Insights Imaging (2018) 9:47–58

https://doi.org/10.1007/s13244-017-0579-9

REVIEW

Role of magnetic resonance imaging in the management

of perianal Crohn’s disease
Jose C. Gallego 1 & Ana Echarri 2

Received: 30 May 2017 / Revised: 5 September 2017 / Accepted: 26 September 2017 / Published online: 15 November 2017
# The Author(s) 2017. This article is an open access publication

Abstract
Perianal fistulas are a major problem in many patients with Teaching points
Crohn’s disease. These are usually complex fistulas that ad- • MRI is the method of choice with which to evaluate perianal
versely affect patients’ quality of life, and their clinical man- fistulas.
agement is difficult. Medical treatment sometimes achieves • In perianal Crohn’s disease, MRI is a fundamental means of
cessation of discharge and closure of the external opening; patient monitoring.
however, it is difficult to assess the status of the rest of the • The usefulness of the Van Assche score for patient monitor-
fistula tract. Magnetic resonance imaging is the method of ing remains unclear.
choice with which to evaluate the condition of perianal fistulas • New MRI sequences' diffusion, perfusion, and magnetisation
and allows for assessment of the status of inaccessible areas. transfer may serve as biomarkers.
Magnetic resonance imaging also allows the clinician to eval-
uate other perianal manifestations of Crohn’s disease that dif- Keywords Crohn’s disease . Magnetic resonance imaging .
fer from the fistulas. This imaging technique is therefore a Perianal disorders . Diffusion magnetic resonance . Perfusion
fundamental means of patient monitoring. When used in con- imaging
junction with assessment of the patient’s morphological find-
ings, it provides information that allows for both quantifica-
tion of disease severity and evaluation of the response to treat-
ment. New types of magnetic resonance sequences are emerg- Introduction
ing, such as diffusion, perfusion, and magnetisation transfer.
These sequences may serve as biomarkers because they pro- Perianal fistulas are frequent manifestations of Crohn’s dis-
vide information reflecting the changes taking place at the ease (CD) and cause significant morbidity, often with serious
molecular level. This will help to shape a new scenario in impairment of the patient’s abilities. Fistulas are abnormal
the early assessment of the response to treatments such as communications formed by granulation tissue and are
anti-tumour necrosis factor drugs. established between two epithelial surfaces. Perianal fistulas
form between the inner surface of the anal canal (internal
opening) and the skin (external opening). These fistulas are
the main, but not the only, manifestation of perianal CD
* Jose C. Gallego (PCD). Their treatment is difficult and sometimes requires
[email protected] colostomy or even proctectomy, although this does not resolve
the problem in all patients.
1
Department of Radiology, Complexo Hospitalario Universitario de Although CD-associated perianal fistulas only account for
Ferrol, Av. da Residencia s/n, 15405 Ferrol, Spain around 1% of all perianal fistulas [1], they appear in 30% to
2
Department of Gastroenterology, Complexo Hospitalario 50% of patients with CD and can be the first manifestation of
Universitario de Ferrol, Ferrol, Spain CD as well as precede the finding of intestinal lesions in up to
48 Insights Imaging (2018) 9:47–58

30% of patients [2]. Additionally, the presence of such fistulas specificity of MRI is substantially higher [9]. Transperineal
is associated with the most severe forms of CD [3]. US is also a very useful technique for the study of perianal
Perianal fistulas seem to be caused by different pathogenet- fistulas. This technique is more comfortable and less invasive
ic mechanisms than enteric fistulas. In fact, up to 80% of than endoanal sonography. However, although it was first de-
patients with PCD do not have enteric fistulas [4]. scribed some years ago, it is only performed in a few centres.
Therefore, PCD is considered a different phenotype by the Bor et al. [10] stated that for patients with PCD, the accuracy
Montreal Classification, in which it appears as a differentiated of transperineal US is identical to that of MRI or endoanal
subclassification [5]. It is possible that these lesions result sonography.
from the deepening of distal rectal ulcers or fissures secondary Several scientific societies and groups of experts are clearly
to the forces exerted during defecation. It also seems clear that committed to MRI for the study of PCD. The European
microbiological, immune, and genetic factors are involved in Society of Crohn’s and Colitis (ECCO) [11] stated in 2010
the aetiology of perianal fistulas [6]. regarding the diagnosis:
Imaging techniques provide information on the anatomical
and functional aspects of fistulas, which are often difficult to – 9A. Pelvic MRI should be the initial procedure because it
obtain through clinical examination, especially in patients is accurate and non-invasive, although it is not needed
with inflammation or abscesses. In addition, magnetic reso- routinely in simple fistulae [EL2b., RG B].
nance imaging (MRI) makes it possible to evaluate the pres- – 9C. Anorectal ultrasound requires expertise, but can be
ence or absence of healing of the deep tissues of the fistula equivalent to pelvic MRI in completing examination un-
paths, a key aspect in the follow-up of patients. der anaesthesia if rectal stenosis has been excluded [EL5,
RG D]. Fistulography is not recommended [EL3, RG C].

