Diffusion-weighted imaging complements T2-weighted MRI for tumour response assessment in squamous anal carcinoma

Abstract

Objectives:
A published tumour regression grade (TRG) score for squamous anal carcinoma treated with definitive chemoradiotherapy based on T2-weighted MRI yields a high proportion of indeterminate responses (TRG-3). We investigate whether the addition of diffusion-weighted imaging (DWI) improves tumour response assessment in the early post treatment period.

Materials and methods:
This retrospective observational study included squamous anal carcinoma patients undergoing MRI before and within 3 months of completing chemoradiotherapy from 2009 to 2020. Four independent radiologists (1-20 years' experience) scored MRI studies using a 5-point TRG system (1 = complete response; 5 = no response) based on T2-weighted sequences alone, and then after a 12-week washout period, using a 5-point DWI-TRG system based on T2-weighted and DWI. Scoring confidence was recorded on a 5-point scale (1 = low; 5 = high) for each reading and compared using the Wilcoxon test. Indeterminate scores (TRG-3) from each reading session were compared using the McNemar test. Interobserver agreement was assessed using kappa statistics.

Results:
Eighty-five patients were included (mean age, 59 years ± 12 [SD]; 55 women). T2-weighted TRG-3 scores from all readers combined halved from 24% (82/340) to 12% (41/340) with DWI (p < 0.001). TRG-3 scores changed most frequently (41%, 34/82) to DWI-TRG-2 (excellent response). Complete tumour response was recorded clinically in 77/85 patients (91%). Scoring confidence increased using DWI (p < 0.001), with scores of 4 or 5 in 84% (287/340). Interobserver agreement remained fair to moderate (kappa range, 0.28-0.58).

Conclusion:
DWI complements T2-weighted MRI by reducing the number of indeterminate tumour responses (TRG-3). DWI increases radiologist's scoring confidence.

Clinical relevance statement:
Diffusion-weighted imaging improves T2-weighted tumour response assessment in squamous anal cancer, halving the number of indeterminate responses in the early post treatment period, and increases radiologists' confidence.

Key points:
Tumour response based on T2-weighted MRI is often indeterminate in squamous anal carcinoma. Diffusion-weighted imaging alongside T2-weighted MRI halved indeterminate tumour regression grade scores assigned by four radiologists from 24 to 12%. Scoring confidence of expert and non-expert radiologists increased with the inclusion of diffusion-weighted imaging.

Full text

Vol.:(0123456789)
1 3
European Radiology
https://doi.org/10.1007/s00330-023-09942-0
GASTROINTESTINAL
Diffusion‑weighted imaging complements T2‑weighted MRI
fortumour response assessment insquamous anal carcinoma
DavidePrezzi1,2 · KeerthiniMuthuswamy2· AshikAmlani2· KasiaOwczarczyk2· AhmedElowaidy2· TinaMistry2·
PaulBassett3· VickyGoh1,2
© The Author(s) 2023
Abstract
Objectives A published tumour regression grade (TRG) score for squamous anal carcinoma treated with definitive chemora-
diotherapy based on T2-weighted MRI yields a high proportion of indeterminate responses (TRG-3). We investigate whether
the addition of diffusion-weighted imaging (DWI) improves tumour response assessment in the early post treatment period.
Materials and methods This retrospective observational study included squamous anal carcinoma patients undergoing MRI
before and within 3 months of completing chemoradiotherapy from 2009 to 2020. Four independent radiologists (1–20 years’
experience) scored MRI studies using a 5-point TRG system (1 = complete response; 5 = no response) based on T2-weighted
sequences alone, and then after a 12-week washout period, using a 5-point DWI-TRG system based on T2-weighted and DWI.
Scoring confidence was recorded on a 5-point scale (1 = low; 5 = high) for each reading and compared using the Wilcoxon
test. Indeterminate scores (TRG-3) from each reading session were compared using the McNemar test. Interobserver agree-
ment was assessed using kappa statistics.
Results Eighty-five patients were included (mean age, 59 years ± 12 [SD]; 55 women). T2-weighted TRG-3 scores from all
readers combined halved from 24% (82/340) to 12% (41/340) with DWI (p < 0.001). TRG-3 scores changed most frequently
(41%, 34/82) to DWI-TRG-2 (excellent response). Complete tumour response was recorded clinically in 77/85 patients
(91%). Scoring confidence increased using DWI (p < 0.001), with scores of 4 or 5 in 84% (287/340). Interobserver agree-
ment remained fair to moderate (kappa range, 0.28–0.58).
Conclusion DWI complements T2-weighted MRI by reducing the number of indeterminate tumour responses (TRG-3). DWI
increases radiologist’s scoring confidence.
Clinical relevance statement Diffusion-weighted imaging improves T2-weighted tumour response assessment in squamous
anal cancer, halving the number of indeterminate responses in the early post treatment period, and increases radiologists’
confidence.
Key Points
• Tumour response based on T2-weighted MRI is often indeterminate in squamous anal carcinoma.
• Diffusion-weighted imaging alongside T2-weighted MRI halved indeterminate tumour regression grade scores assigned
by four radiologists from 24 to 12%.
• Scoring confidence of expert and non-expert radiologists increased with the inclusion of diffusion-weighted imaging.
Keywords Magnetic resonance imaging· Diffusion magnetic resonance imaging· Anus neoplasms·
Carcinoma, squamous cell· Chemoradiotherapy
Abbreviations
CR Complete response
DWI Diffusion-weighted imaging
LRec Local recurrence
PD Progressive disease
PR Partial response
TRG Tumour regression grade
* Davide Prezzi
davide.prezzi@kcl.ac.uk
1 School ofBiomedical Engineering andImaging Sciences,
King’s College London, King’s Health Partners, London, UK
2 Department ofRadiology, Guy’s andSt Thomas’ NHS
Foundation Trust, London, UK
3 Statsconsultancy Ltd., Amersham, UK

