The British Journal of Radiology, 81 (2008), 91-98

Pelvic phased-array MR imaging of anal carcinoma before and

after chemoradiation
^D M K O H , MRCP, FRCR, ^A DZIK-JURASZ, FRCR, PhD. O'NEILL, MRCP, ^ D TAIT, MD, FRCR, ^J E HUSBAND,
DBE, FRCR, FRCP a n d ^G B R O W N , M D , FRCR

^Academic Department of Radiology and ^Department of Radiation Oncology, Royal Marsden Hospital, Sutton, UK

ABSTRACT. The aim of this study was to evaluate the MR findings of anal carcinoma '
using an external pelvic phased-array coil before and after chemoradiation treatment.
15 patients with carcinoma of the anal canal underwent T2 weighted and short-tau
inversion recovery (STIR) imaging before and after chemoradiation. Images were
reviewed in consensus by two radiologists. At pre-treatment imaging, the tumour size
and stage, signal intensity and infiltration of adjacent structures were recorded. MR
imaging was repeated immediately after chemoradiation, every 6 months for the first '
year and then yearly. Tumour response was assessed by recording change in tumour |
size and signal intensity. Prior to treatment, the mean tumour size was 3.9 cm (range,
1.8-6.4 cm). Tumours appeared mildly hyperintense at T2 weighted and STIR imaging.
There was good agreement in T staging between clinical examination and MR imaging
(kappa-0.68). In 12 responders with long disease remission, a greater percentage
reduction in the size of MR signal abnormality in the tumour area was observed at 6
months (mean 54.7%; 46-62%) than immediately after treatment (mean 38.6%; 30- Received 17 November
46%) (p = 0.002, t-test). 7/12 showed stabilization of T2 signal reduction in the tumour 2006
Revised 12 April 2007
area after 1 year, and 5/12 showed complete resolution of signal alterations at 2 years. Accepted 30 April 2007
Pelvic phased-array MR imaging is useful for local staging of anal carcinoma and
assessing treatment response. After treatment, a decrease in tumour size accompanied DOI: 10.1259/bjr/96187638
by reduction and stability of the MR T2 signal characteristics at 1 year after
© 2008 The Bfitish Institute of
chemoradiation treatment was associated with favourable outcome.
Radiology

Carcinoma of the anal canal is uncommon, accounting the local extent of pelvic disease both for primary anal
for less than 2% of large bowel malignancies and 1-6'^ of cancers and for recurrent diseases [8]. Clear delineation
anorectal tumours [1]. Its incidence has been reported to of the anatomical boundaries of local disease enables
be approximately 0.5 per 100 000 in men and 1.0 per optimal planning of radiation fields. Following chemor-
100 000 in women [2]. adiation treatment, an understanding of the MR pattern
Anal carcinoma originates between the anorectal of tumour regression would allow careful non-invasive
junction above and the anal verge below. Not surpris- monitoring of treatment response. In addition, appre-
ingly, the majority of anal canal cancers are squamous ciation of the range of post-treatment appearances can
coll carcinoma [3]. Treatment using a combination of also aid in the early detection of disease relapse. To our
chemotherapy and radiotherapy is usually curative [4, 5]. knowledge, the application of MR imaging for mon-
However, radical surgery, such as abdomino-perincal itoring of effects of chemoradiation treatment has not
resection, may still be necessary to treat local failure or been previously reported in the published literature.
recurrence after chemoradiation treatment. Hence, the aim of this study was to evaluate the MR
Imaging performed prior to treatment provides assess- imaging findings of anal carcinoma using an external
ment of the local disease extent and nodal involvement, pelvic phased-array coil before and after chemoradia-
which is helpful towards treatment planning. Although tion treatment.
endoanal sonography has been used for local staging and
disease prognostication [6], the introduction of the endoanal
probe can be painful and the imaging field of view is limited. Methods and materials
Mesorectal lymph nodes located far from the endoanai
probe may be missed and mguinal nodes in the groin area The study was approved by our institution research
cannot be simultaneously assessed. Such limitations are also review board and ethics committee. Written informed
encountered by the use of endoanal MR imaging alone 17]. consent was obtained from all patients prior to the study.
More recently, MR imaging using an external pelvic
phased-array coil has been found useful for demonstrating
Patient characteristics
Address correspondence to: Dr D M Koh, Academic Department of
Radiology, Royal Marsden Hospital, Cancer Research UK Magnetic 15 consecutive patients with anal carcinoma were
Ri?yonance Group, institute of Cancer Research, Downs Road, prospectively evaluated. The inclusion criteria were (a)
Sutton, SM2 3PT. E-mai!: [email protected] biopsy-proven carcinoma of the anal canal, (b) patients