Imaging studies The ECCO-European Society of Gastrointestinal and

Abdominal Radiology (ESGAR) Consensus [12] subsequent-
MRI and ultrasound (US), both endoanal and transperineal, ly reaffirmed MRI as the most accurate imaging technique
are imaging methods that show the anal anatomy in detail. with which to study perianal fistulas, recognising that it also
They allow the clinician to assess the extent and complexity allows for the detection of many lesions that go unnoticed
of disease as well as the presence of abscesses; such informa- during an examination under anaesthesia and emphasising
tion is essential in choosing the most appropriate treatment. its role as a method for assessing the response to treatment:
The higher anatomical resolution of these methods also con-
tributes to successful surgical outcomes by decreasing the risk – ECCO-ESGAR statement 5A. MRI is the most accurate
of incontinence or other complications. Unfortunately, this imaging test for perianal CD with accuracy surpassing
grade of precision cannot be achieved by other imaging examination under anaesthesia, and is recommended
methods such as fistulography or computed tomography. during the initial diagnosis unless there is a need for
Importantly, correct evaluation of patients with PCD intermediate drainage of sepsis [EL 1].
should include an endoscopic study to evaluate luminal dis- – ECCO-ESGAR statement 5B. Undetected or untreated
ease, specifically the condition of the rectum; a local imaging fistulae extensions and abscesses are the major cause of
study, either MRI or US; and an examination under anaesthe- treatment failure. Imaging, particularly using MRI, is
sia, during which abscesses can be drained, seton stitches can highly accurate in detecting such complications and for
be placed in situ, and other necessary measures can be under- treatment planning [EL 2].
taken [7]. With respect to imaging, MRI is a more objective – ECCO-ESGAR statement 5C. MRI and endosonography
method than US, is more easily understood by are both superior to simple clinical evaluation at
nonradiologists, and allows for a more accurate comparison assessing treatment response, particularly for detecting
of the follow-up studies of each patient. In 2008, Sahni et al. residual abscesses, and either should be considered prior
[8] published a comprehensive review using methods of to significant changes in, or cessation of, surgical or
evidence-based medicine, i.e., consultation of guides from medical therapy [EL 2].
the American College of Radiologists and the American
Gastroenterological Association, a review of the literature, The Shanghai Group, a group of experts comprising mem-
and consultation with experts. The authors concluded that bers of the World Gastroenterology Organisation,
MRI is the best method with which to distinguish simple from International Organisation for Inflammatory Bowel Diseases,
complex fistulas, surpassing both endoanal US and physical European Society of Coloproctology, and Robarts Clinical
examination. Conversely, a meta-analysis comparing MRI Trials, stated in 2014 that US must be supplemented with
and endoanal US for detection of perianal fistulas concluded MRI in many cases and that they consider MRI the gold stan-
that the sensitivity of both methods is similar, although the dard imaging technique [13]:
Insights Imaging (2018) 9:47–58 49

Statement 8.1. Pelvic MRI is a highly accurate non- & Grade 2: Intersphincteric fistula with abscess or sec-
invasive modality for the diagnosis and classification ondary tract. Although complications occur, the dis-
of perianal fistulas; therefore it is considered the gold ease never surpasses the outer sphincter. The
standard imaging technique for perianal CD. MRI pro- branching or secondary tract can surpass the midline
vides additional detailed information on luminal disease and show contralateral extension by adopting a
location, disease severity, and fluid collections. Bhorseshoe^ appearance.
& Grade 3: Trans-sphincteric fistula. The fistula tract passes
More detailed information on all recommendations of various through both sphincters and extends toward the skin
consensus groups is provided in a review by de Groove et al. through the ischioanal fossa.
[14]. Notably, these authors stated that the use of MRI has & Grade 4: Trans-sphincteric fistula with abscesses or sec-
reached a generalised consensus. ondary tracts in the ischioanal fossa. The tract shows ab-
However, pelvic MRI also has its limitations. One of most scess formation, generally in the ischioanal fossa although
important of its disadvantages is the difficulty that is often sometimes also in the intersphincteric region, by adopting
encountered in identifying the internal opening of the fistula. an Bhourglass^ shape.
Other disadvantages include its high cost and contraindication & Grade 5: Fistulas that extend over the levator ani.
in patients with pacemakers or other devices, claustrophobic Suprasphincteric fistulas run through the intersphincteric
patients, and patients with various other conditions. space to the highest point of the levator ani muscle, then
pass through it, extending to the skin through the
ischioanal fossa. Extrasphicteric fistulas originate from a
MRI findings pelvic organ, usually the rectum, and pass through the
levator ani toward the skin, also through the ischioanal
Generally, the correct identification of perianal fistulas is fossa. In any of these situations, contralateral extension
a complex process that requires an experienced radiologist may appear.
[15]. The technical aspects of pelvic MRI when per-
formed for evaluation of perianal fistulas have been wide- Fistulas in patients with PCD are almost always
ly published [16, 17]. It is important to obtain high- Bcomplex.^ According to the American Gastroenterological
resolution images, tilt the oblique-axial and oblique- Association [19], complex fistulas are defined as high
coronal planes according to the anal canal, and perform fistulas, intersphincteric and trans-sphincteric fistulas,
a T2-weighted sequence with a fat-suppression technique those that cross the levator ani muscle, and those with
(e.g., in the axial plane) to enable easier identification of secondary tracts (Fig. 1). These characteristics increase
the fistula tracts and fluid collections. However, consider- the risk of complications. This is especially true for fis-
ing the advances during the last several years, two points tulas with secondary pathways; these tracts, when blind,
could be modified: three-dimensional T2-weighted se- can become complicated and lead to abscesses (Fig. 2).
quences could be performed instead of sagittal, oblique- The most frequent complication is the presence of a
axial, and oblique-coronal high-resolution T2-weighted branch coursing in the cranial direction from the highest
sequences, and diffusion-weighted sequences could be point of a trans-sphincteric path, running toward the roof
added. The first change decreases the number of se- of the ischioanal fossa and even crossing the levator ani
quences and provides data for post-processing reformation muscle. Contralateral tracts to the other ischioanal fossa,
of the images in any desired plane. The usefulness of or Bhorseshoe^ extensions, may also be present on both
diffusion-weighted sequences will be discussed later; sides from the internal opening.
however, we believe that no study protocols should lack Anovaginal fistulas (Fig. 3) also have special charac-
the use of such sequences (Table 1). teristics. Although they account for 10% of all fistulas in
Some additional findings of fistulas that are identifiable on patients with PCD, they cause important problems associ-
MRI, such as the presence of abscesses, branching, and other ated with discomfort, infections, dyspareunia, and other
features, are not included in other classifications. This led to conditions. They also have a worse prognosis because
the development of a radiological classification known as the medical treatments are effective in only a low percentage
St. James Hospital classification [18], which comprises five of cases, the placement of seton stitches does not seem to
grades: be useful, and surgical treatment is associated with a high
rate of recurrence [20]. In MRI, this may be the only
& Grade 1: Simple linear intersphincteric fistula. The tract indication for the use of endoanal coils because the path-
runs between the skin and the anal canal, does not show ways are better demonstrated given their smaller exten-
branching, and does not surpass the outer sphincter or sion (2.0–2.5 cm) and the proximity to the coil [21].
affect the ischioanal fossa. However, the clinician must remember that other perianal
50 Insights Imaging (2018) 9:47–58