European Radiology
1 3
Introduction
Squamous carcinoma of the anal canal is on the rise worldwide,
with an annual incidence of 0.5–2 in 100,000 [1]. Definitive
radiotherapy with concomitant mitomycin C and 5-fluoro-
uracil (or capecitabine) is the therapy of choice for localised
disease, with good outcomes [2, 3]. Timely identification
of locoregional treatment failure, occurring in a minority of
cases (11–14%) [4, 5], allows these patients to be considered
for salvage surgery, which in turn leads to local pelvic control
in approximately 60% of cases and to a 5-year survival rate of
30–60% [2]. Early detection of salvageable local disease relapse
during imaging response assessment and surveillance is key.
MRI is recommended for locoregional staging and
response assessment [2, 3, 6, 7], and has a growing role in
radiotherapy planning [8]. High-resolution T2-weighted
sequences are typically obtained in planes parallel and per-
pendicular to the anal canal. MRI tumour response assess-
ment based on T2-weighted sequences can be challenging in
the early post treatment period, due to the overlapping fea-
tures between therapy-induced inflammation (hyperintense
tissue oedema mixed with hypointense fibrosis, anatomical
distortion) and residual tumour, typically intermediate in
signal [9]. A 5-point MRI tumour regression grade (TRG)
system based on T2-weighted sequences has been utilised to
classify squamous anal carcinoma tumour response to chem-
oradiotherapy [10]. In a prospective single-centre cohort, the
number of indeterminate TRG scores (TRG-3) was consider-
able in the early post treatment period, corresponding to 58%
of the total at 3 months and to 26% of the total at 6 months
post chemoradiotherapy, emphasising the problem posed by
treatment-related inflammation.
Diffusion-weighted imaging (DWI) is used routinely to aid
the assessment of a variety of abdominal malignancies [11, 12].
Specifically, in rectal cancer treated with neoadjuvant chemo-
radiotherapy, DWI increases diagnostic accuracy in the evalu-
ation of complete response [13] and detection of small-volume
residual tumour before endoscopy [14]. We hypothesised that
DWI could benefit early tumour response evaluation in squa-
mous anal carcinoma and improve the diagnostic confidence of
non-expert radiologists. Thus, the primary aim of this study was
to determine whether DWI improves tumour response assess-
ment by reducing indeterminate responses (TRG-3) in the early
post treatment period. Secondary aims were to assess the impact
of DWI on subjective TRG scoring confidence and interobserver
agreement amongst expert and non-expert radiologists.
Materials andmethods
Patients
Institutional board waiver of informed consent was obtained
for this retrospective study of consecutive MRI data
obtained as part of the standard care pathway. Patients with
biopsy-proven squamous cell carcinoma of the anal canal
undergoing treatment with definitive chemoradiotherapy
between February 2009 and May 2020 were identified from
the Picture Archiving and Communication System (PACS)
and electronic patient record (EPR) of a tertiary care institu-
tion (Guy’s and St Thomas’ NHS Foundation Trust). Inclu-
sion criteria were baseline and post treatment MRI (within
3 months of treatment completion) available from PACS;
TNM 8th ed. T2 stage; or greater tumours [15], equivalent
to tumour diameter >2 cm. Exclusion criteria were absence
of DWI on baseline or post treatment MRI; DWI of insuffi-
cient diagnostic quality; no visible tumour on baseline MRI;
and prior tumour surgical excision. The patient flowchart
is shown in Fig.1.
Treatment andclinical follow‑up
Radiotherapy was delivered to a mean dose of 50.86 Gy
(range 50.4–54 Gy) using a linear accelerator (Elekta or Var-
ian) applying a 3D conformal or intensity-modulated tech-
nique. Concomitant chemotherapy consisted of mitomycin C
(12 mg/m2 on day 1) with either 5-fluorouracil (1000 mg/m2/
day, continuous venous infusion, on days 1–4 and 29–32) or
capecitabine orally (825 mg/m2 twice a day on radiation days).
Following completion of chemoradiotherapy, patients
were clinically assessed at 8–10 weeks, then every 3 months
for the first 2 years, and every 6 months afterwards as per
standard institutional practice. Endoscopic evaluation ±
MRI/CT was undertaken if canal ± locoregional/distant
recurrence was suspected clinically.
MRI acquisition
Patients were scanned supine on one of five 1.5- or 3.0-
Tesla MRI scanners (Magnetom Avanto, Aera or Skyra,
Siemens Healthineers) using an 18-channel pelvic phased
array coil. The examination protocol included axial and
sagittal T2-weighted turbo spin-echo (TSE) sequences
covering the whole pelvis, and high-resolution small-
field-of-view T2-weighted TSE sequences perpendicu-
lar and parallel to the anal canal. DWI consisted of a
single-shot spin-echo echo planar imaging (EPI) axial
diffusion-weighted sequence encompassing the pelvis
with three b-values (0, 100, 800 s/mm2). Apparent dif-
fusion coefficient (ADC0-800) maps and calculated high
b-value images (b = 1400 or 1600 s/mm2) were created
automatically at the time of acquisition. Patients did not
undergo any additional preparation prior to the examina-
tion. Typical acquisition parameters are summarised in
Supplemental Table1.