The British Journal of Radiology, February 2008 91

 D M Koh, A Dzik-Jurasz. B O'Neill et al

were selected for primary chemoradiation treatment and MR signal intensity, infiltration of adjacent structures
(c) all patients had MR imaging before and after and the presence of nodal disease.
chemoradiation treatment. These patients were diagnosed
with anal carcinoma between 1999 and 2001. Patients with Tumour size and stage. The maximum diameter of the
contraindications to MR imaging were excluded. tumour was measured in either the axial or coronal plane
There were seven men and eight women with a mean on the T2 weighted MR image to the nearest millimetre.
age of 66.2 years (range, 40-81 years). All patients were The maximum dimension was used to determine the
assessed clinically by digital examination at initial local tumour T stage as defined by the TNM staging
presentation by the primary physician and the tumour system [9] (Table 1). The estimated tumour size by
stage recorded according to the TNM classification clinical evaluation was correlated and compared with
system 19]. the MR assessment of tumour size.
Tumour sig}inl i)itcnsit\f. On TT weighted imaging and
STIR imaging, the signal of the tumour was visually
Chemoradiation treatment compared with the signal intensity of gluteus muscles
and recorded as predominantly hypointense, isointense
All patients received a standard regime of chemo- or hyperintense relative to the muscle signal intensity.
radiation comprising 43 Gy of radiotherapy in 25 fractions liifiltmlion of adjacent structures. Tumour extension to
with or without an additional 15 Gy in 6 fractions to the involve adjacent structures was evaluated on the T2
perianal region. 5-Fluorouracil (100 mg m~^) and mito- weighted MR images [10]. The presence or absence of
mycin G (7 mg m~^) was administered systemically in tumour involvement of the rectum, external anal sphinc-
combination with the radiotherapy. ter, puborectalis muscle, levator ani muscle, superficial
transverse perineal muscle, coccygeus muscle, ischiorectal
fossa, anterior urogenital triangle (vagina/prostate), retro-
MR imaging pubic space, perianal subcutaneous tissue and perianal
skin were assessed and recorded.
Baseline MR imaging was performed prior to treat- Nodal disease. A lymph node was considered malignant
ment on a 1.5 T MR system (Magnetom Vision, Siemens, if it measured greater than 5 mm in maximum short axis
Erlangen, Germariy) employing a pelvic phased-array diameter in the peri-rectal area or greater than 10 mm in
coil with the patient in the supine position. maximum short axis diameter over the inguinal region or
T] weighted (repetition time (TR) = 128 ms, echo time along the pelvic sidewall. Disease was classified using
(TE)^20ms, matrix=256x256, 300 cm field of view the TNM system.
(FOV), 6 mm thickness) and T2 weighted (TR
>4000 ms, TE=120 ms, matrix-256 x 256, 300 cm FOV,
6 mm thickness) images of the whole pelvis were first MR imaging following chemoradiotherapy
acquired. These were supplemented with small FOV T2 The MR images were assessed to evaluate the degree
weighted (TR >4000 ms, TE=120 ms, matrix=256 x 256, of tumour response by recording the following;
140 cm FOV, number of excitations (NEX)^4, 4 mm Tumour size. At each visit, the maximum diameter of
thickness) and short-tau inversion recovery (STIR) the tumour was measured on TT weighted imaging in
(TR=4890 ms, TE-60 ms, T, =150 ms, matrix-256 x 256, either the axial or coronal planes. Where no definite
140 cm FOV, NEX^2, 4 mm thickness) imaging in the tumour was identified at imaging, the maximum
axial and coronal planes at the level of the anal canal. diameter of any focal signal change within the anal
Axial imaging was performed perpendicular to and canal at T2 weighted imaging was recorded and charted.
coronal imaging acquired parallel to the long axis of the
anal canal. The total MR examination time was approxi- Signal intensity and appearance. The signal intensity of
mately 30 min. tumour or focal signal change on T2 weighted imaging was
recorded relative to the gluteus muscle. Any distortion in
MR examination was repeated using the same imaging the anal canal or sphincter complex was also noted.
protocol after completion of chemoradiation and 6 Infiltration of adjacent structure. Following chemoradia-
monthly afterwards for the first year after treatment. If tion, the degree of involvement of adjacent structures
there was no clinical or radiological evidence of disease was assessed at 6 months after treatment.
relapse, yearly surveillance MR imaging was undertaken Nodal disease. Regression or enlargement of malignant
for up to 3 years. nodes seen on pre-treatment imaging was recorded.