Table 1 Possible protocols for MRI acquisition

Sequence FOV Slice thick Slice Acc Fat

(mm) gap factor sat
Philips Siemens GE Toshiba Hitachi

Sagittal T2-weighted TSE TSE FSE FSE FSE 260 × 260 ≤4 0 2 NO

Oblique axial T2-weighted TSE TSE FSE FSE FSE 260 × 260 ≤4 20% 2 NO
Oblique axial T2- weighted SPIR-TSE FS-FSE CS-FSE MSOFT-FSE FS-FSE 260 × 260 ≤4 20% 2 YES
with fat saturation
Oblique coronal T2-weighted TSE FSE FSE FSE FSE 260 × 260 ≤4 20% 2 NO
Oblique axial diffusion weighted Diffusion-weighted imaging 380 × 380 ≤4 10% 2 YES
imaging EPI planar
Oblique axial 3D T1 W GE with fat THRIVE VIBE LAVA QUICK 3D TIGRE 380X380 ≤3 0 3 YES
saturation (Gd)
3D T2 weighted sequence VISTA SPACE CUBE 3D MVOX isoFSE 260 × 260 ≤1.5 0 3 NO

FSE, fast spin echo; TSE, turbo spin echo; SPIR-TSE, spectral saturation with inversion recovery turbo spin echo; FS-TSE, fat-saturated turbo spin echo;
FS-FSE, fat saturated fast spin echo; THRIVE, T1-weighted high-resolution isotropic volume examination; VIBE, volumetric interpolated breath-hold
examination; LAVA, liver acquisition with volume acceleration-extended volume; QUICK 3D, Toshiba name of the sequence, not an acronym; TIGRE,
T1-weighted gradient echo with RF fat saturation; DWI, diffusion-weighted imaging; VISTA, volume isotropic turbo spin echo acquisition; SPACE,
sampling perfection with application optimised contrasts using different flip angle evolution; CUBE, GE name of the sequence, not an acronym; 3D
MVOX, 3D multivoxel; isoFSE, iso fast spin echo

fistulas may coexist; in such cases, it is mandatory to 1. Identify each path and follow it throughout its course. If
perform a second study with a conventional external coil. more than one path is present, it is very important to look
Given the complexity of fistulas in patients with PCD, for possible communications among them.
Horsthuis et al. [22] proposed a series of five steps to ensure 2. Look for abscesses and blind paths. They are more easily
an appropriate approach to pelvic MRI in patients with CD: identified in T2-weighted sequences with fat saturation.

Fig. 1 Complex bilateral

transsphinteric fistula cursing
across both ischioanal fossae
(arrows). a Oblique-coronal T2-
weighted image. b Oblique-
transverse fat-supressed T2-
weighted image. c Post-
gadolinium oblique-transverse
fat-supressed gradient echo T1-
weighted image. d Native
oblique-transverse native
diffusion-weighted image with
800 s/mm2 b factor
Insights Imaging (2018) 9:47–58 51

Fig. 2 Right suprasphicteric fistula (arrows) with retroanal abscess (asterisk). a Oblique-coronal T2-weighted image. b Oblique-transverse T2-weighted
image. c Native oblique-transverse native diffusion-weighted image with 800 s/mm2 b factor

