-
Imaging of Fistula in Ano1
Steve Halligan, MD, FRCP, FRCRJaap Stoker, MD, PhD Fistula in
ano is a common condition that often recurs
despite seemingly adequate surgery, usually because ofinfection
that was missed at surgery. It is now increasinglyrecognized that
preoperative imaging can help identifyinfection that would have
otherwise gone unidentified. Inparticular, magnetic resonance (MR)
imaging findingshave been shown to influence surgery and markedly
dimin-ish the chance of recurrence; thus, preoperative imagingwill
become increasingly routine in the future. In this arti-cle, the
authors describe the pathogenesis, classification,and imaging of
fistula in ano, with an emphasis on MRimaging. Most important, the
authors describe how theradiologist is well placed to answer the
surgical riddles thatmust be solved for treatment to be
effective.
� RSNA, 2006
1 From the Department of Specialist Radiology, UniversityCollege
Hospital, Level 2, Podium, 235 Euston Rd, Lon-don NW1 2BU, England
(S.H.), and Department of Radiol-ogy, Academic Medical Center,
Amsterdam, the Nether-lands (J.S.). Received June 11, 2004;
revision requestedAugust 24; revision received October 15; accepted
Janu-ary 10, 2005; final version accepted March 2; updatedSeptember
29. Address correspondence to S.H.
� RSNA, 2006
REVI
EWS
AND
COM
MEN
TARY
�ST
ATE
OFTH
EAR
T
18 Radiology: Volume 239: Number 1—April 2006
-
F istula in ano (anal fistula) is a com-mon condition that has a
tendencyto recur despite seemingly ade-quate surgery. Recurrence is
usuallydue to infection that has escaped surgi-cal detection and
has thus gone un-treated. It is now increasingly recog-nized that
preoperative imaging, nota-bly magnetic resonance (MR) imaging,can
help identify infected tracts and ab-scesses that would otherwise
have beenmissed. Preoperative MR imaging find-ings have been shown
to influence sub-sequent surgery and markedly diminishthe chance of
recurrent disease as aresult. Because of this, preoperative
im-aging is likely to become increasinglyroutine in the future,
especially in pa-tients with recurrent disease.
This review will detail the pathogen-esis of fistula in ano,
explain how patho-genesis causes the different types of fis-tula
encountered, and describe howthese types can be imaged, with the
em-phasis on MR. Most important, we willdescribe how the
radiologist is wellplaced to answer the surgical questionsthat must
be solved for treatment to beeffective.
Anatomy and Etiology
To fully understand the role of imagingwith regard to fistula in
ano, an appreci-ation of its etiology and how the variousfistula
types are defined by anatomicboundaries is mandatory.
The anal canal is essentially a cylin-
der surrounded by two muscularsphincters, the internal and
externalanal sphincters, which are composed ofsmooth and striated
muscle, respec-tively (Fig 1). The external sphincterhas posterior
attachments to the ano-coccygeal ligament and anterior attach-ments
to the perineal body and urogen-ital diaphragm (and
bulbocavernousmuscle in boys and men) and mergesproximally with the
slinglike puborec-talis muscle (which defines the anorec-tal
junction), which itself merges withthe levator plate of the pelvic
floor.
The internal sphincter is the distaltermination of the circular
muscle of thegut tube. The rectal longitudinal smoothmuscle
interdigitates between the inter-nal and external sphincters and
isthought to have no obvious sphinctericeffect; rather, its role is
probably tobind the anus together (1).
The intersphincteric space is thesurgical plane of dissection
between theinternal and external sphincters and ismost frequently
found between the lon-gitudinal muscle and external sphincter,where
it exists as a sheet of fat contain-ing loose areolar tissue. The
fat-filledischioanal fossa lies lateral to thesphincter complex and
is traversed by anetwork of fibroelastic connective tissuefibers.
We prefer the term ischioanalfossa to ischiorectal fossa because
thisspace predominantly surrounds theanus rather than the rectum.
However,the two terms are interchangeable.
With regard to the lining of the analcanal, the proximal half is
characterizedby longitudinal mucosal folds, the analcolumns of
Morgagni (2). The distal as-pect of each column is linked to
itsneighbor by a small semilunar fold (theanal valves), which in
turn forms smallpockets (the anal sinuses, or crypts ofMorgagni).
The distal undulating limit ofthese valves is the dentate
(pectinate)line, which also marks the most distalaspect of the anal
transitional zone, ahistologic junction between anal squa-mous
epithelium and rectal columnarepithelium.
The dentate line lies approximately2 cm proximal to the anal
verge and is acrucial landmark in fistula in ano be-cause the anal
glands empty into the
crypts that lie proximal to the valves.These glands were first
linked to thegenesis of fistula in ano by Chiari (3),who suggested
that they were thesource of infection. Their purpose is un-clear,
although they may help lubricatethe anus by secreting mucus into
theanal crypts. The origin of the analglands within the surrounding
tissues isvariable. For example, they are presentin the
subepithelium and may bepresent in the internal sphincter,
andapproximately one- to two-thirds ofthese glands are deeply sited
within theintersphincteric space (4) (Fig 1). Mostauthorities
believe that it is infection ofthese intersphincteric glands that
is theinitiating event in fistula in ano, in aprocess known as the
“cryptoglandularhypothesis” (5). Furthermore, lym-phoid aggregates
surround the analglands, which may partly explain the in-creased
incidence of anal fistula inCrohn disease (6,7).
It is believed that gland infection re-sults in an
intersphincteric abscess if thedraining duct becomes blocked by
in-fected debris. This abscess may resolveby means of spontaneous
drainage intothe anal canal or may progress to anacute anorectal
abscess, which is a com-mon acute surgical emergency and isfamiliar
to all general and coloprocto-logic surgeons (8). Treatment
generallyconsists of incision and drainage of themost fluctuant
part of the abscess; how-ever, this procedure does not pay
dueattention to the source of infection inthe intersphincteric
space, with the re-sult that as many as 87% of patientswith an
acute abscess may subsequentlydevelop a fistula (9). Acute
anorectalabscess and fistula in ano are, there-fore, generally
believed to be acute andchronic manifestations, respectively, ofthe
same disease. Because of this, thesearch for intersphincteric
infection andan anal canal internal opening followed
Published online10.1148/radiol.2391041043
Radiology 2006; 239:18–33
Abbreviations:EUA � examination under anesthesiaSTIR � short tau
inversion recovery
Essentials
� Recurrent fistula in ano is usuallycaused by infection that
wasmissed during surgical explora-tion.
� MR imaging depicts remote foci ofinfection better than any
othermodality, including surgical explo-ration.
� MR image–guided surgery helpsreduce postoperative recurrenceby
75% in patients with complexdisease.
� Anal endosonography is a viablealternative to MR imaging
whenthe latter is not available.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
Radiology: Volume 239: Number 1—April 2006 19
-
by treatment of these at the time ofacute manifestation has been
advocatedin an attempt to reduce the incidence ofsubsequent fistula
(10).
Fistula in ano develops when an in-tersphincteric infection is
allowed tocontinue unabated. It has a prevalenceof approximately
0.01% and predomi-nantly affects young adults (11). Fistulain ano
is commoner in men, who domi-nate in all published series, with a
male-to-female ratio of approximately 2:1(11). Patients most
commonly presentwith discharge (65%), but local paindue to
inflammation is also frequent(12). However, some fistulas may
beentirely asymptomatic.
Fistula Classification
By definition, a fistula is an abnormal tractthat connects two
epithelial surfaces. Theanatomic course of an anal fistula will
bedictated by the location of the infectedanal gland and the
anatomic planes andboundaries that surround it. There willusually
be an internal enteric opening inthe anal canal at the level of the
dentateline—that is, at the original site of theduct draining the
infected gland. In mostcases this is at the 6-o’clock position,
be-cause anal glands are more abundant pos-teriorly (radial
positions around the anusare referenced with respect to a
clockface, with 12 o’clock being directly ante-rior) (13). The
fistula can reach the peri-anal skin by a variety of routes,
some
more tortuous than others, and by pene-trating and involving the
muscles of theanal sphincter and surrounding tissues toa variable
degree. Fistulas may thus beclassified according to the route taken
bythis “primary tract” that links the internaland external
openings.
Furthermore, classification largelydetermines treatment. In
1934, Milliganand Morgan (14) stressed the importanceof the
“anorectal ring” (anatomically, thepuborectalis muscle) and
categorized fis-tulas as those that entered the anal canalabove or
below this structure, warningthat postoperative incontinence
washighly likely if high fistulas were surgicallydivided without
due attention. This classi-fication was subsequently modified
andrefined by other authors, but the mostcomprehensive and
practical classifica-tion in use today is that of Parks et al
(6),who carefully analyzed a consecutive se-ries of 400 patients
referred to the sur-geons of St Mark’s Hospital, London, En-gland.
Parks et al found that they wereable to assign all fistulas into
one of fourgroups; intersphincteric, transsphinc-teric,
suprasphincteric, and extrasphinc-teric (6). Of importance, most of
thesegroupings could be explained in terms ofthe cryptoglandular
hypothesis.