Image interpretation and analysis Statistical analysis

The MR images were reviewed in consensus by two The tumour size determined by clinical assessment
radiologists with more than 10 years' experience in was correlated with the measurements obtained by MR
pelvic MR imaging blinded to the clinical outcomes of imaging. The tumour size measurements were also
the patients. The pre-treatment and post-treatment MR compared using the paired f-test. The degree of agree-
images were reviewed sequentially. ment between the local T staging of anal cancer by MR
imaging and clinical assessment was determined by
Pre-treatment MR imaging kappa statistics.
Pre-treatment MR imaging evaluation of the anal The mean maximum diameter of the tumour before and
carcinoma was evaluated for tumour size and stage. following chemoradiation was compared using the paired

92 The British Journal of Radiology, February 2008

Pelvic phased-array MRI of anal carcinoma

Table 1. TNM staging of carcinoma of the anal canal

Primary tumour (T)

TO No evidence of primary tumour
Tl Tumour 2 cm or less in greatest dimension
T2 Tumour more than 2 cm but not more than 5 cm in greatest dimension
T3 Tumour more than 5 cm in greatest dimension
T4 Tumour of any size invading adjacent organ{s), e.g. vagina, urethra, bladder
Regional lymph nodes {N)
NO No regional lymph node metastasis
Nl Metastasis in perirectal lymph node
N2 Metastasis in unilateral internal iliac and/or inguinal lymph node
N3 Metastasis in internal iliac and perirectal lymph nodes and/or bilateral internal iliac
and/or bilateral inguinal lymph nodes
Distant metastasis (M)
MO No distant metastasis
Ml Distant metastasis

t-test. The percentage tumour size regression immediately resection (Patient 4). Another patient showed initial
after chemoradiation and at 6 months after chemoradia- response but had a significant residual mass after
tion was also compared using the paired t-test. completion of chemoradiotherapy. An abdomino-peri-
neal resection was performed but no viable tumour was
found at histopathology (Patient 3). One patient died of
Results disseminated disease at 12 months after completion of
chemoradiotherapy treatment, with persistent disease in
At histology, eight patients had well-differentiated the pelvis (Patient 8).
squamous cell carcinoma; four had moderately differ-
entiated squamous cell carcinoma; two had poorly
differentiated squamous cell carcinoma and one had
transitional basaloid cell carcinoma. Baseline MR imaging
Clinical follow-up information was available in all
patients for at least 3 years after chemoradiation Tumour size and stage
treatment (excluding patient who died). Of the 15 The mean maximum dimension of the tumour
patients, 11/15 had MR imaging follow up for 3 years measured on T2 weighted MR imaging was 3.9 cm
following chemoradiation treatment and 2/15 had MR (range 1,8-6.4). The mean maximum tumour size by
imaging for only 1 year because they were subsequently clinical assessment was 3.6 cm (range 1.0-7.0 cm). There
followed up elsewhere. Two patients underwent surgery was no significant difference in the maximum tumour
after completion of chemoradiation and did not undergo size as determined by MR imaging or clinical assessment
further MR examinations. (paired /-test, p>0.05). A good correlation was found in
12 patients showed a good response to chemoradiation the tumour size measurement by MR imaging and
and were alive and well at 3 years after treatment clinical assessment (r^O.76, p^O.001). On MR imaging,
(Table 2). One patient had a poor response to chemo- the tumour stage was found to be 7% Tl (1/15), 60% T2
mdiation and underwent abdomino-perineal surgical (9/15), 7% T3 (1/15) and 26% T4 (4/15). There was good