European Radiology
1 3
Image analysis
Baseline and post treatment MRI were evaluated sequen-
tially by four independent observers, blinded to clinical
outcome: two senior radiology residents (K.M., A.A.), with
1 year’s experience in oncologic pelvic MRI but no specific
experience in staging squamous anal carcinoma (referred
to as ‘non-expert observers’) and two subspecialist radiolo-
gists (D.P., V.G.) with over 10 years’ experience (‘expert
observers’).
Anonymised scans were presented in a randomised
order. Post treatment MRIs were assessed next to
Pa ts treatedwithdefini ve
chemoradiotherapy forsquamousanal
carcinomabetween February 2009and
May2020(n=218)
Pa ts with ≥T2 disease undergoing baseline
andposttreatment MRI (n=111)
Finalpant cohort (n=85)
Excluded
1) DWInot acquiredon
baseline or post
treatmentMRI (n=14)
2) Insufficient qualityDWI
(n=3)
3) No assessable tumour on
baseline MRI(n=4)
4) Surgical tumour excision
(n=1)
5) Post treatmentMRI me
frameout of inclusion
criteria (n=4)
Fig. 1 Patient flowchart
Table 1 Summary of tumour regression grade (TRG) systems applied
Grade Description
A. Published TRG system based on T2-weighted MRI only
1 Complete response with no evidence of tumour and normal appearances of the anus
2 Excellent response with only low signal post treatment fibrotic change and no
evidence of tumour
3 Moderate response with indeterminate heterogeneous signal intensity at the tumour
site
4 Minimal response with reduction in size but evidence of intermediate tumour signal
in keeping with residual disease
5 No response of the primary tumour or frank tumour progression
B. Proposed DWI-TRG system based on T2-weighted plus diffusion-weighted MRI
1 Complete response with no evidence of tumour and normal appearances of the anus
2 Excellent response with low T2 signal fibrosis only. No areas of diffusion restric-
tion, excluding linear restriction spatially matched to anorectal mucosa
3 Indeterminate response with heterogeneous T2 signal intensity at the tumour site.
Indeterminate non-linear diffusion restriction or linear restriction, not spatially
matched to anorectal mucosa
4 Minimal response with reduction in size but persistent intermediate T2 tumour
signal matched by restricted diffusion
5 No response of the primary tumour or frank tumour progression

European Radiology
1 3
baseline MRIs on dual monitors (Sectra IDS7 worksta-
tions, Sectra AB). The first reading session was based
on multiplanar T2-weighted sequences alone. Tumour
response was evaluated according to a previously pub-
lished 5-point tumour regression grade (TRG) score
[10], outlined in Table1. In addition to TRG, the fol-
lowing primary tumour characteristics were recorded:
size (maximum diameter in any plane); location (lower
canal, mid canal, upper canal/anorectum); invasion of
adjacent structures (prostate, vagina, ischioanal fossa).
The second reading session, separated from the first by
a 12-week wash out period, included T2-weighted and
DWI assessed in conjunction. Multiplanar T2 sequences,
acquired/calculated b-value images and corresponding
ADC maps were displayed simultaneously. DWI image
quality was scored subjectively as inadequate, adequate
or good by each observer, documenting the nature of
image degradation as free text, when present. Tumour
response was re-scored according to a modified DWI-
TRG system, outlined in Table1. Post treatment DWI
images were regarded positive for residual disease when
diffusion restriction (signal hyperintensity on high
b-value images matched by hypointensity on ADC map)
remained present at the site of the tumour, excluding
linear diffusion restriction spatially matched to anorectal
mucosa. A single-slice, free-hand region of interest was
drawn around the tumour on ADC maps, with reference
to the corresponding DWI and T2-weighted images, and
the mean tumour ADC value recorded for baseline MRI.
During both reading sessions, observers scored their
subjective confidence in assessing tumour response on a
scale of 1 (low) to 5 (high).
Statistical analysis
Statistical analyses were performed by a senior statisti-
cian (P.B.) using Stata (v15.1; StataCorp LP). Normally
distributed variables were expressed as mean ± stand-
ard deviation. Categorical variables were expressed as
absolute numbers and their percentages. The McNemar
test was used to compare the number of indeterminate
TRG vs. other TRG scores between reading sessions.
Interobserver agreement was assessed using the kappa
statistics (kappa < 0.21 = poor agreement; 0.21–0.40 =
fair; 0.41–0.60 = moderate; 0.61–0.80 = good; > 0.80 =
excellent). Kappa values and their standard errors were
used to perform a z-test to compare the level of agreement
between reading sessions. The Wilcoxon matched-pairs
test was used to compare observer confidence scores.
Analyses were performed for each observer separately,
and for all observers combined. A p value <.05 was taken
to represent statistical significance.
Results
Patients andclinical response
Baseline patient characteristics are summarised in
Table2. The final cohort consisted of 85 patients, rang-
ing in age between 34 and 86 years (mean, 59 years ±
12 [SD]; 55 women). Mean tumour size was 5.1 ± 2.1
cm. A large proportion of patients had locally advanced
disease at baseline (53%, 45/85), defined as T3 stage or
greater, and/or tumours located in the upper canal/anorec-
tum (52%, 44/85). Mean clinical follow-up duration was
32 ± 18 months. Clinical disease recurrence was recorded
in 36% (31/85) of patients: local, in 18% (15); nodal, in
9% (8); metastatic, in 14% (12). Clinical complete tumour
response was recorded in 91% (77/85) of patients at
8–10 weeks from the end of treatment. There was partial
response in 6% (5/85). Progressive disease was recorded
in 4% (3/85). Local recurrence was documented in 8%
Table 2 Summary of patient and tumour characteristics. Note. Values
are mean ± standard deviation or number of participants with per-
centage of total in parentheses
Characteristic Patients (n = 85)
Sex
Female 55 (65)
Male 30 (35)
Age (year) 59 ± 12
Body mass index (kg/m2) 25.4 ± 5.6
TNM 8th ed. stage
T2 40 (47)
T3 21 (25)
T4 24 (28)
N0 27 (32)
N1 58 (68)
M0 83 (98)
M1 2 (2)
Tumour size (cm) 5.1 ± 2.1
Tumour location
Lower canal 28 (33)
Mid canal 13 (15)
Upper canal/anorectum 44 (52)
Tumour invasion of adjacent structures
Prostate 7 (8)
Vagina 17 (20)
Ischioanal fossa 6 (7)
Tumour ADC (mm2/s × 10−3) 0.910 ± 0.182
Time interval between baseline and post treatment
MRI (days)
168 ± 24
Time interval between end of treatment and post
treatment MRI (days)
77 ± 17