The path of least resistance for fester-ing intersphincteric
infection is straightdown the intersphincteric space, whichcreates
an intersphincteric fistula; thistype of fistula composed 45% of
cases inthe series of Parks et al (6) (Fig 2). Of
importance, this fistula does not pene-trate the adjacent
external sphincter,which forms a relative barrier to
spread.However, some truculent fistulas cancross the external
sphincter and reach theischioanal fossa by doing so (Fig 2).
Thisresults in a transsphincteric fistula, whichcomposed 30% of
cases in the series ofParks et al. Other fistulas may spread
up-ward in the intersphincteric space andarch over the puborectalis
muscle, wherethey must cross the levator plate to reachthe perianal
skin. This type, the supras-phincteric fistula (Fig 2), composed
20%of cases in the series of Parks et al.
Parks et al (6) also noted a fourthtype of fistula in 5% of
cases. This typewas characterized by the surprising ab-sence of
intersphincteric infection. In-stead, the fistula entered the
rectum oranorectal junction directly (Fig 2).Parks et al termed
these extrasphinc-teric fistulas. Clearly, infection of theanal
glands cannot explain this type offistula, and Parks et al stressed
that pri-mary rectal or pelvic disease (eg, diver-ticular disease,
rectal Crohn disease,carcinoma) should be sought when thistype was
encountered. It should benoted that the Parks et al series
inevita-bly suffered spectrum bias owing to thespecialized nature
of St Mark’s Hospital,a bias acknowledged by Parks himself,with the
result that complex fistulaswere almost certainly
overrepresented.For example, Parks et al did not de-scribe
submucosal fistulas, which are
Figure 1
Figure 1: Illustration of anal canal anatomy inthe coronal
plane. EAS � external anal sphincter,IAS � internal anal sphincter,
IS � intersphinc-teric space, LA � levator ani, LM �
longitudinalmuscle, PR � puborectalis muscle.
Figure 2
Figure 2: Illustrations in coronal plane show classification of
fistula in ano according to Parks et al (6):intersphincteric,
transsphincteric, suprasphincteric, and extrasphincteric.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
20 Radiology: Volume 239: Number 1—April 2006
-
very superficial and do not involve thesphincter at all.
While most fistulas probably start asa simple single primary
tract, unabatedinfection may result in ramificationsthat branch
away from this. These sec-ondary tracts are generally known asknown
as “extensions” (Fig 3). Exten-sions may be intersphincteric,
ischio-anal, or supralevator (pararectal), andtheir morphology may
suggest tracts orabscesses. Exactly when a tract be-comes an
abscess is not precisely de-fined, but both terms describe
regionsof infection. The ischioanal fossa is thecommonest site for
an extension (6),especially one that arises from the apexof a
transsphincteric fistula (A in Fig 3).Extensions also occur in the
horizontalplane and are known as “horseshoes” ifthere is
ramification each side of theinternal opening (6) (Fig 3). The
ana-tomic description of the path taken bythe primary fistula tract
and the locationof any associated extension forms
its“classification.”
Assessment and Treatment
Treatment is usually straightforwardand involves laying open the
fistula by
means of surgical incision, usually bycutting down onto a metal
probe thathas been inserted into the primarytract. Indeed, James
Syme (in 1838)thought laying open a fistula to bestraightforward
and wondered why theprocedure was performed in a theaterwith “all
the pomp and circumstance ofa great operation” (15).
Unfortunately,he was mistaken; there are many trapswaiting for the
unwary. Inappropriateincision and injudicious exploration canall
too easily convert a simple fistula intoa surgical nightmare, with
disastrousconsequences for the patient. The pri-mary objectives are
to eradicate thetract and drain all associated sites ofinfection
while simultaneously preserv-ing anal continence.
It is the right balance between eradi-cation of infection and
preservation offunction that is the art of fistula surgery.To
achieve this, two surgical questionsneed to be answered
preoperatively:(a) What is the relationship between thefistula and
the anal sphincter (ie, can thetract be safely laid open with only
a lowrisk of postoperative incontinence), and(b) are there any
extensions from the pri-mary tract that need to be treated to
pre-vent recurrence, and, if so,where are they?
Surgeons have traditionally reliedon examination of patients who
havebeen administered a general anesthetic,a procedure referred to
as examinationunder anesthesia (EUA), to answerthese questions. At
EUA, the surgeonattempts to classify the fistula by palpat-ing it
to determine its relationship to thesphincter. However, anesthesia
andconsequent loss of tone impair preciseidentification of
underlying muscles. Ametal probe is inserted into the
externalopening and is directed toward the den-tate line to find
the internal opening.This is also not as straightforward as
itsounds. For example, the internal open-ing is frequently not
obvious, and thesurgeon may need to inject hydrogenperoxide into
the external openingwhile inspecting the anal canal.
The height of the internal openingrelative to the anal canal and
sphinctersis also crucial; the higher the opening,the more
sphincter will be divided. Al-though Milligan and Morgan (14)
real-
ized the importance of the puborectalismuscle for continence as
long ago as1934—“Whereas, if this ring be cut,loss of control
surely results”—they sug-gested that most of the remainingsphincter
could be sacrificed (14). How-ever, it is now generally accepted
thatfunctional disability can follow evenminimal division of the
sphincter. Per-haps most important, injudicious prob-ing during EUA
can create new second-ary tracts very easily. For example,forceful
probing in the roof of the ischio-anal fossa can rupture through
the leva-tor plate, causing a supralevator exten-sion or even a
rupture into the rectum,which would cause an
extrasphinctericfistula (6,12).
Identification of extensions at EUAis central to curing the
fistula. It is wellrecognized that missed extensions arethe
commonest cause of recurrence,which has reached 25% in some
series(16). Extensions require specific treat-ment and inevitably
necessitate moreextensive surgery. For example, supral-evator
extensions are particularly diffi-cult both to diagnose (because
they arehigh above the anal canal) and to treat(because the levator
plate forms a bar-rier to drainage).
The net result is that it can be verydifficult at EUA to
classify the primarytract with confidence, and there is am-ple
opportunity to make matters worse.Patients with recurrent disease
are aparticular case in point. Such patientsare more likely to
harbor missed dis-ease but are also difficult to assess. Mul-tiple
failed operations are the rulerather than the exception, with the
re-sult that digital palpation frequentlycannot help distinguish
between scar-ring due to repeated surgery and indu-ration due to an
underlying extension.Furthermore, patients in this group arealso
more likely to have extensions thattravel several centimeters away
fromthe primary tract, frustratingly in al-most any direction,
which further ham-pers detection of the extensions. Themore chronic
the fistula, the more com-plicated the associated extensions tendto
be. The inevitable result is that thesepatients become
progressively more dif-ficult to treat, with both patient and
sur-
Figure 3
Figure 3: Illustration in coronal plane showsfistula extensions
(secondary tracts): A � exten-sion into roof of ischioanal fossa,
arising fromapex of a transsphincteric fistula; B � supraleva-tor
pararectal extension arising from apex of atranssphincteric
fistula; C � supralevator exten-sion originating from
intersphincteric plane, D �intersphincteric horseshoe.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
Radiology: Volume 239: Number 1—April 2006 21
-
geon becoming ever more exasperated.The key to breaking this
loop is accuratepreoperative assessment.
Imaging Fistula in Ano: Fistulography
For many years radiologists have at-tempted to help answer the
surgicalquestions posed above, with varying de-grees of success.
Contrast material–en-hanced fistulography was the first mo-dality
used. In fistulography, the exter-nal opening is catheterized with
a finecannula, and a water-soluble contrastagent is injected gently
to define the fis-tula tract (Fig 4).
Unfortunately, fistulography has twomajor drawbacks. First,
extensions fromthe primary tract may fail to fill with con-trast
material if they are plugged with de-bris, are very remote, or
there is exces-sive contrast material reflux from eitherthe
internal or external opening. Second,the sphincter muscles
themselves are notdirectly imaged, which means that the
re-lationship between any tract and thesphincter must be guessed.
Furthermore,an inability to visualize the levator platemeans that
it can be difficult to decidewhether an extension has a supra- or
aninfralevator location. Similarly, the exactlevel of the internal
opening in the analcanal is often impossible to determinewith
sufficient accuracy to help the sur-geon. The net result is that
fistulographicfindings are both difficult to interpret
andunreliable.
Very little has been written on fistu-lography for fistula in
ano, probably be-cause the modality is so fraught witherrors.
Kuijpers and Schulpen (17) at-tempted to determine its value by
retro-spectively reviewing fistulographic im-ages in 25 patients.
They found that theinternal opening and associated exten-sions were
demonstrated and correctlyinterpreted in only four (16%)
subjects.Moreover, false-positive diagnoses ofrectal openings and
supralevator exten-sions were made in three (12%) pa-tients, which
would have resulted in se-rious surgical errors if acted on.
Theauthors concluded that fistulographywas “inaccurate and
unreliable,” al-though they admitted prior bias againstthe
technique (17). In contrast, Weis-
man and co-workers (18) found fistulog-raphy to be more useful
in that it pro-vided helpful information in nearly halfof the 27
subjects in their study.
It has been suggested that the primereason why fistulography is
generally un-helpful is that radiologists are not familiarwith the
concepts of fistula pathogenesisand anatomy and the relevant
surgicalquestions (17). One of the most hazard-ous
misinterpretations is to diagnose adirect rectal opening merely
becausethere is contrast material in the rectallumen; usually, the
contrast agent hasmerely refluxed up from the internal analopening.