Table 2. Tumour size and signal characteristics at MR imaging before and after chemoradiation
Patient Sex Age Histology Size (cm) and signal intensity at Tj weighted MR imaging
years Pre- Immediate 6 months 1 year 2 years 3 years
treatment post-treatment
1 Male 58 Squamous cell 2.7a 1.6a 1.2— I.OT
2 Male 11 Squamous ceil 3.9a 3.2— 1.6T 1.2T 0 0
3 Male 63 Squamous cell 5.8a 3.8a
4 Female 67 Squamous cell 3.2(7 3.8a
5 Female 40 Basaloid cell 4.9a 2.2f7 1.4T 1.2T i.lT l.Or
6 Male 81 Squamous cell 1.8a 1.1(7 0.5T 0.51 0 0
7 Female 52 Squamous cell 2.8— 1.3ff 0.8T 0.8T 0 0
8 Male 51 Squamous cell 6.0a 5.3(7 5.8— 6.3(7
9 Female 53 Squamous cell 3.9a 2.5a 1.3T 1.0T I.IT l.Or
10 Female 69 Squamous cell 5.2a 26— 2.4T 2.2T 2.2T 2.0T
11 Female S7 Squamous cell 2.9a 1.4a I.IT 0 0 0
12 Female 62 Squamous cell 4.4a 2.8— 2.5T I.BT 1.3r l.Or
13 Female 66 Squamous cell 2.0a 1.5a I.OT I.IT 0 0
14 Male 73 Squamous cell 6.4(7 3.6— 3.2T 2.8T 2.9t 3.0T
15 Male 71 Squamous cell 4.5r7 3.4— 3.2— 3.4T 3.2T 3.2r
a, hyperintense; —, isointense; r, hypointense.

The British Journal of Radiology, February 2008 93

 D M Koh, A Dzik-Jurasz, B O'Neill et al

Table 3. Agreement between clinical staging and MR (Figure 1). The frequencies of involvement of adjacent
imaging staging of anai carcinoma structures following consensual assessment are tabulated
in Table 3. Tumour extension frequently involved the
MR staging Clinical staging Total
sphincter complex {6O'/'<>). The levator ani muscle was
Tl T2 T3 T4 involved in 40%. Extramural spread of disease was
usually anterior into the urogenital triangle (27%) to
Tl 1 0 0 0 1
T2 2 6 1 0 9 involve the vagina, bladder or urethra. There was
T3 0 0 1 0 1 inferior extension to involve the peri-anal subcutaneous
T4 0 0 2 2 4 tissue in 20%. Superior extension to involve the rectum
Total 3 6 4 2 15 and mesorectum occurred in 33% and 27% respectively.
Lateral spread of tumour into the ischiorectal fossa was
Weighted kappa=a.68 {p<O.Q^).
not observed in patients in our series (Figure 2).

Nodal disease
agreement in the T staging of the tumour by the MR Two patients had stage Nl disease within the
imaging and clinical assessment (weighted kappa=0.68, mesorectum. Two patients had stage N2 nodal disease
p<Q.O1) (Table 3). However, clinical examination along the left pelvic sidewall and inguinal area
appeared to under-stage two cases of T4 disease (Figure 3).
(invasion of urogenital structures) and two cases of stage
T2 disease determined by MR imaging.
MR imaging following chemoradiotherapy
Tumour signal intensity
14 out of the 15 tumours appeared hyperintense to Tumour size
gluteus muscle on both Tj weighted and STIR imaging. In the 12 patients who showed long-term remission
Tn the remaining case, the tumour was isointense to after chemoradiation, there was initially a 38.6% mean
muscle on Tj weighted and STIR imaging, and was reduction (95% confidence interval (CI) 30-46"/.) in the
recognized by distortion of the anal canal. In seven maximum diameter of the signal abnormality on T2
patients, heterogeneity of signal intensity within the weighted imaging at the tumour site immediately after
tumour was visible on both Tz weighted and STIR chemoradiation. However, a greater 54.7% mean reduc-
imaging. tion in the maximum diameter of the tumour (95% CI:
46-62%) was observed at 6 months after chemoradiation
Infiltration of adjacent structures compared with the baseline imaging (;)^0.002, paired i-
Tumours were better delineated with respect to test). There was stabilization in the size of the residual
adjacent structures on T2 weighted imaging since STIR change by 1 year after treatment in 7/12 (58%) patients
imaging diminished the contrast between soft tissues, (Figure 4). The remaining 5/12 (42%) showed complete
making it more difficult to define anatomical boundaries regression of any signal change and no measurable

(ai

Figure 1. (a) T2 weighted and (b) STIR images of a stage T4 anal canal carcinoma. Note the mass arising from the right anal
canal, which is mildly hyperintense compared with the gluteus muscle. The mass is infiltrating into the external sphincter (*) and
puborectalis muscle (arrowhead). There is also tumour extension anteriorly involving the posterior vaginal wall (white arrows).
Anatomical structures are better delineated on the T2 weighted image than on the STIR image.