European Radiology
1 3
(7/85) of patients: 5% (4), corresponding to late recur-
rence, beyond 12 months from the end of treatment; 4%
(3), corresponding to microscopic subclinical recurrence
at 6–9 months from the end of treatment. Four patients
underwent salvage surgery by means of abdominoperineal
excision of the rectum.
MRI andtumour regression grade
DWI image quality was scored as inadequate in 3 cases,
which were excluded; adequate in 26% (22/85); and good
in 74% (63/85). The most common problems affecting DWI
quality were low signal-to-noise ratio and susceptibility
artefacts, particularly from air/gas at the anal margin and
in the rectal lumen. Calculated b-value (b1400 or b1600 s/
mm2) images were available in 70/85 (82%) cases. Baseline
tumour conspicuity on b800 images and ADC maps was
high in all included cases, with mean tumour ADC values
of 0.910 ± 0.182 × 10−3 mm2/s.
TRG scores from the four radiologists are shown in
Fig.2. With the inclusion of DWI, the number of indeter-
minate TRG-3 scores decreased significantly for three of the
four radiologists examined individually (difference range,
11–19%; p range, < 0.001–0.04), and for all radiologists
combined (difference, 12%; p < 0.001) (Table3). For all
observers combined, the number of TRG-3 cases halved
from 24% (82/340) of the total based on T2-weighted MRI
alone to 12% (41/340) based on T2-weighted plus DWI.
Indeterminate TRG-3 scores changed most frequently to
DWI-TRG-2 (41%, 34/82), corresponding to excellent
response (Figs.3 and 4); 9% (7/82) changed to DWI-TRG-4,
corresponding to minimal response (Fig.5). The remaining
50% of TRG-3 (41/82) corresponded to indeterminate DWI-
TRG-3 scores (Fig.6).
Subjective TRG scoring confidence
Observers’ confidence in assessing response increased with
the addition of DWI. Scores were higher for each of the
four observers (p < 0.001), and for all observers combined
(Supplemental Table2). For all observers combined, 84%
(287/340) of confidence scores were 4 or 5 for T2-weighted
plus DWI, compared to 55% (188/340) for T2-weighted MRI
alone.
0
10
20
30
40
50
60
70
80
12345
No. of Cases
TRG
T2
Observer 1Observer 2Observer 3Observer 4
0
10
20
30
40
50
60
70
80
12345
No. of Cases
DWI-TRG
T2 + DWI
Observer 1Observer 2Observer 3Observer 4
Fig. 2 Tumour regression grade (TRG) scores amongst observers. Note. Observers 1 and 2 are expert observers. Observers 3 and 4 are non-
expert observers
Table 3 Change in the number
of indeterminate tumour
responses between T2-weighted
MRI alone (TRG-3) and
T2-weighted MRI plus DWI
(DWI-TRG-3) per observer
*Data are number of cases over total, with percentage of total in parentheses
#Differences calculated as value for T2+DWI minus value for T2
Observers 1 and 2 are expert observers. Observers 3 and 4 are non-expert observers
T2* T2 + DWI* Difference # (95% CI) p value
Observer 1 25/85 (29%) 15/85 (18%) −12% (−21%, −3%) 0.003
Observer 2 21/85 (25%) 5/85 (6%) −19% (−28%, −9%) < 0.001
Observer 3 19/85 (22%) 10/85 (12%) −11% (−20%, −1%) 0.04
Observer 4 17/85 (20%) 11/85 (13%) −7% (−18%, 4%) 0.16
All combined 82/340 (24%) 41/340 (12%) −12% (−17%, −7%) < 0.001