Such radiology reports only en-courage the surgeon to look for
nonexist-ent openings and extensions, which canresult in iatrogenic
secondary tracts.
Imaging Fistula in Ano: CT
Computed tomography (CT) may depictfistula in ano, especially if
rectal and intra-venous contrast material are used, andinitial
reports were encouraging (19–21).However, fistula depiction is not
enough;fistulas must be classified correctly, andmore recent and
better data suggest thatCT cannot be used for this purpose
withsufficient accuracy. This is because theCT attenuation of the
anal sphincter andpelvic floor is similar to that of the
fistulaitself, unless the latter contains air or con-trast
material. This is compounded by theinability to image in the
surgically relevantcoronal plane. Comparative studies aresparse: A
study of 25 patients with 17fistulas found that CT could be used
tocorrectly classify only four fistulas, in con-trast to the 14
correct classificationsachieved with endosonography (22).
Po-tentially, the disadvantages of CT mightbe overcome by using
multi–detector rowCT fistulography, which offers the possi-bility
of isotropic voxels and multiplanarimaging, but results are awaited
and themotivation to perform these studies islikely limited by the
ready availability ofMR imaging. It should be borne in mindthat CT
is frequently used to search forabscesses in the context of Crohn
disease,and fistula in ano may be encountered atthe time of
examination. At present, how-ever, accurate classification of these
fistu-las is best left to MR imaging or anal en-
dosonography, with CT probably limitedto the diagnosis of
fistula-associated pel-vic abscesses where other imaging is
un-available or cannot be tolerated.
Imaging Fistula in Ano: AnalEndosonography
Anal endosonography, developed by CliveBartram, FRCP, FRCS,
FRCR, was thefirst technique to directly depict the analsphincter
complex in detail (23). Simplemodification of a rotating rectal
endo-probe by covering it with a nondeform-able plastic cone
allowed the transducerto be withdrawn through the anal canalitself,
situating it very close to the targetstructures and thus providing
images ofhigh spatial resolution. The technique hasattracted
considerable attention becauseof its ability to demonstrate the
presenceand extent of anal sphincter disruption,notably after
vaginal delivery (24). Analendosonography has also been
exten-sively used for the preoperative classifica-tion of fistula
in ano.
Figure 4
Figure 4: Fistulography in a male patient.Coronal image shows
that it is obvious that thereare several high extensions (arrows)
surroundingthe anorectal junction; however, the exact ana-tomic
location of these is unclear because thepelvic floor (ie, levator
ani in this case) cannot bedirectly visualized. Definition of
extension location(supra- or infralevator) is central to
surgicalmanagement.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
22 Radiology: Volume 239: Number 1—April 2006
-
The examination is simple, rapid, andwell tolerated by patients.
The patient liesin the left lateral position or in the
proneposition if female (25). The probe isgently inserted into the
distal rectum andthen withdrawn through the anal canal.The internal
sphincter is visualized as ahypoechoic ring encircling the anal
canal,whereas the external sphincter is ofmixed echogenicity (Fig
5). The inter-sphincteric space and longitudinal musclelie between
these and are of mixed echo-genicity and are easily identified by
usingmodern 10-MHz transducers (26).
Endosonography is particularly wellsuited to identification of
the internalopening, because this opening is usuallypositioned
right at the probe surface. It isimportant to realize, however,
that a tractextending up to the anal mucosal surfaceis rarely seen.
Although a breach in thesubepithelial layer of the anal canal is
oc-casionally present, it is more common forthe position of the
internal opening to berevealed as a hypoechoic focus in the
in-tersphincteric space that abuts the inter-nal sphincter, often
with a small corre-sponding defect in the internal sphincter(Fig
6). Because intersphincteric fistulasdo not stray beyond the
intersphinctericspace, they are usually very well visual-ized at
anal endosonography. Trans-sphincteric fistulas are revealed by
tracts
that cross the external sphincter to reachthe ischioanal fossa
(Fig 6). As would beexpected, extensions are revealed as
hy-poechoic fluid collections (Fig 7).
It was initially hoped that anal en-dosonography would
revolutionize preop-erative fistula classification, a view
sup-ported by the results of early studies (27).However, subsequent
work has been in-conclusive. For example, some investiga-tors
(28,29) have found the technique tobe useful, while others (30)
have found itto be no better than digital rectal exami-nation.
Furthermore, in direct compari-sons with MR imaging, anal
endosonogra-phy has been variously found to be supe-rior (31),
equivalent (32,33), or inferior(34,35). Much of this discrepancy is
prob-ably related to operator expertise, sinceanal endosonography,
similar to fistulog-raphy, is highly operator dependent.
However, there are undoubtedly sev-eral areas where anal
endosonographyhas specific disadvantages. For example,insufficient
penetration of the ultrasoundbeam beyond the external sphincter,
es-pecially with high-frequency transducers,limits the ability to
resolve ischioanal andsupralevator infections, with the resultthat
extensions from the primary tractmay be missed at endosonography.
Also,anal endosonography cannot be used toreliably distinguish
infection from fibro-sis, because both have a hypoechoic
ap-pearance (30). This causes particular dif-ficulties in patients
with recurrent dis-ease, since infected tracts and fibroticscars
are frequently combined. Attemptshave been made to clarify the
course ofpatent tracts by injecting hydrogen perox-ide or
sonography contrast agents intothe external opening during
examination(34,35). However, gas formed within thetract as a result
may cause acoustic shad-owing that mimics an extension. Indeed,this
phenomenon can occur with any tractthat contains air, leading, for
example, tointersphincteric fistulas being inadvertentlyclassified
as transsphincteric (Fig 8).
Anal endosonography is also disad-vantaged by the inability to
image in thesurgically important coronal plane, sothat it may be
very difficult to distin-guish supra- from infralevator exten-sions
(Fig 7). Some investigators (36)have attempted to overcome this
disad-
Figure 5
Figure 5: Intersphincteric fistula. Anal en-dosonogram in
transverse plane at mid–anal canallevel in a male patient shows
fistula with hypo-echoic tract located in intersphincteric plane
be-tween external (EAS) and internal (IAS) analsphincters. Internal
sphincter is markedly hypo-echoic. At surgery, the internal opening
was lo-cated at 6 o’clock posteriorly and was correctlypredicted
from anal endosonographic visualiza-tion because of the radial
position of the fistulawithin the intersphincteric plane.
Figure 6
Figure 6: Transsphincteric fistula shown onanal endosonogram in
the transverse plane at themid–anal canal level in a female
patient. In con-trast to Figure 5, the fistula (�) has penetrated
theexternal anal sphincter (EAS). The internal open-ing was
correctly predicted at 7-o’clock position.Note that the internal
sphincter is relatively thinnedhere, which is a clue to the site of
the internalopening, but there is no tract extending to the
analmucosa.
Figure 7
Figure 7: Anal endosonogram at upper analcanal level in a male
patient shows extensive hy-poechoic horseshoe extension (�).
Because en-dosonography is limited to the transverse plane, itis
difficult to determine whether this extension isinfra- or
supralevator.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
Radiology: Volume 239: Number 1—April 2006 23
-
vantage by employing a three-dimen-sional acquisition (Fig 9),
but this tech-nique remains relatively experimental.
There is no doubt that anal en-dosonography is a valuable
technique inthe right hands. A recent study (37) inwhich anal
endosonography was com-pared with digital rectal evaluation andMR
imaging in 108 primary tracts foundthat digital evaluation resulted
in cor-rect classification of 61% of fistulas; analendosonography,
81%; and MR imag-ing, 90%. Anal endosonography wasparticularly
adept with regard to thesite of the internal opening, with
correctprediction in 91% of cases comparedwith 97% for MR imaging
(37). How-ever, there is little doubt that MR imag-ing is a
superior technique overall and isnow also generally available.
Given this,the major role of anal endosonographyin fistula disease
is probably in the as-sessment of the degree of sphincter
dis-ruption in patients who become anallyincontinent after surgery
for a fistula.Endosonography has high spatial reso-lution, so it
also has a particular role inpatients who potentially have a
smallintersphincteric abscess that might bedifficult to resolve by
using standardbody- or surface-coil MR imaging.
Imaging Fistula in Ano: MR Imaging
In recent years, MR imaging has emergedas the leading contender
for preoperativeclassification of fistula in ano. The abilityof MR
imaging to help not only accuratelyclassify tracts but also
identify diseasethat otherwise would have been missedhas had a
palpable effect on surgical treat-ment and, ultimately, patient
outcome.
TechniqueMagnetic field strength does not appearto be a critical
factor for good results(38). Although higher field strengthmight be
relevant for more subtle differ-entiation of sphincter anatomy
andtracts, no definite diagnostic benefit hasbeen demonstrated, to
our knowledge.
Authors of initial reports (39–43) onMR imaging necessarily used
the bodycoil, with good results. However, au-thors of some
subsequent studies (44–45) reported lower accuracy when a
body coil was used; this was thought tobe possibly due to lower
spatial resolu-tion. The introduction of externalphased-array
surface coils increased thesignal-to-noise ratio and spatial
resolu-tion, to good effect (46,47), and thesecoils quickly became
generally available.