94 The British Journal of Radiology, February 2008

Pelvic phased-array MRI of anal carcinoma

Figure 2. Local tumour extension. T2 weighted (a) axial and (b) coronal images showing a carcinoma arising from the right side
of the anal canal. There is anterior tumour extension which is inseparable from the vagina (arrow). On the coronal image (b),
note the lateral extension of tumour involving the right levator ani, puborectalis and external sphincter (arrow), but there is ncj
further infiltration into the right ischiorectal fossa. Note also the superior tumour extension to involve the distal rectum (*).

abnormality was observed at 2 years after treatment. One a reduchcm in the tumour size and associated signal
patient did not respond to treatment and showed an intensity. At 6 months following chemoradiation, the
increase in size of the tumour during treatmer\t. Another degree of involvement of adjacent structures among the
showed 4O''ii reduction in size of the primary tumour 13 patients with follow-up MR imaging is summarized in
after chemoradiation, but substantial residual abnorm- Table 4.
ality remained which prompted surgery. Disease pro-
gression was encountered in one patient who developed
Nodal disease
metastatic disease and persistent disease in the pelvis.
There was regression of all malignant nodes visualized
at MR imaging prior to treatment. No suspicious nodes
Signal characteristic and appearance were detected after chemoradiation.
Immediately following chemoradiation, 7/12 patients
showed hyperintensity and 5/12 patients showed iso-
intensity at the tumour site at T^ weighted MR imaging.
At 6 months, 10/12 responders showed hypointensity at
the site of tumour, but 2/12 showed isointensity. There
was, however, variable hypointensity at the site of
treated tumour at 1 year following chemoradiation.
This was accompanied by distortion of the anal canal
in 33%, which may be attributed to fibrosis (Figure 5).
Variable hypointense signal intensity persisted
unchanged in 7/12 {5S%) but complete resolution of
the altered signal intensity occurred in 5/12 (42'/.>) at 2
years, paralleling the reduction in size of the signal
abnormality. In the three patients who relapsed or
showed disease progression, there was persistent hyper-
intensity at the site of the tumour.

infiitration of adjacent structures

Following chemoradiation, there was a reduction in Figure 3. T2 weighted MR image showing a large 4 cm left
the apparent infiltration of adjacent structures, based on internal Iliac nodal mass (arrow).

The British Journal of Radiology, February 2008 95

 D M Koh, A Dzik-Jurasz, B O'Neiii et al

(cm) Pre- Post- 6 months 1 year 7 years 3 years Figure 4. Change in size of the
treatment treatment visible abnormality on V^ weighted
MR imaging at the site of the tumour
Hme in relation to chemoradiation treatment before and after chemoradiation.