3. Check, preferably in the coronal sequence, if the path- symptomatic and may result in the formation of a fistula
ways reach or surpass the levator ani muscle. (Fig. 5). When deep ulcers are present, there are almost
4. Identify the internal opening. It is usually located at the always signs of proctitis. These ulcers are easily detected
level of the dentate line (i.e., about 2 cm from the by MRI because they are usually associated with inflam-
anocutaneous margin), although it can be located at any matory infiltrates. They may be confused with infectious
site. lesions, post-radiotherapy lesions, or ulcerated cancers.
5. Identify ancillary findings such as inflammation of other – Stenosis: These lesions may be either inflammatory ste-
tissues (proctitis, infiltrates, bone oedema) or cancer. noses caused by anal spasm (type I) or true fibrous scar
tissue (type II). They are usually asymptomatic until they
Differential diagnoses should include pilonidal sinuses, reach a high degree of severity. A typical appearance of
haemorrhoids, and especially hidradenitis suppurativa, a dis- type II has been described as perianal hypointensity on
ease that is associated with and may coexist with PCD, induc- T2-weighted images and peripheral anal enhancement af-
ing clinical and histological confusion between the two dis- ter gadolinium administration [25] (Fig. 6).
eases. The coexistence of abscesses in other locations such as – Cutaneous flaps: The cause of these flaps is lymphedema
the groin or axillae and the presence of multiple fistulas and secondary to lymphatic obstruction, and 30% of flaps
blind pathways without a clear origin in the anorectal region contain non-caseiform granulomas. They are usually lo-
suggest hidradenitis suppurativa. Pelvic MRI may reveal this cated near the margins of superficial fissures and are al-
disease if skin thickening and subcutaneous induration are most always asymptomatic. Two types of cutaneous flaps
present in the perianal area, anal cleft, and perineum (Fig. 4) exist. The first type of flap is large, oedematous, and
[23, 24]. cyanotic and typically appears alongside a healed ulcer.
The second is called an Belephant ear^ flap and is flat,
soft, and painless. The flaps are usually hyperintense on
Other possible lesions in patients with PCD T2-weighted images and show poor enhancement after
gadolinium administration (Fig. 7).
In addition to fistulas, other less well-known manifestations of – Neoplastic lesions: Patients with highly evolved perianal
PCD also exist. Most result from the primary lesions caused disease may develop malignant lesions such as anal squa-
by the disease [25]. Such manifestations include: mous carcinoma or adenocarcinoma of the distal rectum,
but the risk for such neoplasia seems quite low. In these
– Ulcerations: These lesions comprise both fissures and cases, imaging studies do not substantially help in early
deeper cavitated ulcers. The latter can be very detection. Therefore, although no increase in the

Fig. 3 Anovaginal fistula. A thick gauge fistula exibited abscess formation involving the recto vaginal septum (arrows). a Sagittal T2-weighted image. b
Oblique-transverse T2-weighted image. c Post-gadolinium oblique-transverse fat-supressed gradient echo T1-weighted image
52 Insights Imaging (2018) 9:47–58

Fig. 4 Hidradenitis suppurativa. Skin and subcutaneous abscesses transverse fat-supressed gradient echo T1-weighted image. c Post-
(arrows) are present in the anal cleft without anal involvement. a gadolinium oblique-coronal fat-supressed gradient echo T1-weighted
Oblique-transverse T2-weighted image. b Post-gadolinium oblique- image

incidence of cancer has been demonstrated in patients based on the evaluation of secretion, pain, restriction of sexual
with chronic PCD treated with anti-tumour necrosis fac- activity, type of perianal fistula, and degree of induration [27].
tor (anti-TNF) drugs, careful inspection is recommended, A simpler method is fistula drainage assessment [28].
and anal biopsies under anaesthesia may even be needed Gentle compression of the fistula path allows for characteri-
before starting treatment with this type of drug [26]. sation of the fistula as open or closed. A fistula is considered
open when content comes out upon compression, while a
closed fistula is characterised by the absence of content (al-
though it is better to describe closed fistulas as those Bwithout
drainage^). Using this method, treated patients can be classi-
fied as responders (when the drainage ceases), in remission
Determination of activity (when the drainage decreases by >50%), or nonresponders.
In the daily clinical setting, the use of these methods in the
Several PCD activity measures are used in the clinical setting; physical examination is generally adequate. However, pelvic
the most well known is the perianal disease activity index. It is MRI is increasingly requested both in the initial evaluation

Fig. 5 Anal ulcer. Diffuse

hyperintensity of the left anal
margin is present on T2-weighted
images, and enhancement is
present in the post-gadolinium
sequence (arrows). A true fistula
developed during follow-up
(open arrows). a Oblique-
transverse fat-supressed T2-
weighted image. b Oblique-
coronal T2-weighted image. c
Oblique-transverse fat-supressed
T2-weighted image. d Oblique-
coronal T2-weighted image
Insights Imaging (2018) 9:47–58 53

Fig. 6 Anal fibrous stenosis. The

internal anal sphincter shows
hypointensity on T2-weighted
sequences and diffuse
enhancement after gadolinium
administration (arrows); the
native diffusion-weighted image
and the ADC map show
hypointensity. a Oblique-coronal
T2-weighted image. b Post-
gadolinium oblique-transverse
fat-supressed gradient echo T1-
weighted image. c Native
oblique-transverse native
diffusion-weighted image with
800 s/mm2 b factor. d ADC map