The best spatial resolution is achievedby using a dedicated
endoluminal anal coil(48), which may be combined with a sur-face
coil to increase the field of view (49).It should be noted that
these endoluminalcoils are not the same as a rectal coil butare of
smaller diameter and are designedto be placed in the anus. The
receiver coilis generally housed in a plastic cover thatallows
placement across the anal canal(Fig 10). The external diameter
generallyranges between 12 and 19 mm, althoughsmaller coils have
been used for pediatricexaminations (50,51). Endoluminal coilsare
susceptible to motion artifact, but thiscan be reduced by means of
careful pa-tient preparation. For example, patientsshould be asked
to try and relax thesphincter and pelvic floor as much as
pos-sible, and due attention should be paid tocomfort, including
support for the coiland patient with pads (50). Spasmolyticssuch as
20 mg of hyoscine butylbromide(Buscopan; Boehringer Ingelheim,
In-gelheim, Germany) or 1 mg of glucagonadministered
intramuscularly may help toreduce motion-induced artifacts.
(Hyo-
scine butylbromide is not licensed for usein the United
States.)
The choice of coil depends on per-sonal preference,
availability, the pa-tient group studied, and the clinicalquestion
in each patient. In a study of 10patients with cryptoglandular
fistula(45), an endoluminal coil was found tobe superior to a
surface coil; in a subse-quent study of 30 patients (52), how-ever,
a body coil was found to be supe-rior overall because the limited
field ofview inevitable with endoluminal imag-ing meant that
distant extensions weremissed, a phenomenon that is especially
Figure 8
Figure 8: Transverse anal endosonogram atupper anal canal level
in a female patient showsintersphincteric horseshoe extenstion
(arrows).Gas in the fistula causes acoustic shadowing (�),which
could be mistaken for transsphincterictracts.
Figure 9
Figure 9: Three-dimensional anal endosono-gram (coronal view) in
a female patient showstranssphincteric tract (arrows) that has been
out-lined by injecting hydrogen peroxide into the ex-ternal
opening.
Figure 10
Figure 10: Endoanal receiver coil for MRimaging.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
24 Radiology: Volume 239: Number 1—April 2006
-
frequent in patients with Crohn disease(53). In a third study
(49), endoluminaland phased-array coils were comparedin 20
patients, and the investigatorsfound that while the endoluminal
coilwas superior for classification of the pri-mary tract,
extensions were better im-aged by using the superior field of
viewof the external coil.
These results clearly suggest that alarge field of view is
necessary when-ever extensions are suspected—for ex-ample, in
patients with recurrent fistulaor Crohn disease. The high spatial
reso-lution of endoluminal coils makes themideal for precise
demonstration of thelocation and height of the internal open-ing,
and they may have a special role fordemonstrating ano- or
rectovaginal fis-tulas, which are notoriously difficult toimage
(54). Endoluminal coils are alsovaluable when simultaneous
informa-tion on the degree of sphincter disrup-tion is needed,
which may be the case inpatients who have undergone
previoussurgery. Endoluminal coils are some-times difficult to
place owing to anal ste-nosis or local pain as a result of
exten-sive infection. Halligan and Bartram(52) found that an
endoluminal coilcould not be placed in 17% of their pa-tients,
whereas Stoker and colleagues(50) failed to place the coil in only
3%,
which may reflect differences in patientpopulation since the
coil used in bothstudies was identical.
When circumstances allow, it islikely that an optimal
examination willbe achieved by using a combination ofboth external
and endoluminal coils.However, it should be borne in mindthat
accuracy with a body or externalcoil alone remains high
(41–43,47,49,52), and lack of an endoluminalcoil alone is
insufficient reason to avoidpreoperative MR imaging of fistula
inano. Indeed, examination with a bodyor phased-array coil has
become stan-dard practice, not least because endolu-minal coils
specifically designed for anal im-aging remain relatively
unavailable.
MR SequencesVarious investigators have adopteddifferent
strategies with respect to theMR sequences used to image fistula
inano. All agree that anatomic precisionis needed so that the
course of thefistula with respect to adjacent struc-tures can be
judged accurately, and alluse some method with which
infection(usually pus) can be highlighted.These aims can be
achieved by a vari-ety of means. Many investigators usethe rapid
and convenient fast spin-echo T2-weighted sequence, which
provides good contrast between hy-perintense fluid in the tract
and thehypointense fibrous wall of the fistulawhile simultaneously
enabling gooddiscrimination between the severallayers of the anal
sphincter (50,55).Others have used T1-weighted se-quences, which
must be combinedwith intravenous contrast material forthe fistula
to be highlighted (43). Fat-suppression techniques are widelyused
with both T2-weighted (56) andgadolinium-enhanced T1-weighted
se-quences (the latter may be especiallyvaluable in patients with
Crohn dis-ease, to differentiate between fluidand inflammatory
tissue that are bothhyperintense on fat-suppressed T2-weighted
and/or short tau inversion-recovery [STIR] images). Normal
ano-rectal structures do not enhance sub-stantially, except for the
internal analsphincter and blood vessels, includinghemorrhoids.
Investigators have alsosuccessfully employed STIR, a se-quence that
combines fat suppressionwith high conspicuity of active
tracts(41,47,52). Other approaches have in-cluded saline
instillation into the ex-ternal opening in an attempt to in-crease
tract conspicuity (57) or rectalcontrast medium (58,59).
However,such measures increase the complex-
Figure 11
Figure 11: Correct orientation for MR imagingof anal canal.
Sagittal T2-weighted scout imagethrough patient’s midline is used
to plan imagesthat are truly transverse with respect to anal
canal,as shown by white lines. Coronal imaging is thenperformed at
90° to the transverse plane.
Figure 12
Figure 12: Coronal (a) T2-weighted fast spin-echo (2500/70; echo
train length, 16; field of view, 300 mm;matrix, 256 � 512; section
thickness, 4 mm; gap, 0.4 mm) and (b) coronal STIR (4000/42,
inversion time of150 msec; echo train length, 16; matrix, 224 �
256; section thickness, 4 mm; gap 0.4 mm; two signals ac-quired) MR
images acquired with external phased-array coil show complex
transsphincteric fistula with tract(short straight arrows) in left
ischioanal fossa that extends below ischial bone (I) toward the
upper leg (notshown). At the ischial tuberosity, bone marrow edema
(long straight arrow) is visible on b. Arrowhead � ex-ternal
opening, curved arrow � small abscess, AS � anal sphincter.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
Radiology: Volume 239: Number 1—April 2006 25
-
ity of the examination in the face of thealready excellent
results achieved withmore standard procedures. Imagingprotocols are
detailed in the Table.
Imaging PlanesIt is central to success that imagingplanes are
correctly aligned with re-spect to the organ of interest, namelythe
anal canal. Because the anal canal istilted forward from the
vertical by ap-
proximately 45°, straight transverseand coronal images will fail
to achievethis alignment because of marked par-tial volume effect.
Oblique transverseand coronal planes oriented orthogonaland
parallel, respectively, to the analsphincter are therefore
necessary andare most easily planned by using a mid-line sagittal
image (Fig 11). It may benecessary to align supplementary
exam-inations with the rectal axis in complex
cases of an internal opening high in therectum, but this is
seldom needed.
It is important that the imaged volumeextend several centimeters
above the leva-tors and include the whole presacral space,both of
which are common sites for exten-sions. The entire perineum should
also beincluded. On occasion, tracts may extendfor several
centimeters, even leaving thepelvis or reaching the legs, and any
tractvisible must be followed to its termination if
Figure 13
Figure 13: Transsphincteric fistula in a man with Crohn disease.
(a, b) Transverse T2-weighted fast spin-echo MR images (echo train
length, 16; field of view, 300mm; matrix 256 � 512; section
thickness, 4 mm; gap, 0.4 mm; two signals acquired) obtained (a)
without (2500/70) and (b) with (4000/85) fat saturation and (c)
trans-verse fat-saturated contrast-enhanced T1-weighted fast
spin-echo image (see Table for parameters) show two separate
fistula tracts (straight and curved arrows) in leftposterior
ischioanal space, close to the anal sphincter (A). Both tracts show
confluent high signal intensity centrally, which represents pus in
the tract lumen. On a and b,the surrounding inflammatory tissue
(arrowheads) is of low signal intensity (a), which increases with
fat-saturation (b) and especially with contrast enhancement
(c).Anterior tract (curved arrow) demonstrates more adjacent
inflammation (arrowheads) than does posterior tract (straight
arrow).
External Phased-Array 1.5-T MR Protocols for Imaging Perianal
Fistula
Sequence and Plane* TR/TE†Echo TrainLength
Field ofView (mm) Matrix
Section Thickness/Intersection Gap(mm)‡
T2-weighted fast spin echo without fat suppressionSagittal,
coronal, transverse 2500/70 10 300§ 256 � 512 3/0.3
T2-weighted fast spin echo with fat suppressionTransverse
2500/70 10 300§ 256 � 512 3/0.3
T1-weighted fast spin echo with fat suppressionTransverse#
600/minimal 3 450 256 � 256 4/0.4
Note.—Parameters are typical and were established with EchoSpeed
imager (GE Healthcare, Milwaukee, Wis). Field of view should
include rectum and perineum; at endoanal MR, volume shouldencompass
entire sensitive region of coil. For all sequences, two signals are
acquired and bandwidth is 20.83 kHz.
