Original Article

DOI: 10.1111/vco.12012

The utility of staging in canine mast cell

tumours
J. Warland1 , I. Amores-Fuster2 , W. Newbury3 , M. Brearley1 and J. Dobson1
1 Department

of Veterinary Medicine, Queens Veterinary School Hospital, University of Cambridge,

Cambridge, UK
2
School of Veterinary Science, Small Animal Teaching Hospital, University of Liverpool, Liverpool, UK
3 Abbey Veterinary Centre, Grimsby, UK

Abstract

Keywords
lymph node, metastasis,
sentinel, stage

Current staging of canine mast cell tumours (MCTs) practiced by many veterinarians involves a
minimum of lymph node (LN) assessment, abdominal ultrasound and thoracic radiography.
Historically, some have advocated buffy coat and bone marrow evaluation. Two hundred and twenty
dogs with MCT seen at a referral clinic were staged using LN palpation/cytology, thoracic radiography
and abdominal ultrasound. The utility of each method was evaluated by considering prevalence of
spread and future behaviour. At presentation, 30.9% of dogs had metastases to the local LN; 6.8% of
all the dogs also had distant metastases. No dog had or developed distant metastasis in the absence
of LN metastasis. No dog had convincing evidence of pulmonary metastasis. In this series, the local
LN was sentinel to metastasis and in the absence of local LN metastasis, the utility of further staging
was low. Thoracic radiography was not useful in the staging of canine MCT.

Introduction

Correspondence address:
J. Warland
Queens Veterinary School
Hospital
University of Cambridge
Madingley Road,
Cambridge CB3 0ES, UK
e-mail:
JHW36@CAM.AC.UK

Mast cell tumours (MCTs) are common tumours

of the canine skin, estimated to represent up to 20%
of all skin tumours in this species.1 In the UK, MCT
is the second most common canine malignancy,
after soft tissue sarcoma, with an incidence of 129
per 100 000 insured dogs per year.2 Cutaneous
MCTs are typically solitary lesions, but their clinical
appearance can be variable and some dogs can
develop more than one apparently unrelated MCT.3
The clinical behaviour of MCTs is also variable,
ranging from benign to highly malignant. It is this
variable clinical behaviour and the potential for
associated paraneoplastic syndromes that can make
MCTs challenging to manage.
Currently, predictions of behaviour are made
using histological grading. Most commonly, the
Patnaik grading system separates the tumours into
Presented at World Veterinary Cancer Congress 2012, Paris,

March 13, 2012.

2012 Blackwell Publishing Ltd

low, intermediate and high grades.4 The metastatic

potential and prognosis is significantly correlated
to the grade: low-grade tumours are unlikely to
metastasize and carry a good prognosis; high-grade
tumours frequently metastasize and have a poor
prognosis.3 5 The intermediate grade tumours
do not usually metastasize,6 but a subset will
show metastatic behaviour. Differentiating which
intermediate tumours carry a worse prognosis is
one of the main challenges in MCT research,
with markers such as the mitotic index7 as well
as immunohistochemistry of Ki67,8 and of minichromosome-maintenance proteins (MCM7),9 all
being shown to provide prognostic information
beyond that provided by grade alone. Recently, a
new two-tier grading system has been proposed
that appears to predict the behaviour of MCTs
more accurately, by not using an intermediate
grade.10
Complimentary to histological grading, the
clinical approach to determining the extent and
possible prognosis of a MCT is by its clinical stage

287

288

J. Warland et al.

in particular the presence of lymph node (LN)

involvement or distant metastases.
Until recently, clinical staging of dogs with MCTs
has involved palpation and cytological sampling
of the local LNs, and thoracic and abdominal
imaging (by radiography and ultrasound) to assess
the lungs, liver and spleen. Historically, some
authors also recommended examination of buffy
coat smears for presence of circulating mast
cells and to biopsy or aspirate bone marrow to
evaluate for mast cell infiltration.11 The benefits
of these latter investigations have been shown to
be limited; mastocytaemia can be detected in dogs
without MCTs making buffy coat interpretation
unreliable,12 and the incidence of bone marrow
infiltration by cutaneous MCTs is very low.
Several recent textbooks have recommended a more
measured approach to clinical staging, only advising
full staging (LN cytology, thoracic radiography and
abdominal ultrasound) in the presence of negative
prognostic factors.13,14 Despite this, it is the authors
experience that the majority of veterinarians would
still consider it best practice to carry out full clinical
staging for all grade II and III MCTs.
It is our hypothesis that canine MCTs typically
metastasize to local LNs before spreading more
widely. Thus, the purpose of this study was to
determine the value of full clinical staging in a dog
with a MCT if there is no evidence of spread to the
local draining LN.

Materials and methods

The study consisted of a retrospective analysis of
an Oncology Database on which clinical details
of all oncology cases referred to a university
teaching hospital are entered contemporaneously.
The database was searched for all canine skin
(cutaneous and subcutaneous) MCTs seen between
1 January 1997 and 31 December 2008. The breed,
age, sex and neutering status of each dog as well as
the location and grade of the tumour were recorded
on the database. Tumour location was classified
as being on the head, upper limb (shoulderelbow/hip-stifle), lower limb (elbow-carpus/stifletarsus), foot, trunk, tail and perineum/scrotum.
Grade was taken from the original histopathology
report.

