis 2011;17: 1759–68. 8. Ordas I, Rimola J, Rodrıguez S, et al.
Accuracy of magnetic resonance enterography in assessing response to therapy and mucosal healing in patients with Crohn’s disease. Gastroenterology 2014;146: 374–82.
9. Langhorst J, K€uhle CA, Ajaj W, et al. MR colonography without bowel purgation for the assessment of inflammatory bowel diseases: diagnostic accuracy and patient acceptance. Inflamm Bowel Dis 2007; 13: 1001–8.
10. Kiryu S, Watanabe M, Kabasawa H, et al. Evaluation of super paramagnetic iron oxide-enhanced diffusion-weighted PROPELLER T2-fast spin echo magnetic resonance imaging:
Preliminary experience. J Comput Assist Tomogr 2006;30: 197–200.
11. Kiryu S, Dodanuki K, Takao H, et al. Free-breathing diffusion-weighted imaging for the assessment of inflammatory activity in Crohn’s disease. J Magn Reson Imaging JMRI 2009;29: 880–6.
12. Oto A, Zhu F, Kulkarni K, et al. Evaluation of diffusion-weighted MR imaging for detection of bowel inflammation in patients with Crohn’s disease. Acad Radiol 2009;16: 597–603. 13. Oussalah A, Laurent V, Bruot O, et al.
Diffusion-weighted magnetic resonance without bowel preparation for detecting colonic inflammation in inflammatory bowel disease. Gut 2010;59: 1056–65. 14. Buisson A, Joubert A, Montoriol P-F,
et al. Diffusion-weighted magnetic resonance imaging for detecting and assessing ileal inflammation in Crohn’s disease. Aliment Pharmacol Ther 2013; 37: 537–45.
15. Hordonneau C, Buisson A, Scanzi J, et al. Diffusion-weighted magnetic resonance imaging in ileocolonic Crohn’s disease: validation of quantitative index of activity. Am J Gastroenterol 2014;109: 89–98. 16. Caruso A, DʼInca R, Scarpa M, et al.
Diffusion-weighted magnetic resonance for assessing ileal Crohn’s disease activity. Inflamm Bowel Dis 2014; 20: 1575–83.
17. Lennard-Jones JE. Classification of inflammatory bowel disease. Scand J Gastroenterol Suppl 1989;170: 2–6; discussion 16–19.
18. Mary JY, Modigliani R. Development and validation of an endoscopic index of the severity for Crohn’s disease: a prospective multicentre study. Groupe d’Etudes Therapeutiques des Affections Inflammatoires du Tube Digestif (GETAID). Gut 1989;30: 983–9. 19. Daperno M, D’Haens G, Van Assche G,
et al. Development and validation of a new, simplified endoscopic activity score for Crohn’s disease: the SES-CD. Gastrointest Endosc 2004;60: 505–12. 20. Sato H, Tamura C, Narimatsu K, et al.
Magnetic resonance enterocolonography is highly detectable of erosion and redness in intestinal mucosa of patients
with Crohn’s disease. J Gastroenterol Hepatol 2015;30: 667–73.
21. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009;42: 377–81.
22. Rothman KJ. No adjustments are needed for multiple comparisons. Epidemiol Camb Mass 1990;1: 43–6. 23. Feise RJ. Do multiple outcome
measures require p-value adjustment? BMC Med Res Methodol 2002;2: 8. 24. Oto A, Kayhan A, Williams JTB, et al.
Active Crohn’s disease in the small bowel: evaluation by diffusion weighted imaging and quantitative dynamic contrast enhanced MR imaging. J Magn Reson Imaging JMRI 2011;33: 615–24. 25. Sakuraba H, Ishiguro Y, Hasui K, et al.
Prediction of maintained mucosal healing in patients with Crohn’s disease under treatment with infliximab using diffusion-weighted magnetic resonance imaging. Digestion 2014;89: 49–54. 26. Friedrich C, Fajfar A, Pawlik M, et al.
Magnetic resonance enterography with and without biphasic contrast agent enema compared to conventional ileocolonoscopy in patients with Crohn’s disease. Inflamm Bowel Dis 2012;18: 1842–8.
27. Fiorino G, Bonifacio C, Padrenostro M, et al. Comparison between 1.5 and 3.0 Tesla magnetic resonance enterography for the assessment of disease activity and complications in ileo-colonic Crohn’s disease. Dig Dis Sci 2013; 58: 3246–55.
28. Ferrante M, Colombel J-F, Sandborn WJ, et al. Validation of endoscopic activity scores in patients with Crohn’s disease based on a post hoc analysis of data from SONIC. Gastroenterology 2013;145: 978–86.
29. Kim K-J, Lee Y, Park SH, et al. Diffusion-weighted MR enterography for evaluating crohn’s disease: how does it add diagnostically to conventional MR enterography? Inflamm Bowel Dis 2015;21: 101–9.