and during follow-up [12], and whenever studies are per- One of the first studies of the use of MRI for monitoring
formed for monitoring treatment, it should be mandatory [11]. patients undergoing treatment with anti-TNF drugs was per-
formed by Van Assche et al. [31]. They also assessed the use
of a numerical scale of severity and extension based on mor-
phological findings and the presence of signs of inflammation.
Role of MRI in follow-up The scale was used to evaluate the number and complexity of
fistulas, the degree of hyperintensity in T2-weighted se-
MRI studies have shown that closure of the external opening quences, the presence of abscesses, and the presence of signs
does not always indicate that the fistula is fully healed because of rectal inflammation (Table 2). However, although the resul-
signs of inflammation can persist in the internal tissues of the tant score (Van Assche score) provides a quick quantitative
fistula [29] (Figs. 8 and 9). This can also be demonstrated with idea of the complexity and severity of PCD, its usefulness for
US [30] and indicates the importance of evaluating perianal patient monitoring remains unclear. Karmiris et al. [32], part
fistulas using imaging methods not only for surgical planning of this same group of authors, later evaluated 59 patients treat-
but also for treatment monitoring. When using anti-TNF ed with infliximab and found a significant decreases in their
drugs, rigorous patient monitoring is needed not only because scores in the short term (p < 0.002) and medium term
such drugs are expensive but also because they are not free of (p < 0.0001), but not in the long term (e.g., 1 year). In another
side effects such as infection, hypersensitivity, and others. long-term follow-up study, Ng et al. [33] evaluated 34 patients

Fig. 7 Cutaneous flap showing hyperintensity on T2-weighted images c Post-gadolinium oblique-transverse fat-supressed gradient echo T1-
and a slight post-gadolinium rim enhancement (arrows). a Oblique- weighted image
coronal T2-weighted image. b Oblique-transverse T2-weighted image.
54 Insights Imaging (2018) 9:47–58

Fig. 8 Complete response after anti-TNF therapy. a Axial T2-weighted 24 weeks of treatment show that all paths have decreased in thickness
sequences at baseline show a complex left trans-sphincteric fistula with and exhibit hypointensity (arrow)
hyperintense paths (arrow); b Axial T2-weighted sequences after

treated with anti-TNF drugs and found that the Van Assche expected that the superior signal-to-noise ratio of 3-T equip-
score was not significantly different between responders and ment will allow for improved spatial resolution compared with
nonresponders. Likewise, in a study carried out by Savoye- 1.5-T equipment as well as easier identification of fistulas and
Collet et al. [34], more than 20 patients were assessed at base- greater accuracy in three-dimensional reconstructions.
line and after 1 year of treatment with anti-TNF drugs. The The apparent limitations of the conventional MRI approach
authors found no significant variations in the Van Assche to PCD led several groups of investigators to assess other MRI
score among responders, patients in remission, and nonre- parameters that could be used to quantitatively evaluate dis-
sponders. Finally, Hortshuis et al. [35] studied 16 patients ease severity and variations that reflect the treatment effects.
treated with infliximab and found no significant variations in The aforementioned study by Savoye-Collet et al. [34] proved
the Van Assche score after treatment, even when clinical im- that decreased intensity in T2-weighted sequences (p < 0.01)
provement or decreases in biological markers such as C- and a subjective decrease in enhancement after administration
reactive protein had occurred. of intravenous gadolinium (p < 0.02) occurred in patients who
exhibited a response or remission after treatment. The authors
also observed that the disappearance of the post-gadolinium
New MRI tools enhancement predicted clinical remission.
Another MRI tool is diffusion imaging. This type of se-
New high-field MRI equipment can provide better perfor- quence reflects the restrictions on the free movement of water
mance in the study of perianal fistulas. Publications regarding molecules in the tissues that occur due to ischaemia, increased
the performance of 3-T machines in the study of pelvic dis- cellularity, or the presence of macromolecules. Although such
eases such as gynaecological, prostatic, and rectal cancers sequences have little spatial resolution, they show greater con-
have been extensively published. However, we found no re- trast between the tissues, making the lesions easier to identify.
ports comparing 3- and 1.5-T machines in the study of This restriction can be measured because the image we obtain
perianal fistulas. Despite this lack of information, it is has a quantifiable apparent diffusion coefficient (ADC). The

Fig. 9 Partial response after anti-TNF therapy. a Axial T2-weighted Axial T2-weighted sequences after 24 weeks of treatment reveal that
sequences at baseline show a complex bilateral trans-sphincteric fistula the distal part of the fistula shows signs of healing (hypointensity); how-
with posterior Bhorseshoe^ fluid. Hyperintense paths are present both ever, the intersphincteric fluid collection persists (arrow)
near the external opening and at the intersphincteric level (arrows). b
Insights Imaging (2018) 9:47–58 55