European Radiology
1 3
Interobserver agreement
Interobserver agreement was between fair and moderate
(Supplemental Table3), with kappa values ranging between
0.28 and 0.58. The highest agreement was achieved by non-
expert observers assessing response on T2-weighted plus
DWI. No significant differences in interobserver agreement
were found between the two response assessment methods
(p = 0.16–0.40).
Correspondence betweenMRI TRG andclinical
tumour response
Correspondence between MRI TRG scores and clini-
cal tumour response is summarised in Table4. Patients
with complete clinical response (n = 77) were assigned
TRG scores of 1 or 2 in 88% (270/308) of cases using
T2-weighted plus DWI, versus 73% (226/308) based on
T2-weighted MRI alone. Patients with partial clinical
response (n = 5) were assigned scores of 3 or above in
75% (15/20) and 65% (13/20) of cases respectively.
Eleven out of 12 TRG scores in 3 patients with progres-
sive disease were ‘4’ or ‘5’. In 7 patients with documented
local recurrence after an initial complete response, 64%
(18/28) TRG-2 and 93% (26/28) DWI-TRG-2 scores were
recorded, in line with complete responders without sub-
sequent recurrence.
Discussion
Assessing tumour response in the initial period following
chemoradiotherapy can be challenging in patients with
squamous anal carcinoma. An MRI tumour regression
grade (TRG) system based on multiplanar T2-weighted
sequences alone has been proposed to standardise
assessment, yielding over 50% of indeterminate TRG-3
responses at 3 months from the end of treatment [10].
Fig. 3 Images in a 57-year-
old man with squamous anal
carcinoma. T2-weighted MRI
(right column), high b-value
DWI (middle column) and DWI
ADC map (left column). At
baseline MRI, tumour staged as
T4 invades the pelvic sidewall
and prostate (upper row). After
treatment (lower row), a region
of indeterminate intermedi-
ate T2 signal in the lower
rectal wall (TRG-3) does not
correspond to restricted diffu-
sion (arrows); linear diffusion
restriction is spatially matched
to anorectal mucosa (DWI-
TRG-2)
Fig. 4 Images in a 53-year-old
woman with squamous anal
carcinoma. T2-weighted MRI
(right column), high b-value
DWI (middle column) and DWI
ADC map (left column). At
baseline MRI, tumour of the
lower canal staged as T2 (upper
row). Persistence of intermedi-
ate T2 signal (TRG-3) at the
site of tumour (arrows) does not
correspond to restricted diffu-
sion, in keeping with excellent
response (DWI-TRG-2)

European Radiology
1 3
In our study, we found that DWI as a complement to
T2-weighted sequences improved early MRI response
assessment (1–3 months post treatment) by halving the
total number of indeterminate responses from 24 to 12%.
Qualitative evaluation of DWI signal changes, specifically
the resolution of tumour hyperintensity on high b-value
images, increased the subjective TRG scoring confidence
of both expert and non-expert observers. Indeterminate
TRG-3 scores changed most frequently (41%) to DWI-
TRG-2, corresponding to excellent response. TRG-3
scores changed to DWI-TRG-4 in 9% of cases, indicating
minimal response/residual disease, potentially allowing
earlier consideration of salvage surgery, associated with
favourable 5-year survival rates as high as 64% [16].
To our knowledge, external validation of the previously
proposed TRG system for squamous anal carcinoma has
yet to be undertaken. Our findings highlight the potential
for DWI to provide early reassurance on the presence of a
favourable response to definitive chemoradiotherapy, and
to lower the number of patients referred for examination
under anaesthesia and biopsy, given the majority were
downgraded to DWI-TRG-2.
The value of DWI has already been demonstrated in
rectal adenocarcinoma, where active surveillance may
be considered for complete responders following neoad-
juvant chemoradiotherapy [17]. DWI has higher sensi-
tivity in restaging versus T2-weighted MRI alone (84%
vs. 50%) [18]. As observed in our study, DWI improves
Fig. 5 Images in a 41-year-old woman with squamous anal carci-
noma. T2-weighted MRI (right column), high b-value DWI (mid-
dle column) and DWI ADC map (left column). At baseline MRI,
tumour staged as T4 invades the vagina posteriorly and left pelvic
sidewall anteriorly (upper row). After treatment (lower row), the
presence of both linear and nodular diffusion restriction within
a cavity left by tumour shrinkage, not spatially matched to ano-
rectal mucosa (arrows), was deemed indeterminate (TRG-3 and
DWI-TRG-3) by most observers. Complete response was recorded
clinically
Fig. 6 Images in a 64-year-old
woman with squamous anal
carcinoma. T2-weighted MRI
(right column), high b-value
DWI (middle column) and
DWI ADC map (left column).
At baseline MRI, tumour of
the lower canal staged as T3
(upper row). After treatment
(lower row), a small nodule
of restricted diffusion in the 7
o’clock position (arrow) lies
within a broader area of indeter-
minate intermediate T2 signal
(TRG-3), confirming residual
disease (DWI-TRG-4). Incom-
plete response was recorded
clinically