* Transverse and coronal planes are off axis, orthogonal and
parallel, respectively, to anal canal.† TR/TE � repetition time
(msec)/echo time (msec).‡ For endoanal MR, transverse section
thickness and gap are 2–3 mm and 0.2–0.3 mm, respectively.§ For
endoanal MR, field of view is 160 mm for coronal and sagittal
planes and 100 mm for transverse plane.# Additional value of
routine use of this contrast-enhanced sequence is not yet
determined, but it may help differentiate inflammation from
abscess.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
26 Radiology: Volume 239: Number 1—April 2006
-
this has not been included on the standardimage volume (Fig 12).
The imaged volumeshould encompass the whole sensitive re-gion of
the coil when an endoanal receiveris used. The precise location of
the primarytract (eg, ischioanal or intersphincteric) isusually
most easily appreciated by usingtransverse images; the radial site
of the in-ternal opening is also well seen on images inthis plane.
Coronal images best depict thelevator plate,which helps distinguish
supra-from infralevator infection. The height ofthe internal
opening may also be best ap-preciated on coronal images, with the
ca-veat that the anal canal must be imagedalong its entire
craniocaudal extent.
The radial plane (as used for imagingmenisci of the knee) is
seldom used butseems attractive because it has the poten-tial to
depict fistulas along their full cranio-caudal extent, something
that is onlyachieved with standard coronal imaging
when the fistula is at the 3- or 9-o’clockposition. Instead of
being obtained parallelto the anal canal, the image planes
radiatefrom it like the spokes of awheel (60). Littleresearch has
specifically addressed the ben-efits of various imaging planes, but
in astudyof 20patients (60) investigators foundthat the combination
of a transverse seriesand a longitudinal series (coronal,
sagittal,radial, or a combination) provided all nec-essary
information for successful interpre-tation.
InterpretationThe success of MR imaging for preoper-ative
classification of fistula in ano is adirect result of the
sensitivity of MR fortracts and abscesses combined withhigh
anatomic precision and the abilityto image in surgically relevant
planes.Accurate preoperative classification isachieved by correctly
relating the im-aged fistula to the anal sphincter.
Primary TractActive tracts are filled with pus andgranulation
tissue and, thus, appear as
hyperintense longitudinal structures onT2-weighted or STIR
images (Fig 12).On contrast-enhanced T1-weighted im-ages, active
granulation tissue will en-hance while fluid in the tract itself
re-mains hypointense (Fig 13). Activetracts are often surrounded by
hypoin-tense fibrous walls (Fig 13), which canbe relatively thick,
especially in patientswith recurrent disease and previoussurgery.
Occasionally, some hyperin-tensity in this fibrous area may be
seen,probably reflecting edema. Hyperinten-sity may also extend
beyond the tractand its fibrous sleeve, where it repre-sents
adjacent inflammation (Fig 13).
The external anal sphincter is clearlyvisualized by using MR
imaging. It is rela-tively hypointense, and its lateral
bordercontrasts against the fat in the ischioanalfossa, both on
STIR (Fig 14) and especiallyon fast T2-weighted MR studies (Fig
13a).Thus, it is relatively easy to determinewhether a fistula is
contained by the exter-nal sphincter or has extended beyond it. If
afistula remains contained by the externalsphincter throughout its
course, then it is
Figure 14
Figure 14: Intersphincteric fistula in a malepatient. Transverse
STIR MR image (1500/15; fieldof view, 375 mm; matrix, 256 � 256;
sectionthickness, 4 mm; gap, 1 mm; four signals ac-quired) shows
that lateral margin of externalsphincter (long arrow) contrasts
against fat in theischioanal fossa (�). Fistula (short arrow) is in
theintersphincteric space posteriorly at 6 o’clock andis contained
by the external sphincter. There is notract in the ischioanal
fossa.
Figure 15
Figure 15: Fistula classified as suprasphincteric on coronal
STIR MR images (same parameters as for Fig14) in a female patient.
(a) Primary tracts in right (long arrow) and left (short arrow)
ischioanal fossae areshown. (b) Image obtained just posterior to a
shows that right-sided primary tract (white arrows) arches
overpuborectalis muscle (�) to reach a lower internal opening at
the dentate line level (black arrow).
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
Radiology: Volume 239: Number 1—April 2006 27
-
highly likely to be intersphincteric (Fig 14).In contrast, any
evidence of a tract in theischioanal fossa effectively excludes an
in-tersphincteric fistula. However, trans-sphincteric,
suprasphincteric, and extra-sphincteric fistulas all share the
commonfeature of a tract that lies beyond the con-fines of the
external sphincter. While atranssphincteric fistula will be the
common-est cause of a tract in the ischioanal fossa(Fig 13),
differentiation between thesethree fistulas is only possible by
locating theinternal opening and determining thecourse between this
and the primary tract.
Internal OpeningThe exact location of the internal openingcan be
difficult to define, whatever theimaging modality used. Two
questionsneed to be answered. What is the radialsite of the
internal opening, and what isits level? The vast majority of anal
fistulasopen into the anal canal at the level of thedentate line,
commensurate with thecryptoglandular hypothesis of
fistulapathogenesis. Furthermore, most fistulasalso enter
posteriorly, at the 6-o’clock po-sition. Unfortunately, the dentate
linecannot be identified as a discrete ana-tomic entity, even when
endoanal re-ceiver coils are used, but its general posi-tion can be
estimated with sufficient pre-cision for the imaging assessment to
beworthwhile. The dentate line lies at ap-proximately the mid–anal
canal level.This is generally midway between the su-perior border
of the puborectalis muscleand the most caudal extent of the
subcu-taneous external sphincter. These land-marks define the
“surgical” anal canal (asdistinct from the “anatomic” anal
canal,which is shorter and is defined as thecanal caudal to the
anal valves). The den-tate level is probably best appreciated
oncoronal views, which allow the craniocau-dal extent of the
puborectalis muscle andexternal sphincter to be appreciated;
withexperience, however, its location can beestimated with
reasonable precision byusing transverse views.
It should be noted that in some pa-tients the puborectalis
muscle is rathergracile, unlike the bulky muscle sug-gested in many
anatomy texts, and itfrequently and imperceptibly seguesinto the
external sphincter, all of which
hampers precise identification of themid–anal canal level.
Nevertheless, withexperience it is possible to estimate theexact
height of the internal opening withreasonable precision (47).
Any tract that penetrates the pelvicfloor above the level of the
puborectalismuscle is potentially a suprasphinctericor
extrasphincteric fistula. The level ofthe internal opening
distinguishes be-tween these types of fistula; specifically,the
internal opening is anal in supra-sphincteric fistulas (Fig 15) and
rectal inextrasphincteric fistulas (Fig 16).
Transsphincteric fistulas penetratethe external sphincter, a
feature that canbe easily appreciated on transverse (Fig17) or
coronal views. However, recentstudies (61) in which MR imaging
wasused have revealed that a transsphinc-teric tract may cross the
sphincter at avariety of angles. For example, it may
Figure 16
Figure 16: Fistula classified as extrasphinc-teric in a female
patient on coronal T2-weightedMR image (4563/150; field of view,
350 mm; ma-trix, 256 � 256; section thickness, 6 mm; gap, 0.6mm;
four signals acquired). Fistula tract (horizon-tal white arrows) is
seen in left ischioanal fossa.Levator plates (vertical white
arrows) are well de-picted bilaterally. Tract penetrates the left
levatorplate, and the internal opening (top horizontalwhite arrow)
is into the rectum, above the level ofthe puborectalis muscle (�)
and well above thedentate line (black arrow).
Figure 17
Figure 17: Transsphincteric fistula in a malepatient. Transverse
STIR MR image (same param-eters as for Fig 14) shows primary tract
(verticalarrow) in right ischioanal fossa, where it can beclearly
seen to penetrate external sphincter (�) toreach the
intersphincteric space. Internal openingis posterior at 6 o’clock
(horizontal arrow), at den-tate line level.
Figure 18
Figure 18: Transsphincteric fistula in a malepatient. Transverse
STIR MR image (same param-eters as for Fig 14) at level of the
internal openingshows primary tract (vertical arrow) at 4
–5o’clock. Unlike Figure 17, the tract cannot betraced right to the
anal mucosa, and the adjacentinternal sphincter (horizontal arrow)
appears in-tact. However, an internal opening at 4 –5 o’clockwas
reported because this position indicated siteof maximal infection
in the intersphincteric plane.The internal opening was confirmed at
this siteduring subsequent EUA. Intersphincteric plane iswell seen
in this patient between hyperintenseinternal sphincter and the
external sphincter.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
28 Radiology: Volume 239: Number 1—April 2006
-
arch upward as it passes through the ex-ternal sphincter and
thus cross the mus-cle at a higher level than would be de-duced
merely by inspecting the level ofthe internal opening. This is
important,because such tracts will require a greaterdegree of
sphincter incision during fistu-lotomy, with a corresponding
increase inthe risk of postoperative incontinence.MR imaging in the
coronal plane is bestfor estimates of the precise angulation ofthe
tract with respect to the surroundingmusculature (61).