Table 2 Magnetic these types of studies, especially for assessing the evolution
resonance imaging- Number of fistula tracks
of the ADC during treatment.
based score for severity None 0
of perianal Crohn’s MRI perfusion studies (dynamic contrast-enhanced
Single, unbranched 1
disease as described by MRI) reflect the degree of tissue inflammation with great
Van Assche et al. [31] Single, branched 2
fidelity. These techniques are based on the acquisition of
Multiple 3
images at the moment of passage of the contrast medium
Location
by the organ of interest, thus reflecting the dynamic re-
Extra- or intersphicteric 1
sponse of this tissue to the arrival of blood and its subse-
Transsphicteric 2
quent distribution in the extracellular space. The analysis
Suprasphinteric 3
of signal changes as a function of time can be carried out
Extension
by studying time-intensity curves from a qualitative view-
Infraelevatoric 1
point or using specific software that provides information
Supraelevatoric 2
on semiquantitative or quantitative parameters. The most
Hyperintensity on T2-weighted images useful semiquantitative parameters obtained from the
Absent 0 analysis of time-intensity curves are the maximum en-
Mild 4 hancement, rate of ascent of the curve, time for the max-
Pronounced 8 imum value, and area under the curve (Fig 11).
Collections (cavities >3 mm diameter) Quantitative parameters are based on models described
Absent 0 by Tofts et al. [41] and include the transferability of gad-
Present 4 olinium through the vascular endothelium (K trans), the
Rectal wall involvement fractional volume of the extracellular space (ve), and the
Normal 0 relationship between these two parameters (kep).
Thickening 2 Horsthuis et al. [42] performed the first studies using
these types of sequences and observed that greater num-
bers of pixels were rapidly enhanced in patients with more
inflamed tissues usually show diffusion restriction (Fig. 10); clinically severe disease, although they did not observe a
thus, these sequences seem useful for the detection of perianal correlation between the semiquantitative parameters and
fistulas [36]. Additionally, because abscesses show low the perianal disease activity index. However, in a later
ADCs, they can be detected with diffusion imaging, particu- study, Ziech et al. [43] did observe this correlation, al-
larly when the use of intravenous gadolinium is contraindi- though they only evaluated a small group of patients.
cated or otherwise not possible. Dohan et al. [37] found that an Nevertheless, their study showed that 6 weeks after initi-
ADC of <1.18 μm2/s can be used as a reference for diagnosing ation of treatment with anti-TNF drugs, the quantitative
perianal abscesses with a sensitivity of 100% and specificity parameter (Ktrans) had significantly decreased in treatment
of 90%. However, whether diffusion sequences reflect the responders, indicating that it may be a predictive param-
degree of inflammatory activity in patients with fistulas that eter of the response to treatment.
have not been complicated by abscesses remains unclear. Finally, in the field of non-routine MRI sequences,
Although Yoshizako et al. [38] found significant differences the most recent contribution to the assessment of the
in the mean ADC between active and inactive fistulas, other degree of PCD activity was the use of magnetisation
researchers did not [39, 40]. It seems necessary to expand transfer sequences. By varying the phase-coding

Fig. 10 Oblique-transverse T2-

weighted image (a) and merged
view of a colour-coded map
derived from an 800 s/mm2 b
factor and a T2-weighted image
(b). Right intersphinteric path
(arrow) is clearly more
conspicuous in image B
56 Insights Imaging (2018) 9:47–58

Fig. 11 Effect of treatment on

semiquantitative perfusion
parameters. a Variation of the
maximum relative enhancement
at baseline. b Variation of the
maximum relative enhancement
at 4 weeks of treatment

frequencies applied, the signal is altered in accordance being studied (Fig. 12). Pinson et al. [44] found that in
with the amount of macromolecules present in the tissue a group of 29 patients with CPE, the mean values of

Fig. 12 Calculation of the

magnetisation transfer ratio of a
fistula. a Signal intensity on an
oblique-transverse gradient-echo
image encoded at 800 off-
resonance radio frequency
saturation. b Signal intensity on
an oblique-transverse gradient-
echo image without radio
frequency saturation. The
magnetisation transfer ratio (%)
is = 1- (Msat image / Mnonsat
image) × 100
Insights Imaging (2018) 9:47–58 57

relative magnetisation transfer in the non-active group Open Access This article is distributed under the terms of the Creative
Commons Attribution 4.0 International License (http://
were significantly higher than those in the active group
creativecommons.org/licenses/by/4.0/), which permits unrestricted use,
(p < 0.02); additionally, the values were correlated with distribution, and reproduction in any medium, provided you give appro-
those of the Van Assche scale (p < 0.05). priate credit to the original author(s) and the source, provide a link to the
Creative Commons license, and indicate if changes were made.