Discussion iscbiorectai fat was not seen in our study population,

suggesting that the levator ani muscle may act as a
Squamous cell carcinoma of the anus is an uncommon relative barrier to lateral tumour growth.
cancer but is eminently curable using a combination of Following cbemoradiation treatment, the majority of
chemotherapy and radiotherapy [4]. MR imaging v^-ith a patients (12/15) in our study showed a good response to
pelvic phased-array coil using TT weighted and STTR treatment. Tbis was' seen as a reduction in tbe tumour
imaging is useful for identifying tumours within the anal size accompanied by signal intensity change at T2
canal and can be used to define the extent of disease [8] weighted MR imaging. Interestingly, in the 12 patients
prior to non-invasive therapy. who responded favourably to cbemoradiation treatment,
Prior to treatment, anal carcinomas appeared hyper- tbere was greater size involution observed on MR
intense on both T2 weighted and STIR imaging. imaging at 6 months following chemoradiation than on
However, Tj weighted imaging was better in demon- imaging performed soon after completion of treatment.
strating the relationship of the tumour to adjacent This observation was perhaps not surprising since
structures since there was greater homogenization of radiotherapy could provoke inflammatory reactions that
the soft tissue signal intensity on STIR imaging, making were superimposed on the treated disease and may
it more difficult to delineate normal anatomical bound- regress slowly.
aries. Thus, we would recommend the use of T2 In tbe 12 patients who bad sustained remission (>2
weighted MR imaging rather than STIR imaging for the years) following cbemoradiation, the majority (58%)
evaluation of anal carcinoma. By performing T2 showed reduction and stabilization in tbe T2 signal
weighted imaging in the axial and coronal planes, the intensity and tbe size of any residual change at 1 year
anal canal can be practically assessed in less than 30 min. after chemoradiation. However, continued and complete
In our study, we found good agreement (kappa=0.68, regression of any residual changes occurred in the
;?<0.01) in the local staging of tumour between clinical remaining 42% at 2 years following chemoradiation.
examination and MR imaging. Interestingly, there was Hence, it would appear that stabilization of any visible
no significant difference in the assessment of tumour size residual abnormality more than 1 year after chemoradia-
by both methods. This may reflect the relative superficial tion was associated with a favourable outcome. This is
location of these tumours, making them amenable to potentially important since criteria for non-invasive
direct measurements. However, clinical examination assessment of disease response using MR imaging have
under-staged four cases compared with MR imaging. not been previously described. However, these observa-
Although this did not affect treatment planning in our tions were made in a relatively small number of patients
patients, if is conceivable that, with tbe increasing use of and their clinical utility needs to be further verified in
conformal radiotherapy, accurate delineation of local future studies.
disease would be important for defining radiotherapy Following chemoradiation, we found that the reduc-
boundaries in the future. tion in the size of tumour appeared to parallel the
TT weighted MR imaging was able to define local reduction in signal intensity on T2 weighted MR
tumour infiltration by visualizing tumour signal inten- imaging. Extrapolating from observations made in rectal
sity extending into adjacent structures. In our study there cancer following chemoradiation, tbe appearance of low
was frequent involvement of the sphincter complex signal intensify with the treated anal canal on T2
(external sphincter, levator ani muscle and the pubo- weighted MR imaging is likely to represent fibrosis
rectalis muscle). In addition, tumour extension into the [10]. However, it would be impossible for MR to detect
urogenital triangle was common (27%) and this may be foci of microscopic disease witbin these fibrotic tissues,
explained by tbe relatively thin anatomical separation and follow-up imaging is thus important for ensuring
between the anal canal and the anterior urogenital stability of appearance, and for the detection of early
triangle. Interestingly, tumour extension into the relapse.

96 The British Journal of Radiology, February 2008

Pelvic phased-array MRI of anal carcinoma

Figure 5. Response to chemoradiotherapy. (a) Pre-treatment T^ weighted MR image showing hyperintense irreguiar thickening
of the iower anai canai with extension into the right externai sphincter. The tumour was stage T2 on MR imaging, (b)
immediately foliowing chemoradiation, residual isointense signal intensity was noted at the site of previous tumour, (c) At 1
year foliowing radiotherapy, there was further regression in the size of the abnormality accompanied by reduction in the signal
intensity on T2 weighted MR imaging, (d) This appearance was unchanged on T2 weighted MR imaging obtained 2 years after
treatment consistent with post-treatment fibrosis.

Currently, the routine use of MR imaging for the acute drug toxicity using combination chemoradiation.
management of anal cancer is still not widely practised However, MR imaging has not been specified as a
[II]. Treatment and follow-up is still predominantly method for the assessment of tumour stage or tumour
reliant on clinical evaluation of local disease. The ACT II response. Based on our study findings, we believe that
Trial is a randomized phase III clinical trial that is MR imaging should be considered as an imaging tool for
currently in progress for patients with carcinoma of the the assessment of tumour stage prior to treatment and
anal canal or margin. The objective of this trial is to for the assessment of treatment effects. MR imaging may
improve complete response rates and recurrence-free also facilitate more accurate radiotherapy delivery by
survival without significantly increasing the rates of providing detailed maps of local disease extent.