European Radiology
1 3
performance by differentiating post-radiation fibrosis
from viable tumour [13, 19, 20]. In rectal adenocarci-
noma, active surveillance with DWI is now an alternative
to surgery following neoadjuvant therapy with complete
response [14].
Digital rectal examination has traditionally been the
mainstay for determining complete local response in
squamous anal carcinoma, and there is ongoing debate
as to the benefit of imaging versus clinical evaluation.
Treatment-related oedema and/or fibrosis can be difficult
to distinguish from persistent active disease clinically.
Treatment-related effects may even complicate the inter-
pretation of post treatment biopsies. Proximal anorectal
squamous carcinomas and locally advanced tumours rep-
resent a further challenge for clinical assessment, as their
extent may not be fully appreciable by rectal examination,
even under general anaesthesia [3]. It remains accepted
that it may take up to 6 months for complete tumour reso-
lution to occur. In the ACT II trial, the optimum time to
assess complete response was reported as 26 weeks based
on digital rectal examination and abdominopelvic CT [21].
Our findings suggest that T2-weighted MRI plus DWI may
allow for earlier evaluation.
In a previous study by Kochhar etal [10], a high num-
ber of indeterminate TRG-3 scores were found in the early
post treatment period (3 months post chemoradiotherapy)
based on T2-weighted sequences alone, corresponding to
58% of the total. In our study, this proportion was lower,
corresponding to 24% of the total based on T2-weighted
sequences alone. Such difference highlights a variability
in local practice, even between large-volume centres, and
emphasises the need for consensus radiological guidelines
for tumour response assessment.
Interobserver agreement amongst the four observers was
only fair to moderate, with kappa values ranging between
0.28 and 0.58. No significant improvement in agreement
was found by combining T2-weighted sequences with DWI.
Importantly, non-expert observers did not show less agree-
ment that might suggest a difficulty interpreting DWI. On
the contrary, they reached the highest agreement assessing
response on T2-weighted plus DWI.
A 5-point TRG system may be redundant for squamous
carcinoma treated with chemoradiotherapy: only 12 patients
were classified as having a TRG score of 1 or 5, as cur-
rently defined. In the study by Kochhar etal [10], no patients
were scored as TRG-5 and only 2/74 patients were scored as
TRG-1. A modified 3-point TRG incorporating DWI may be
worth assessing in future prospective studies, as proposed
for rectal cancer [22].
The value of DWI in SCCA has been investigated to date in a
small number of studies, assessing its role in tumour volumetry
and staging [23], and its predictive and prognostic value [24,
25]. Non-specialist or non-expert radiologists should familiarise
themselves with the common interpretation pitfalls associated
with DWI [26–28]. In line with previous publications, the most
common problems affecting DWI quality in our study were
low signal-to-noise ratio and susceptibility artefacts from air/
gas at the anal margin and in the rectal lumen. All observers in
our study acknowledged the value of calculated high b-value
images in terms of T2 shine-through reduction [29].
We acknowledge several limitations. First, given the ret-
rospective nature of our study, selection bias might have
affected our results. Complex and advanced cases are
referred to our tertiary surgical oncology centre. Likely
because of this, a high proportion (53%) of locally advanced
tumours was included in our sample. Second, due to the
definitive nature of chemoradiotherapy and the high num-
ber of complete responders, it was not possible to correlate
imaging with tumour histopathology after treatment in the
majority of cases. Third, interpretation of post treatment
MRI was particularly challenging, due to the presence of
marked anatomical distortion, tumour cavities and fistulous
tracts. Fourth, minor variations in the imaging acquisi-
tion across multiple 1.5- and 3.0-Tesla scanners could not
be avoided. There was also some variability in the tim-
ing of post treatment MRI, which ranged between 1 and 3
months. Fifth, unlike Kochhar etal previously, we performed
response MRI at a single early time point (1–3 months),
instead of two time points (3 and 6 months). It is difficult
to speculate how many of the TRG-3 would have resolved
to TRG-2 on T2-weighted sequences alone at 6 months.
Finally, our response assessment methods did not consider
Table 4 Correspondence
between MRI tumour regression
grade (TRG) scores and
clinical response amongst four
observers
Values are number of TRG scores with percentage of total in parentheses
CR, complete response; PR, partial response; PD, progressive disease; LRec, local recurrence
CR (n = 77) PR (n = 5) PD (n = 3) LRec (n = 7)
TRG T2 T2+DWI T2 T2+DWI T2 T2+DWI T2 T2+DWI
1 6/308 (2) 6/308 (2) 0/20 (0) 0/20 (0) 0/12 (0) 0/12 (0) 0/28 (0) 0/28 (0)
2 220/308 (71) 264/308 (86) 7/20 (35) 5/20 (25) 0/12 (0) 0/12 (0) 18/28 (64) 26/28 (93)
3 70/308 (23) 31/308 (10) 11/20 (55) 10/20 (50) 1/12 (8) 0/12 (0) 9/28 (32) 2/28 (7)
4 12/308 (4) 7/308 (2) 2/20 (10) 5/20 (25) 5/12 (42) 6/12 (50) 1/28 (4) 0/28 (0)
5 0/308 (0) 0/308 (0) 0/20 (0) 0/20 (0) 6/12 (50) 6/12 (50) 0/28 (0) 0/28 (0)