Timing of MRI examination References

In patients undergoing medical treatment for perianal fis- 1. Sainio P (1984) Fistula-in-ano in a defined population. Incidence
tulas, the clinical response is usually detected before com- and epidemiological aspects. Ann Chir Gynecol 73:219–224
plete disappearance of the MRI abnormalities. Several 2. Keighley MR, Allan RN (1986) Current status and influence of
studies have shown that during the first weeks of treat- operation on perianal Crohn’s disease. Int J Color Dis 1:104–107
3. Beaugerie L, Sokol H (2012) Clinical, serological and genetic pre-
ment, changes occur that result in lower scores; however,
dictors of inflammatory bowel disease course. World J
it is only in the medium term (26–52 weeks) that this Gastroenterol 18:3806–3813
improvement is observed or even increased in treatment 4. Sachar DB, Bodian CA, Goldstein ES, Present DH, Bayless TM,
responders [32]. Likewise, another study showed that pa- Picco M et al (2005) Is perianal Crohn’s disease associated with
tients who were in remission at 6 months continued to be intestinal fistulization? Am J Gastroenterol 100:1547–1549
5. Silverberg MS, Satsangi J, Ahmad T, Arnott ID, Bernstein CN,
in remission 1 year later despite the fact that they had Brant SR et al (2005) Toward an integrated clinical, molecular
abandoned treatment [33]. and serological classification of inflammatory bowel disease: report
In summary, it seems appropriate to propose a baseline of a working party of the 2005 Montreal World Congress of
pelvic MRI study before starting treatment and to repeat the Gastroenterology. Can J Gastroenterol 19(Suppl A):5–36
MRI study 6 months later to assess the response. The result 6. Tozer PJ, Whelan K, Phillips RKS, Hart AL (2009) Etiology of
perianal Crohn’s disease: role of genetic, microbiological, and im-
will be of prognostic value for the long-term disease evolu- munological factors. Inflamm Bowel Dis 15:1591–1598
tion. If complete healing has not occurred, the treatment 7. Tozer PJ, Burling D, Gupta A, Phillips RKS, Hart AL (2011)
should be continued. Review article: medical, surgical and radiological management of
perianal Crohn’s fistulas. Aliment Pharmacol Ther 33:5–22
8. Sahni V, Ahmad R (2008) Which method is best for imaging of
perianal fistula? Abdom Imaging 33:26–30
Conclusions 9. Siddiqui MHR, Ashrafian H, Tozer P, Daulatzai N, Burling D, Hart
A et al (2012) A Diagnostic accuracy meta-analysis of endoanal
ultrasound and MRI for perianal fistula assessment. Dis Colon
MRI is the imaging method of choice for the initial study of Rectum 5:576–585
patients with PCD. Endoanal or transperineal US is an alter- 10. Bor R, Farkas K, Bàlint A, Szücs M, Ábrahám S, Milassin Á et al
native when experienced professionals are available to per- (2016) Prospective comparison of magnetic resonance imaging,
form it. transrectal and transperineal sonography, and surgical findings in com-
plicated perianal Crohn's disease. J Ultrasound Med 35:2367–2372
Experienced radiologists should perform MRI with an
11. Van Assche G, Dignass A, Panes J, Beaugerie L, Karagiannis J,
external multichannel coil whenever possible. Ideally, a Allez M et al (2010) The second European evidence-based consen-
structured report should describe the path of the fistula sus on the diagnosis and management of Crohn’s disease: special
in relation to the anatomical structures of the anus, with situations. J Crohns Colitis 4:3–101
an attempt to identify the internal orifice and the presence 12. Panés J, Bouhnik Y, Reinisch W, Stoker J, Taylor SA, Baumgart
of abscesses, secondary tracts, or other complications. If DC et al (2013) Imaging techniques for assessment of inflammatory
bowel disease: joint ECCO and ESGAR evidence-based consensus
possible, despite the known limitations of those that are guidelines. J Crohns Colitis 7:556–585
available, a score should be included to quantify the se- 13. Gecse KB, Bemelman W, Kamm MA, Stoker J, Khanna R, Ng SC
verity of the disease. In addition, attention should be paid et al (2014) A global consensus on the diagnosis and multidisci-
to any other perianal manifestations of CD that may be plinary treatment of perianal fistulising Crohn's disease. Gut 63:
present. 1381–1392
14. de Groof EJ, Cabral VN, Buskens CJ, Morton DG, Hahnloser D,
MRI is required while monitoring certain treatments,
Bemelman WA et al (2016) Systematic review of evidence and
such as anti-TNF drugs. In addition to the baseline consensus on perianal fístula: an analysis of national and interna-
study, another MRI examination should be performed tional guidelines. Color Dis 18:O119–O134
at approximately 6 months to establish a prognosis if 15. Buchanan GN, Halligan S, Taylor S, Williams A, Cohen R, Bartram C
possible. The use of new MRI sequences for the early (2004) MRI of fistula in ano: inter- and intraobserver agreement and
effects of directed education. AJR Am J Roentgenol 183:135–140
monitoring of special treatments such as anti-TNF drugs 16. Szurowska E, Wypych J, Izycka-Swieszewska E (2007) Perianal
seems promising, although more extensive studies are fistulas in Crohn's disease: MRI diagnosis and surgical planning.
still required. Abdom Imaging 32:705–718
58 Insights Imaging (2018) 9:47–58