The radial site of the internal openingis easy to identify if
the fistula tract can be
traced right to the anal mucosa (Fig 17).The radial site is
reported with respect toa clock face, with 12 o’clock being
directlyanterior. However, like endosonography,it is frequently
impossible to trace a tractright up to the anal mucosa, especially
ifan endoanal coil has not been used. Insuch cases, an intelligent
deduction mustbe made as to where the internal openingis likely to
be. This is best accomplishedby looking for the area of maximal
inter-sphincteric sepsis, since the internalopening is likely to
lie very close to this(Fig 18). The intersphincteric space
andlongitudinal layer are often seen as a hy-pointense ring between
the internal andexternal sphincters (Fig 18). The internalsphincter
is hyperintense on both T2-weighted fast spin-echo and STIR
images,especially if contrast material has beenused (62).
ExtensionsThe major advantage of MR imaging isthe facility with
which it can demonstrateextensions associated with a primarytract.
Morphologically, extensions fre-quently take the form of complex
tractsystems, regions of which have often be-come dilated to create
an abscess (al-though a precise radiologic distinction be-tween
abscess and a large tract remainselusive). Extensions appear as
hyperin-tense regions on T2-weighted and STIRimages and enhance if
intravenous con-trast material is used. Again,
collateralinflammation can be present to a variableextent.
The commonest type of extension is
one that arises from the apex of a trans-sphincteric tract and
extends into theroof of the ischioanal fossa (Figs 3, 19).The major
benefit of MR imaging find-ings is that they can alert the surgeon
toextensions that would otherwise bemissed. For example, extensions
maybe several centimeters from the pri-mary tract (Fig 20), which
makes themdifficult to detect during clinical exami-nation or EUA.
This is especially thecase when extensions are contralateralto the
primary tract (Fig 21). It is alsoimportant to search for
supralevator ex-tensions (Fig 22), since these are notonly
difficult to detect but pose specificproblems with regard to
treatment.Horseshoe extensions spread acrossboth sides of the
internal opening andare recognized on MR images by theirunique
configuration (Fig 23); horse-shoe extensions may be
intersphinc-teric, ischioanal, or supralevator. Com-plex extensions
are especially commonin patients with recurrent fistula in ano(Fig
21) or in those who have Crohndisease (Fig 13).
Effect of Preoperative MR on Surgeryand OutcomeOver the past few
years, imaging, nota-bly MR, has revolutionized the treat-ment of
patients with fistula in ano. Thisis because MR can be used to
classifyfistulas preoperatively with high accu-racy while also
alerting the surgeon todisease that would otherwise have
beenmissed. While there are reports of thetechnique dating from
1989 (40), it wasnot until the description by Lunniss andco-workers
(41) that the true potentialof MR imaging was fully
appreciated.Lunniss et al imaged 16 patients withcryptoglandular
fistula in ano and com-pared the MR classifications they ob-tained
with those from subsequentEUA. MR imaging proved correct in 14(88%)
cases, which immediately sug-gested that it was the most
accuratepreoperative assessment yet available.However, the
remaining two patients,in whom MR suggested disease but EUAyielded
normal findings, re-presentedsome months later with disease at
thesite initially indicated on MR images.This led the authors to
conclude that
Figure 19
Figure 19: Left-sided transsphincteric tract(short arrow) in a
female patient. Coronal STIR MRimage (same parameters as for Fig
14) showslarge extension (long arrow) from apex of tract intoroof
of ipsilateral ischioanal fossa.
Figure 20
Figure 20: Left-sided trans-sphincteric fistula (short
straightarrows) with internal opening at 6o’clock (long straight
arrow) in afemale patient. Transverse STIRMR image (same parameters
as forFig 14) shows remote extension(curved arrow) into
ipsilateralbuttock that was unsuspected atclinical examination but
is welldemonstrated at MR imaging. Theextension was found at
surgeryguided by MR findings.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
Radiology: Volume 239: Number 1—April 2006 29
-
MR imaging “is the most accuratemethod for determining the
presenceand course of anal fistulae” (41). Thiswork was rapidly
confirmed by othersworking in the field and was subse-quently
elaborated on.
Spencer and colleagues (63) inde-pendently classified 37
patients intothose with simple or those with complexfistulas on the
basis of MR imaging andEUA and found that MR results werethe better
predictor of outcome, withpositive and negative predictive
values,respectively, of 73% and 87% for MRand 57% and 64% for EUA.
These re-sults clearly implied that MR imagingand outcome were
closely related andagain raised the possibility that preop-erative
MR could help identify featuresthat cause postoperative
recurrence.
Beets-Tan and colleagues (46) ex-tended this hypothesis by
investigatingthe therapeutic effect of preoperative MRimaging; the
MR imaging findings in 56patients were revealed to the surgeon
af-ter he or she had completed an initialEUA. MR imaging provided
important ad-ditional information that precipitated fur-ther
surgery in 12 (21%) of 56 patients,predominantly in those with
recurrent fis-tula or Crohn disease (46).
Buchanan and co-workers (47) hypoth-esized that the therapeutic
influence and,thus, beneficial effect of preoperative MRimaging
would be greatest in patients withrecurrent fistula, since these
patients hadthe greatest chance of harboring occult in-fection,
while such fistulas were also themost difficult to evaluate
clinically. After aninitial EUA, Buchanan et al revealed
thefindings of preoperative MR imaging to thesurgeons for 71
patients with recurrent fis-tulas and left any further surgery to
thediscretion of the operating surgeon. Theyfound that
postoperative recurrence wasonly 16% for surgeons who always acted
ifMR findings suggested that areas of infec-tion had been missed,
whereas recurrencewas 57% for those surgeons who insteadalways
chose to ignore imaging results (47).Furthermore, in the 16
patients whoneeded further unplanned surgery, MR ini-tially
correctly predicted the site of this dis-ease in all cases (47).
Using a similar ap-proach, Buchanan and colleagues (64)
alsoinvestigated the effect of preoperative MR
imaging on clinical outcome in patients withfistula in ano at
initial presentation andfound that the scheduled surgical
approachchanged in 10% of this group.
Ever since the results of Lunniss et al(41) suggested that EUA
might be an im-perfect reference standard with which tojudge MR
imaging, comparative studieshave been plagued by the lack of a
genu-ine reference standard. It is now well rec-ognized that
surgical findings at EUA areoften incorrect. In particular, there
arefrequent false-negatives. In a recent com-parative study of
endosonography, MRimaging, and EUA in 34 patients with fis-tula due
to Crohn disease, Schwartz andco-workers (32) found that a
combinationof the results of at least two modalitieswas necessary
to arrive at a correct clas-sification. Indeed, it is well
establishedthat many false-negative surgical resultswill only
reveal themselves during long-term clinical follow-up, and, at this
pointin time, comparative studies that ignoreclinical outcome are
likely to be seriouslyflawed.
Imaging for Differential Diagnosis
Not all cases of perianal sepsis are due tofistula in ano. For
example, acne conglo-bata, hidradenitis suppurativa, pilonidal
sinus, actinomycosis, tuberculosis, proc-titis, human
immunodeficiency virus, lym-phoma, and anal and rectal carcinomamay
all cause perianal infection. While
Figure 21
Figure 21: Transsphincteric primary tract(short arrow) in the
right posterior quadrant of afemale patient. Transverse STIR MR
image (sameparameters as for Fig 14) shows two
left-sidedcontralateral extensions (long arrows) that
wereundetected at EUA until results of patient’s MRexamination were
revealed to the surgeon in theoperating theater.
Figure 22
Figure 22: Bilateral supralevator extensions(long arrows) in a
female patient. Coronal STIRMR image (same parameters as for Fig
14) clearlyshow levator plates (short arrows) bilaterally, sothat
it is easy for the radiologist to be confident thatinfection
extends above them.
Figure 23
Figure 23: Horseshoe extension (arrows) aris-ing from
intersphincteric fistula in a male patient.Transverse STIR MR image
(same parameters asfor Fig 14) shows that, in this case, the
horseshoepractically encircles the anal canal.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
30 Radiology: Volume 239: Number 1—April 2006
-
clinical examination results are often con-clusive, this is not
always the case, andimaging may help with the differential
di-agnosis. The cardinal feature of fistula inano is
intersphincteric infection, which isnot generally found in other
conditions.Whenever imaging suggests that infectionis superficial
rather than deep seated andthat there is no sphincteric
involvement,other conditions such as hidradenitis sup-purativa
should be considered (Fig 24).For example, authors of a recent
study(65) of patients with pilonidal sinus andfistula in ano found
that MR imaging couldbe used to reliably distinguish betweenthe two
on the basis of intersphinctericinfection and an enteric
opening.
The possibility of underlying Crohndisease should always be
considered inpatients who have a particularly com-plex fistula,
especially if the history isrelatively short. Indeed, a perianal
fis-tula is the presenting condition in 5% ofpatients (34), and
30%–40% of patientswith Crohn disease will experience analdisease
at some time (66–68). Specificscoring systems for perianal Crohn
dis-ease have been proposed, such as thePerianal Crohn’s Disease
Activity Index(69) and a recently described systembased solely on
MR imaging findings
(70). Small-bowel imaging may be usedto search for Crohn disease
when it issuspected, and the possibility of under-lying pelvic
disease should be consid-ered in any patient with an
extrasphinc-teric fistula, whether thought due toCrohn disease or
otherwise.
Which Patients Should Be Imaged?