European Radiology
1 3
regional nodal response, an important prognostic factor for
squamous anal carcinoma patients [3]. Combined local and
regional nodal response assessment using both FDG PET-
CT and MRI (including DWI) 3 months after chemoradio-
therapy with curative intent was found to be the strongest
predictor of patient outcome by Adusumilli etal in a single-
centre series of 75 patients [30].
It must be stressed that the predictive and prognostic
value of MRI response assessment against clinical refer-
ence standards remains to be proven from large prospective
series in squamous anal carcinomas [9, 10]. Any conclusion
regarding the predictive or prognostic value of DWI over
T2-weighted sequences is beyond the scope of this study.
We found no obvious correlation between MRI TRG and the
onset of local recurrence after initial response.
In summary, the inclusion of DWI alongside T2-weighted
MRI increases diagnostic confidence and improves early
tumour response assessment in squamous anal carcinoma,
by reducing the number of indeterminate responses follow-
ing chemoradiotherapy.
Supplementary information The online version contains supplemen-
tary material available at https:// doi. org/ 10. 1007/ s00330- 023- 09942-0.
Acknowledgements The authors acknowledge funding support from
the Wellcome/Engineering and Physical Sciences Research Coun-
cil Centre for Medical Engineering at King’s College London (WT
203148/Z/16/Z); Cancer Research UK National Cancer Imaging Trans-
lational Accelerator Award (C4278/A27066); UK Research & Innova-
tion London Medical Imaging and Artificial Intelligence Centre; and
National Institute for Health Research Biomedical Research Centre at
Guy’s & St Thomas’ Hospitals and King’s College London.
Funding The authors state that this work has not received any specific
funding.
Declarations
Guarantor The scientific guarantor of this publication is Dr Davide Prezzi.
Conflict of interest Prof Vicky Goh has received research funding for
a PhD student from Siemens Healthineers, unrelated to this study. The
other authors of this manuscript declare no relationships with any com-
panies, whose products or services may be related to the subject matter
of the article.
Statistics and biometry Mr Paul Bassett kindly provided statistical
advice for this manuscript.
Informed consent Written informed consent was waived by the Insti-
tutional Review Board.
Ethical approval Institutional Review Board approval was obtained.
Study subjects or cohorts overlap Some study subjects (n=40) have
been previously reported in ‘Owczarczyk K, Prezzi D, Cascino M etal
(2019) MRI heterogeneity analysis for prediction of recurrence and
disease free survival in anal cancer. Radiother Oncol 134:119-126’.
The study’s aims differed entirely from those of this manuscript and
there is no overlap of results or conclusions.
Methodology
• retrospective
• observational
• performed at one institution
Open Access This article is licensed under a Creative Commons Attri-
bution 4.0 International License, which permits use, sharing, adapta-
tion, distribution and reproduction in any medium or format, as long
as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons licence, and indicate if changes
were made. The images or other third party material in this article are
included in the article's Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not included in
the article's Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will
need to obtain permission directly from the copyright holder. To view a
copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.
References
1. Islami F, Ferlay J, Lortet-Tieulent J, Bray F, Jemal A (2017) Inter-
national trends in anal cancer incidence rates. Int J Epidemiol
46:924-938
2. Glynne-Jones R, Nilsson PJ, Aschele C etal (2014) Anal cancer:
ESMO-ESSO-ESTRO clinical practice guidelines for diagnosis,
treatment and follow-up. Radiother Oncol 111:330-339
3. Rao S, Guren MG, Khan K etal (2021) Anal cancer: ESMO Clini-
cal Practice Guidelines for diagnosis, treatment and follow-up.
Ann Oncol 32:1087-1100
4. de Meric de Bellefon M, Lemanski C, Castan F etal (2020) Long-
term follow-up experience in anal canal cancer treated with inten-
sity-modulated radiation therapy: clinical outcomes, patterns of
relapse and predictors of failure. Radiother Oncol 144:141-147
5. Shakir R, Adams R, Cooper R etal (2020) Patterns and predic-
tors of relapse following radical chemoradiation therapy delivered
using intensity modulated radiation therapy with a simultaneous
integrated boost in anal squamous cell carcinoma. Int J Radiat
Oncol Biol Phys 106:329-339
6. Gourtsoyianni S, Goh V (2014) MRI of anal cancer: assessing
response to definitive chemoradiotherapy. Abdom Imaging 39:2-17
7. NCCN Clinical Practice Guidelines in Oncology: Anal Carcinoma
(2023) Available via https:// www. nccn. org/ profe ssion als/ physi
cian_ gls/ pdf/ anal. pdf. Accessed 6 Mar 2023
8. Bird D, Henry AM, Sebag-Montefiore D, Buckley DL, Al-Qaisieh B,
Speight R (2019) A systematic review of the clinical implementa-
tion of pelvic magnetic resonance imaging-only planning for external
beam radiation therapy. Int J Radiat Oncol Biol Phys 105:479-492
9. Goh V, Gollub FK, Liaw J etal (2010) Magnetic resonance imag-
ing assessment of squamous cell carcinoma of the anal canal
before and after chemoradiation: can MRI predict for eventual
clinical outcome? Int J Radiat Oncol Biol Phys 78:715-721
10. Kochhar R, Renehan AG, Mullan D, Chakrabarty B, Saunders MP,
Carrington BM (2017) The assessment of local response using
magnetic resonance imaging at 3- and 6-month post chemoradio-
therapy in patients with anal cancer. Eur Radiol 27:607-617
11. Wagner M, Ronot M, Doblas S etal (2016) Assessment of the
residual tumour of colorectal liver metastases after chemother-
apy: diffusion-weighted MR magnetic resonance imaging in the
peripheral and entire tumour. Eur Radiol 26:206-215
12. Bharwani N, Miquel ME, Powles T et al (2014) Diffusion-
weighted and multiphase contrast-enhanced MRI as surrogate
markers of response to neoadjuvant sunitinib in metastatic renal
cell carcinoma. Br J Cancer 110:616-624