17. Torkzad MR, Karlbom U (2010) MRI for assessment of anal fistula. therapy guided by magnetic resonance imaging for Crohn’s perineal
Insights Imaging 1:62–71 fistulas. Am J Gastroenterol 104:2973–2986
18. Morris J, Spencer JA, Ambrose NS (2000) MR imaging classifica- 34. Savoye-Collet C, Savoye G, Koning E, Dacher JN, Lerebours E
tion of perianal fistulas and its implications for patient management. (2011) Fistulizing perianal Crohn’s disease: Contrast-enhanced
Radiographics 20:623–635 magnetic resonance imaging assessment at 1 year on maintenance
19. Sandborn WJ, Fazio VW, Feagan BG, Hanauer SB (2003) AGA anti-TNF-alpha therapy. Inflamm Bowel Dis 17:1751–1758
technical review on perianal Crohn’s disease. Gastroenterology 35. Hortshuis K, Ziech MLW, Bipat S, Spijkerboer AM, de Bruine-
125:1508–1530 Dobben AC, Hommes DW et al (2011) Evaluation of an MRI-
20. Zhu Y-F, Tao G-Q, Zhou N, Xiang C (2011) Current treatment of based score of disease activity in perianal fistulizing Crohn’s dis-
rectovaginal fistula in Crohn’s disease. World J Gastroenterol 17: ease. Clin Imaging 35:360–365
963–967 36. Hori M, Oto A, Orrin S, Suzuki K, Baron RL (2009) Diffusion-
21. Dwarkasing S, Hussain SM, Hop WCJ, Krestin GP (2004) weighted MRI: a new tool for the diagnosis of fistula in ano. J Magn
Anovaginal fistulas: evaluation with endoanal MR imaging. Reson Imaging 30:1021–1026
Radiology 231:123–128 37. Dohan A, Eveno C, Oprea R, Pautrat K, Placé V, Pocard M et al
22. Horsthuis K, Stoker J (2004) MRI of perianal Crohn’s disease. AJR (2014) Diffusion-weighted MR imaging for the diagnosis of ab-
Am J Roentgenol 183:1309–1315 scess complicating fistula-in-ano: preliminary experience. Eur
23. Griffin N, Williams AB, Anderson S, Irwing PM, Sanderson J, Radiol 24:2906–2915
Desai N et al (2014) Hidradenitis suppurativa: MRI features in 38. Yoshizako T, Wada A, Takahara T, Kwee TC, Nakamura M, Uchida
anogenital disease. Dis Colon Rectum 57:762–771 K et al (2011) Diffusion-weighted MRI for evaluating perianal fis-
24. Monnier L, Dohan A, Amara N, Zagdanski AM, Drame M, Soyer P tula activity: Feasibility study. Eur J Radiol 81:2049–2053
et al (2017) Anoperineal disease in hidradenitis suppurativa: MR 39. Gallego JC, Echarri A, Porta A, Fernández-Pérez C. Role of
imaging distinction from perianal Crohn's disease. Eur Radiol. diffusion-weighted MRI in the evaluation of perianal Crohn's dis-
https://doi.org/10.1007/s00330-017-4776-1 ease. In: Final Programme ESGAR 2012, Edimburg: SE-012, p 95.
25. Bouguen G, Siproudhis L, Bretagne J-F, Bigard M-A, Peyrin- http://posters.webges.com/e-Poster/poster.php?id=47119&from=
Biroulet L (2010) Nonfistulizing perianal Crohn’s disease: clinical home. Access 15 May 2017. https://doi.org/10.5444/esgar2012/
features, epidemiology, and treatment. Inflamm Bowel Dis 16: SE-012
1431–1442 40. Bakan S, Olgun DC, Kandemirli SG, Tutar O, Samaci C, Dikici S
26. Ogawa H, Aneda S, Shibata C, Miura K, Nagao M, Ohnuma S et al et al (2015) Perianal fistula with and without abscess: assessment of
(2013) Adenocarcinoma associated with perianal fistulas in Crohn's fistula activity using diffusion-weighted magnetic resonance imag-
disease. Anticancer Res 33:685–690 ing. Iran J Radiol 12:e29084
27. Irvine EJ (1995) Usual therapy improves perianal Crohn's disease 41. Tofts PS, Brix G, Buckley DL, Evelhoch JL, Henderson E, Knopp
as measured by a new disease activity index. Mc Master IBG MV et al (1999) Estimating kinetic parameters for dynamic contrast-
Group. J Clin Gastroenterol 20:27–32 enhanced T(1)-weighted MRI of a diffusable tracer: standardized
28. Present D, Rutgeerts P (1999) Infliximab for the treatment of fistu- quantities and symbols. J Magn Reson Imaging 10:223–232
las in patients with Crohn’s disease. N Engl J Med 340:1398–1405 42. Horsthuis K, Lavini C, Bipat S, Stokkers P, Stoker J (2009) Perianal
29. Bell S, Halligan S, Windsor A (2003) Response of fistulating Crohn disease: evaluation of dynamic contrast-enhanced MR im-
Crohn’s disease to infliximab treatment assessed by magnetic res- aging as an indicator of disease activity. Radiology 251:380–387
onance imaging. Aliment Pharmacol Ther 17:387–393 43. Ziech MLW, Lavini C, Bipat S, Ponsioen CY, Spijkerboer AM,
30. Rasul I, Wilson SR, MacRae H, Irwin S, Greenberg GR (2004) Stokkers PCF et al (2013) Dynamic contrast-enhanced MRI in de-
Clinical and radiological responses after infliximab treatment for termining disease activity in perianal fistulizing crohn disease: a
perianal fistulizing Crohn’s disease. Am J Gastroenterol 99:82–88 pilot study. AJR Am J Roentgenol 200:W170–W177
31. Van Assche G, Vanbeckevoort D, Bielen D, Coremans G, Aerden I, 44. Pinson C, Dolores M, Cruypeninck Y, Koning E, Dacher JN,
Noman M et al (2003) Magnetic resonance imaging of the effects of Savoye G et al (2017) Magnetization transfer ratio for the assess-
infliximab on perianal fistulizing Crohn’s disease. Am J ment of perianal activity in Crohn's disease. Eur Radiol 27:80–87
Gastroenterol 98:332–339
32. Karmiris K, Bielen D, Vanbeckevoort D, Vermeire S, Coremans G,
Rutgeerts P et al (2011) Long-term monitoring of infliximab thera-
py for perianal fistulizing Crohn’s disease by using magnetic reso- Publisher’s Note
nance imaging. Clin Gastroenterol Hepatol 9:130–136
33. Ng SC, Plamondon S, Gupta A, Burling D, Swatton A, Vaizey CJ Springer Nature remains neutral with regard to jurisdictional claims in
et al (2009) Prospective evaluation of anti-tumor necrosis factor published maps and institutional affiliations.