While fistula in ano is simple to diag-nose and simple to treat
in most pa-tients, many other patients will benefitfrom detailed
and accurate preopera-tive investigation. Where there is easyaccess
to MR imaging, it could be ar-gued that all patients should
undergopreoperative imaging. For example, ithas been estimated that
the therapeuticeffect of MR imaging is 10% in patientspresenting
for the first time with aseemingly simple fistula (64). Where
ac-cess to imaging is more restricted, how-ever, the clinician and
radiologist willneed to select those patients who aremost likely to
benefit. Since there is nowoverwhelming evidence that MR imag-ing
alters surgical therapy and improvesclinical outcome in patients
with recur-rent disease, MR should be routine insuch cases.
Patients presenting for thefirst time with a fistula that
appearscomplex at clinical examination shouldalso be referred, as
should patients withknown Crohn disease, since the prepon-derance
of complex fistulas is increased.
There are also surgical situationswhere imaging is likely to be
particularlybeneficial, even when the fistula itself issimple. For
example, the anterior exter-nal sphincter is very short in
women,and division of this sphincter during fis-tula incision is
particularly associatedwith postoperative incontinence, evenwhen
the fistula itself is simple and hasno extensions. Faced with such
a di-lemma, the surgeon may choose to passa thread (seton) through
the tractrather than incise the fistula to providedrainage. The
patient can undergo post-operative imaging so that the
potentialextent of sphincter division can be as-sessed by
visualizing the relationship ofthe seton to the external sphincter.
Adecision can then be made whether toprogress with fistulotomy or
to keep the
seton in place for a few months, afterwhich time the internal
opening can beclosed with a rectal mucosal advance-ment flap. A
seton can also be placed atEUA when the surgeon is uncertainabout
the relationship between the tractand the sphincter; postoperative
imag-ing can then be used to help answer thisquestion. In patients
who undergo noveltherapies, such as use of fibrin glue, im-aging
may be necessary during instilla-tion to be confident that the
whole tracthas been filled (71).
MR imaging is not restricted to sur-gical assessment.
Infliximab, a chimericmonoclonal antibody to human tumornecrosis
factor–�, currently has aprominent role for the medical treat-ment
of Crohn disease, especially in pa-tients with a chronic fistula
(72). How-ever, infliximab therapy is contraindi-cated if an
abscess is present, and MRimaging may be used to search for
this(70,73). Indeed, MR imaging may beused to monitor infliximab
therapy,since it seems that fistulas may persistin the face of
clinical findings that sug-gest remission (74). For example,
inves-tigations (70,73) of MR in patientswhose external opening has
closed haverevealed that underlying infection canstill be present,
indicating a need forcontinued therapy. Further studies areneeded
to determine whether MR mon-itoring improves outcome.
Where MR imaging is unavailable orwhere competent interpretation
of theimages is not possible, then anal en-dosonography is a viable
and useful al-ternative. While recent comparisonshave shown that MR
imaging outper-forms anal endosonography in all re-spects, the
latter is far superior to sim-ple clinical examination, and its
perfor-mance in some areas is very close tothat of MR (37).
Notably, anal en-dosonography is very adept at depictingthe
internal opening.
Conclusion
In those patients with fistula in ano whohave a high likelihood
of complex dis-ease, the evidence that preoperativeMR imaging
influences the surgical ap-proach and the extent of exploration
Figure 24
Figure 24: Hidradenitis suppurativa in a malepatient. Transverse
STIR MR image (same param-eters as for Fig 14) shows extensive
superficialinfection (arrows). Absence of any infection relatedto
anal canal and intersphincteric space meant thatdiagnosis could be
confidently made preopera-tively by using imaging.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
Radiology: Volume 239: Number 1—April 2006 31
-
and improves the ultimate outcome isnow overwhelming. We hope
that thisarticle will stimulate radiologists to pro-vide this
service to their surgeons in theexpectation that this will reduce
the in-cidence of recurrent fistula in ano andthe misery that this
causes.
References1. Lunniss PJ, Phillips RKS. Anatomy and func-
tion of the anal longitudinal muscle. Br J
Surg1992;79:882–884.
2. Woodburne RT. Essentials of human anat-omy. 7th ed. Oxford,
England: Oxford Uni-versity Press, 1983.
3. Chiari H. Uber die analen divertikel der rec-tumschleimhaut
und ihre beziehung zu denanal fisteln. Wien Med Press
1878;19:1482–1483.
4. McColl I. The comparative anatomy and pa-thology of anal
glands. Ann R Coll Surg Engl1967;40:36–67.
5. Parks AG. Pathogenesis and treatment offistula-in-ano. Br Med
J 1961;5224:463–469.
6. Parks AG, Gordon PH, Hardcastle JD. Aclassification of
fistula-in-ano. Br J Surg1976;63:1–12.
7. Langman JM, Rowland R. Density of lym-phoid follicles in the
rectum and at the ano-rectal junction. J Clin Gastroenterol
1992;14:81–84.
8. Buchan R, Grace RH. Anorectal suppu-ration: the results of
treatment and the fac-tors influencing the recurrence rate. Br
JSurg 1973;60:537–540.
9. Fucini C. One stage treatment of anal ab-scesses and
fistulae. A clinical appraisal onthe basis of two different
classifications. Int JColorectal Dis 1991;6:12–16.
10. Eisenhammer S. Advance of anorectal sur-gery with special
reference to ambulatorytreatment. S Afr Med J 1954;28:264–266.
11. Sainio P. Fistula-in-ano in a defined popula-tion. Incidence
and epidemiological aspects.Ann Chir Gynaecol 1984;73:219–224.
12. Practice parameters for treatment of fistula-in-ano:
supporting documentation. The Stan-dards Practice Task Force. The
AmericanSociety of Colon and Rectal Surgeons. DisColon Rectum
1996;39:1363–1372.
13. Kuster GG. Relationship of anal glands andlymphatics. Dis
Colon Rectum 1965;8:329–332.
14. Milligan ET, Morgan CN. Surgical anatomyof the anal canal
with special reference to
anorectal fistulae. Lancet 1934;2:1213–1217.
15. Syme J. On diseases of the rectum. Edin-burgh, Scotland:
Adam and Charles Black,1838; 1–2.
16. Lilius HG. Fistula-in-ano, an investigation ofhuman foetal
anal ducts and intramuscularglands and a clinical study of 150
patients.Acta Chir Scand Suppl 1968;383:7–88.
17. Kuijpers HC, Schulpen T. Fistulography forfistula-in-ano: is
it useful? Dis Colon Rectum1985;28:103–104.
18. Weisman RI, Orsay CP, Pearl RK, et al. Therole of
fistulography in fistula-in-ano: reportof five cases. Dis Colon
Rectum 1991;34:181–184.
19. Guillaumin E, Jeffrey RB, Shea WJ, et al.Perirectal
inflammatory disease: CT find-ings. Radiology 1986;161:153–157.
20. Fishman EK, Wolf EJ, Jones B, Bayless TM,Siegelman SS. CT
evaluation of Crohn’sdisease: effect on patient management. AJRAm J
Roentgenol 1987;148:537–540.
21. Yousem DM, Fishman EK, Jones B. Crohndisease: perirectal and
perianal findings atCT. Radiology 1988;167:331–334.
22. Schratter-Sehn AU, Lochs H, Vogelsang H,Schurawitzki H,
Herold C, Schratter M. En-doscopic ultrasonography versus
computedtomography in the differential diagnosis ofperianorectal
complications in Crohn’s dis-ease. Endoscopy 1993;25:582–586.
23. Law PJ, Bartram CI. Anal endosonography:technique and normal
anatomy. GastrointestRadiol 1989;14:349–535.
24. Sultan AH, Kamm MA, Hudson CN, ThomasJ, Bartram CI. Anal
sphincter disruptionduring vaginal delivery. N Engl J Med
1993;329:1905–1911.
25. Frudinger A, Bartram CI, Halligan S, KammMA. Examination
techniques for anal en-dosonography. Abdom Imaging
1998;23:301–303.
26. Frudinger A, Halligan S, Bartram CI, PriceAB, Kamm MA,
Winter R. Female analsphincter: age-related differences in
asymp-tomatic volunteers with high frequency en-doanal US.
Radiology 2002;224:417–423.
27. Law PJ, Talbot RW, Bartram CI, et al. Analendosonography in
the evaluation of perianalsepsis and fistula-in-ano. Br J Surg
1989;76:752–755.
28. Deen KI, Williams JG, Hutchinson R, Keigh-ley MR, Kumar D.
Fistula in ano: endoanalultrasonographic assessment assists
decisionmaking for surgery. Gut 1994;35:391–394.
29. Lindsey I, Humphreys MM, George BD,
Mortensen NM. The role of anal ultrasoundin the management of
anal fistulas. Colorec-tal Dis 2002;4:118–122.
30. Choen S, Burnett S, Bartram CI, NichollsRJ. Comparison
between anal endosonogra-phy and digital examination in the
evaluationof anal fistulae. Br J Surg 1991;78:445–447.
31. Orsoni P, Barthet M, Portier F, Panuel M,Desjeux A, Grimaud
JC. Prospective com-parison of endosonography, magnetic reso-nance
imaging and surgical findings in ano-rectal fistula and abscess
complicatingCrohn’s disease. Br J Surg 1999;86:360–364.
32. Schwartz DA, Wiersema MJ, Dudiak KM, etal. A comparison of
endoscopic ultrasound,magnetic resonance imaging, and exam un-der
anesthesia for evaluation of Crohn’s peri-anal fistulas.