European Radiology
1 3
13. Kim SH, Lee JM, Hong SH etal (2009) Locally advanced rectal
cancer: added value of diffusion-weighted MR imaging in the
evaluation of tumor response to neoadjuvant chemo- and radiation
therapy. Radiology 253:116-125
14. Gollub MJ, Das JP, Bates DDB etal (2021) Rectal cancer with
complete endoscopic response after neoadjuvant therapy: what is
the meaning of a positive MRI? Eur Radiol 31:4731-4738
15. Amin MB, Edge SB, American Joint Committee on Cancer (2016)
AJCC Cancer Staging Manual, 8th Ed. Springer International
Publishing ISBN 9783319406176
16. Harris DA, Williamson J, Davies M, Evans MD, Drew P, Beynon
J (2013) Outcome of salvage surgery for anal squamous cell car-
cinoma. Colorectal Dis 15:968–973
17. van der Valk MJM, Hilling DE, Bastiaannet E etal (2018) Long-
term outcomes of clinical complete responders after neoadjuvant
treatment for rectal cancer in the International Watch & Wait
Database (IWWD): an international multicentre registry study.
Lancet 391:2537-2545
18. van der Paardt MP, Zagers MB, Beets-Tan RG, Stoker J, Bipat
S (2013) Patients who undergo preoperative chemoradiotherapy
for locally advanced rectal cancer restaged by using diagnostic
MR imaging: a systematic review and meta-analysis. Radiology
269:101-112
19. Maas M, Lambregts DM, Nelemans PJ etal (2015) Assessment of
clinical complete response after chemoradiation for rectal cancer
with digital rectal examination, endoscopy, and MRI: selection
for organ-saving treatment. Ann Surg Oncol 22:3873-3880
20. Sassen S, de Booij M, Sosef M etal (2013) Locally advanced rec-
tal cancer: is diffusion weighted MRI helpful for the identification
of complete responders (ypT0N0) after neoadjuvant chemoradia-
tion therapy? Eur Radiol 23:3440-3449
21. Glynne-Jones R, Sebag-Montefiore D, Meadows HM etal (2017) Best
time to assess complete clinical response after chemoradiotherapy in
squamous cell carcinoma of the anus (ACT II): a post-hoc analysis of
randomised controlled phase 3 trial. Lancet Oncol 18:347-356
22. Lee MA, Cho SH, Seo AN etal (2017) Modified 3-point MRI-
based tumor regression grade incorporating DWI for locally
advanced rectal cancer. AJR Am J Roentgenol 209:1247-1255
23. Prezzi D, Mandegaran R, Gourtsoyianni S etal (2018) The impact
of MRI sequence on tumour staging and gross tumour volume
delineation in squamous cell carcinoma of the anal canal. Eur
Radiol 28:1512-1519
24. Muirhead R, Bulte D, Cooke R etal (2020) A prospective study
of diffusion-weighted magnetic resonance imaging as an early
prognostic biomarker in chemoradiotherapy in squamous cell car-
cinomas of the anus. Clin Oncol 32:874-883
25. Owczarczyk K, Prezzi D, Cascino M etal (2019) MRI heterogene-
ity analysis for prediction of recurrence and disease free survival
in anal cancer. Radiother Oncol 134:119-126
26. van Griethuysen JJM, Bus EM, Hauptmann M etal (2018) Gas-
induced susceptibility artefacts on diffusion-weighted MRI of the
rectum at 1.5 T - effect of applying a micro-enema to improve
image quality. Eur J Radiol 99:131-137
27. Min LA, Vacher YJL, Dewit L etal (2020) Gross tumour volume
delineation in anal cancer on T2-weighted and diffusion-weighted
MRI - reproducibility between radiologists and radiation oncolo-
gists and impact of reader experience level and DWI image qual-
ity. Radiother Oncol 150:81-88
28. Lambregts DMJ, van Heeswijk MM, Delli Pizzi A etal (2017)
Diffusion-weighted MRI to assess response to chemoradiother-
apy in rectal cancer: main interpretation pitfalls and their use for
teaching. Eur Radiol 27:4445-4454
29. Bates DDB, Golia Pernicka JS, Fuqua JL etal (2020) Diagnostic
accuracy of b800 and b1500 DWI-MRI of the pelvis to detect
residual rectal adenocarcinoma: a multi-reader study. Abdom
Radiol (NY) 45:293-300
30. Lambregts DM, Rao SX, Sassen S, Martens MH etal (2014) MRI
and diffusion-weighted MRI volumetry for identification of com-
plete tumor responders after preoperative chemoradiotherapy in
patients with rectal cancer: a bi-institutional validation study. Ann
Surg 262:1034-1039
Publisher’s note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.