The British Journal of Radiology, February 2008 97

 D M Koh. A Dzik-Jurasz, B O'Neill et al

Table 4. Assessment of infiltration of adjacent structures on MR imaging

Tumour involvement on MRI Frequency of observations

Pre-treatment (%) 6 months after chemoradiation

Levator ani muscle 6/15 (40) 1/13 (8)
External anal sphincter 9/15 (60) 2/13 (15)
Puborectalis muscle 6/15 (40) 1/13 (8)
Coccygeus muscle 5/15(33) 0/13 (0)
Transverse perineal muscle 2/15 (13) 0/13 (0)
Rectum 5/15(33) 1/13 (8)
Mesorectum 4/15 (27) 1/13 (8)
Urogenitai triangle (bladder, prostate or 4/15 (27) 1/13 (8)
vagina)
Subcutaneous perianal tissue 3/15 (20) 0/13 (0)

There are a few limitations to the current study. Firstly, chemoradiation, a reduction in tumour size accompanied
this study was conducted in a small population and it is by reduction and stabilization of the T2 signal intensity at
uncertain as to what degree our findings can be the site of tumour at 1 year after treatment were
generalized. However, because carcinoma of the anal associated with a favourable outcome.
canal is rare, it would be difficult to perform a larger
prospective study without involving multiple clinical
centres. The ideal setting for future studies would be to References
apply MR imaging in future multicentre clinical trials in
anal cancer, which would allow the technique to be 1. Deans GT, McAleer JJ, Spence RA. Malignant nnal tumours.
widely evaluated. Secondly, the cielineation of local BrJSurg 1994;81:50n-8.
disease extent and the regression of tumour after 2. Frisch M. C>n the etiology of anal squamous carcinoma. Dan
treatment were based on imaging and clinical findings Med Bull 2002;49:194-209.
3. Williams GR, Talbot IC. Anal carcinoma - a histological
without histopathological confirmation. However, sam-
review. Histopathology 1994;23:507-16.
pling of the tumour area following treatment is not 4. UKCCCR Anal Cancer Trial Working Party. UK Co-
performed routinely and it may be difficult to justify ordinating Committee on Cancer Research. Epidermoid
repeated invasive biopsies. Furthermore, even if biopsies anal cancer: results from the UKCCCR randomised trial of
were taken, these would still be subject to sampling radiotherapy alone versus radiotherapy, .S-fluorouracil, and
error. Thirdly, the MR classification of tumour appear- mitomycin. Lancet 1996;348:!049-.S4.
ance was made according to the visual inspection of the 3. Ryan DP, Mayer RJ. Anal carcinoma: histology, staging,
predominant signal intensity. However, as most tumours epidemiology, treatment. Curr Opin Oncol 2000;! 2:345-52.
show some degree of heterogeneity, it was difficult to 6. Giovannini M, Bardou VJ, Barclay K, Palazzo L, Roseau G,
adequately reflect this by current assessment. Helbert T, et al. Anal carcinoma: prognostic value of
Nevertheless, we found that, by comparing with the endorectai ultrasound (ERUS). Results of a prospective
baseline imaging, a reduction in tumour size, accom- multicenter study. Endoscopy 2001;33:231-6.
panied by a reduction or stability of the MR T2 signal 7. de Souza NM. Current and future perspectives on the role
of internal coils for magnetic resonance imaging of the anal
characteristics at 1 year after chemoradiation treatment, sphincter and rectum. Abdom Imaging 2002;27:432-41.
was associated with a favourable outcome. However, 8. Roach SC, Hulse PA, Moulding F], Wilson R, Carrington
larger prospective studies in the future would be useful BM. Magnetic resonance imaging of anal cancer. Clin
to ascertain the predictive value of the MR observations. Radiol 2005;60:ini-19.
9. Sobin LH, Wittekind C, eds. TNM classification of
malignant tumours (UICCS). Houston: John Wiley and
Sons, 2002.
Conclusions 10 Stollfuss JC, Becker K, Sendler A, Seidl S, Settles M, Auer F,
et al. Rectal carcinoma: high-spatial-resolution MR imaging
Pelvic phased-array MR imaging was useful for the and TT quantification in rectal cancer specimens. Radiology
local staging of anal carcinoma and for the monitoring
of treatment response. Prior to treatment, there was 11. Indinnimeo M, Cicchini C, Stazi A, Ghini C, Mingazzini P,
good agreement between clinical examination and MR Laghi A. Analysis of a follow-up program for anal canal
imaging for the local staging of tumour. Following carcinoma. J Exp Clin Cancer Res 2001 ;20:199-203.

98 The British Journal of Radiology, February 2008