Gastroenterology 2001;121:1064–1072.
33. Gustafsson UM, Kahvecioglu B, Astrom G,Graf W. Endoanal
ultrasound or magneticresonance imaging for preoperative
assess-ment of anal fistula: a comparative study.Colorectal Dis
2001;3:189–197.
34. Kruskal JB, Kane RA, Morrin MM. Perox-ide-enhanced anal
endosonography: tech-nique, image interpretation, and clinical
ap-plications. RadioGraphics 2001;21(SpecIssue):S173–S189.
35. Chew SS, Yang JL, Newstead GL, DouglasPR. Anal fistula:
Levovist-enhanced endoanalultrasound—a pilot study. Dis Colon
Rectum2003;46:377–384.
36. Buchanan GN, Bartram CI, Williams AB,Halligan S, Cohen CR.
Value of hydrogenperoxide enhancement of three-dimensionalendoanal
ultrasound in fistula-in-ano. DisColon Rectum 2005;48:141–147.
37. Buchanan GN, Halligan S, Bartram CI, Wil-liams AB, Tarroni
D, Cohen CRG. Clinicalexamination, endosonography, and MR im-aging
in preoperative assessment of fistula inano: comparison with
outcome-based refer-ence standard. Radiology
2004;233(3):674–681.
38. Madsen SM, Myschetzky PS, Heldmann U,Rasmussen OO, Thomsen
HS. Fistula in ano:evaluation with low-field magnetic
resonanceimaging (0.1 T). Scand J
Gastroenterol1999;34:1253–1256.
39. Fishman-Javitt MC, Lovecchio JL, Javors B,Naidich JB,
McKinley M, Stein HL. The valueof MRI in evaluating perirectal and
pelvicdisease. Magn Reson Imaging 1987;5:371–380.
40. Koelbel G, Schmiedl U, Majer MC, et al.Diagnosis of fistulae
and sinus tracts in pa-tients with Crohn disease: value of MR
imag-
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
32 Radiology: Volume 239: Number 1—April 2006
-
ing. AJR Am J Roentgenol 1989;152:999–1003.
41. Lunniss PJ, Armstrong P, Barker PG, et al.Magnetic resonance
imaging of anal fistulae.Lancet 1992;340:394–396.
42. Barker PG, Lunniss PJ, Armstrong P,Reznek RH, Cottam K,
Phillips RK. Magneticresonance imaging of fistula-in-ano:
tech-nique, interpretation and accuracy. Clin Ra-diol
1994;49:7–13.
43. Spencer JA, Ward J, Beckingham IJ, AdamsC, Ambrose NS.
Dynamic contrast-en-hanced MR imaging of perianal fistulas. AJRAm J
Roentgenol 1996;167:735–741.
44. Van Beers B, Grandin C, Kartheuser A, et al.MRI of
complicated anal fistulae: comparisonwith digital examination. J
Comput AssistTomogr 1994;18:87–90.
45. Stoker J, Hussain SM, van Kempen D, Elev-elt AJ, Laméris
JS. Endoanal coil in MR im-aging of anal fistulas. AJR Am J
Roentgenol1996;166:360–362.
46. Beets-Tan RG, Beets GL, van der Hoop AG,et al. Preoperative
MR imaging of analfistulas: does it really help the surgeon?
Ra-diology 2001;218:75–84.
47. Buchanan G, Halligan S, Williams A, et al.Effect of MRI on
clinical outcome of recur-rent fistula-in-ano. Lancet
2002;360:1661–1662.
48. Hussain SM, Stoker J, Schouten WR, HopWC, Lameris JS.
Fistula in ano: endoanalsonography versus endoanal MR imaging
inclassification. Radiology 1996;200:475–481.
49. deSouza NM, Gilderdale DJ, Coutts GA,Puni R, Steiner RE. MRI
of fistula-in-ano: acomparison of endoanal coil with externalphased
array coil techniques. J Comput As-sist Tomogr 1998;22:357–363.
50. Stoker J, Rociu E, Zwamborn AW, SchoutenWR, Laméris JS.
Endoluminal MR imaging ofthe rectum and anus: technique,
applica-tions, and pitfalls. RadioGraphics 1999;19:383–398.
51. deSouza NM, Gilderdale DJ, MacIver DK,Ward HC.
High-resolution MR imaging ofthe anal sphincter in children: a
pilot study
using endoanal receiver coils. AJR Am JRoentgenol
1997;169:201–206.
52. Halligan S, Bartram CI. MR imaging of fistulain ano: are
endoanal coils the gold standard?AJR Am J Roentgenol
1998;171:407–412.
53. Stoker J, Laméris JS. MR imaging of perianalfistulas using
body and endoanal coil. AJRAm J Roentgenol 1999;172:1139–1140.
54. Stoker J, Rociu E, Schouten WR, LamérisJS. Anovaginal and
rectovaginal fistulas: en-doluminal sonography versus endoluminalMR
imaging. AJR Am J Roentgenol 2002;178:737–741.
55. Maier AG, Funovics MA, Kreuzer SH, et al.Evaluation of
perianal sepsis: Comparison ofanal endosonography and magnetic
reso-nance imaging. J Magn Reson Imaging 2001;14:254–260.
56. Halligan S, Healy JC, Bartram CI. Magneticresonance imaging
of fistula-in-ano: STIR orSPIR? Br J Radiol 1998;71:141–145.
57. Myhr GE, Myrvold HE, Nilsen G, et al. Peri-anal fistulas:
use of MR imaging for diagno-sis. Radiology 1994;191:545–554.
58. Mergo PJ, Helmberger T, Cerda JJ, UrrutiaM, Ros PR. Rectal
perflubron: new applica-tion in MRI of perirectal fistulae. J
ComputAssist Tomogr 1997;21:259–264.
59. Sabir N, Sunguntekin U, Erdem E, Nessar M.Magnetic resonance
imaging with rectal Gd-DTPA: new tool for the diagnosis of
perianalfistula. Int J Colorectal Dis 2000;15:317–322.
60. Stoker J, Jong Tjien Fa VE, Eijkemans MJ,Schouten WR,
Laméris JS. Endoanal MR im-aging of perianal fistulas: the optimal
imagingplanes. Eur Radiol 1998;8:1212–1216.
61. Buchanan GN, Williams AB, Bartram CI,Halligan S, Nicholls
RJ, Cohen CR. Potentialclinical implications of direction of a
trans-sphincteric anal fistula track. Br J Surg
2003;90:1250–1255.
62. deSouza NM, Kmiot WA, Puni R, et al. Highresolution magnetic
resonance imaging ofthe anal sphincter using an internal coil.
Gut1995;37:284–287.
63. Spencer JA, Chapple K, Wilson D, Ward J,
Windsor AC, Ambrose NS. Outcome aftersurgery for perianal
fistula: predictive valueof MR imaging. AJR Am J Roentgenol
1998;171:403–406.
64. Buchanan GN, Halligan S, Williams AB, et al.Magnetic
resonance imaging for primary fis-tula in ano. Br J Surg
2003;90:877–881.
65. Taylor SA, Halligan S, Bartram CI. Pilonidalsinus disease:
MR imaging distinction fromfistula in ano. Radiology
2003;226:662–667.
66. Platell C, Mackay J, Collopy B, Fink R, RyanP, Woods R. Anal
pathology in patients withCrohn’s disease. Aust N Z J Surg
1996;66:5–9.
67. Winter AM, Hanauer SB. Medical manage-ment of perianal
Crohn’s disease. SeminGastrointest Dis 1998;9:10–14.
68. Schwartz DA, Loftus EV Jr, Tremaine WJ, etal. The natural
history of fistulizing Crohn’sdisease in Olmsted County, Minnesota.
Gas-troenterology 2002;122:875–880.
69. Sandborn WJ, Feagan BG, Hanauer SB, etal. A review of
activity indices and efficacyendpoints for clinical trials of
medical ther-apy in adults with Crohn’s disease. Gastro-enterology
2002;122:512–530.
70. van Assche G, Vanbeckevoort D, Bielen D,et al. Magnetic
resonance imaging of the ef-fects of infliximab on perianal
fistulizingCrohn’s disease. Am J Gastroenterol 2003;98:332–339.
71. Buchanan GN, Bartram CI, Phillips RK, et al.Efficacy of
fibrin sealant in the managementof complex anal fistula: a
prospective trial.Dis Colon Rectum 2003;46:1167–1174.
72. Present DH, Rutgeerts P, Targan S, et al.Infliximab for the
treatment of fistulas in pa-tients with Crohn’s disease. N Engl J
Med1999;340:1398–1405.
73. Bell SJ, Halligan S, Windsor AC, WilliamsAB, Wiesel P, Kamm
A. Response of fistulat-ing Crohn’s disease to infliximab
treatmentassessed by magnetic resonance imaging. Al-iment Pharmacol
Ther 2003;17:387–393.
74. van Bodegraven AA, Sloots CE, Felt-BersmaRJ, Meuwissen SG.
Endosonographic evi-dence of persistence of Crohn’s
disease-as-sociated fistulas after infliximab
treatment,irrespective of clinical response. Dis ColonRectum
2002;45:39–46.
STATE OF THE ART: IMAGING OF FISTULA IN ANO Halligan and
Stoker
Radiology: Volume 239: Number 1—April 2006 33