The protocol for clinical staging of MCTs

involved the following: local draining LNs were
examined and any that were palpable (enlarged or
normal) were sampled by fine-needle aspiration
(FNA) and examined cytologically by a clinical
pathologist. The local LN was considered to be the
first LN in the expected lymphatic drainage path.
Where the expected local lymphatic drainage path
included the sublumbar lymphatic chain (medial
iliac, hypogastric, sacral LNs, etc.), this was assessed
by ultrasound; FNA of these nodes was performed
at the discretion of the clinician and radiologist
involved when suspicion for metastasis existed;
normal appearing LNs (size and echotexture) were
considered not to represent metastatic spread.
LN status, positive or negative with respect to
mast cell metastasis, as defined by the clinical
pathologist, was recorded. The method used was
consistent throughout the period of the study.
In each sample, the whole film was scanned
on low power (4 or 10) to make sure that
there were adequate numbers of a mixture of
lymphocytes present to determine that LN was
being examined. Mast cells in clusters, small rafts
and individual numbers were estimated. Criteria
for diagnosis of LN metastasis were increased
numbers of mast cells (more than 3% of the
population or more than 20 cells in total) in
the smear, showing at least three of: poor or
very variable granulation; moderate to marked
anisocytosis; moderate to marked anisokaryosis;
increased nuclear:cytoplasmic ratio; variation in
the nucleus or nucleoli shape or numbers and the
presence of eosinophils or clusters of mast cells.
Where peripheral LNs could not be palpated,
and therefore sampled (e.g. prescapular or inguinal
LNs), these were considered negative for metastasis.
Thoracic radiography (minimum of right and
left lateral views) and abdominal ultrasound, in
some cases with radiography, were performed
on cases routinely, and any evidence of distant
metastasis and its location, as reported in the
detailed radiological report, was recorded. Distant
metastasis was defined as:
Cytological evidence of MCT in LNs other

than the local draining LN (usually a regional

LN and with cytological criteria as described
earlier).

2012 Blackwell Publishing Ltd, Veterinary and Comparative Oncology, 12, 4, 287298

Canine mast cell tumour staging

Ultrasonographic

evidence of visceral
organomegaly accompanied by multiple
hypoechoic nodules or mottled echotexture;
where findings were equivocal, cytology was
used to confirm or refute the presence of
metastatic spread.
Radiological evidence of multiple interstitial
nodules in lung fields consistent with
pulmonary metastasis.

For the purposes of assessing the utility of clinical

staging, any dogs that had either not had their local
LN assessed and its status recorded in the medical
record or had not had investigation for distant
metastasis were excluded.
The medical records of those dogs that met the
inclusion criteria were examined and treatment(s)
received were recorded. In most cases, the outcome
of treatment, date and cause of death were recorded
in the database or the medical record. In cases
where this was not present, follow-up information
including outcome, development of metastasis and
cause of death was obtained from the clinical
records at the referring practice. Animals were
considered lost to follow-up if they had less than
6 months follow-up available.

289

and 2008. Twelve dogs were excluded as they did

not receive full clinical staging. Two dogs with
conjunctival MCTs were excluded as these were not
truly skin associated. Two hundred and twenty dogs
met the inclusion criteria for the study.

Initial presentation
Of the dogs included in the analysis, 123 were female
(101 neutered) and 97 were male (53 neutered).
The mean age was 7.4 years (range 114 years) and
was similar between the sexes. Forty-two different
breeds were represented within the dataset. Five
breeds included more than five cases, which were
Labradors (n = 64), Golden Retrievers (n = 29),
Boxers (n = 22), Staffordshire Bull Terriers (n = 16)
and Jack Russell Terriers (n = 6); there were 23
crossbreed dogs.
The Patnaik system grading4 was recorded for
217 tumours. Twenty-four were considered grade
I, 152 grade II and 20 grade III. Eleven tumours were
graded as III and 10 graded IIIII. The grade was
unavailable in three cases. The grade information is
included in Table 1.

Lymph node status and distant metastases

Statistical analysis
Odds ratios (ORs) were calculated, with 95%
confidence intervals (CIs), for different breeds
with respect to the development of metastases
and multiple tumours. These were calculated
using an online calculator (http://www.hutchon.
net/ConfidOR.htm).
The sensitivity and negative predictive values of LN assessment for metastatic spread
were calculated using another online calculator
(http://www.hutchon.net/EPRval.htm).
Logistic regression analysis was used to establish
what factors, including age, breed, sex, neutering
status, grade and tumour location, influence likelihood of metastasis at presentation. A statistical software package (SPSS version 19, IBM, Armonk, NY,
USA) was used to perform the logistic regression.

Results
Two hundred and thirty-four dogs with MCTs were
entered into the oncology database between 1997

Local LNs were accessible for aspiration in 119 of

220 dogs. The other dogs either had non-palpable
or internal LNs, which could not be sampled.
Sixty-eight (30.9%) dogs had cytological evidence of local LN metastasis at presentation. Fifteen
(6.8% of total) had evidence of distant metastasis
to liver, spleen, skin or distant LNs; all animals
with distant metastasis also had evidence of local
LN metastasis. Overall, 40 dogs had more than
one, apparently unrelated, MCT at presentation;
10 of these had metastatic disease (LN or distant),
whereas 30 had no evidence of metastatic disease. Therefore, the presence of multiple primary
tumours did not increase the risk of metastasis.
Table 2 contains details of the dogs that had
distant metastatic spread.
Two dogs had evidence of pulmonary nodules
consistent with metastatic disease; however, both
these dogs had another unrelated tumour, which
was a more probable cause of the metastases. One
had a suspected primary lung tumour, visible on

2012 Blackwell Publishing Ltd, Veterinary and Comparative Oncology, 12, 4, 287298

290

J. Warland et al.

Table 1. Patnaik histological grades for 220 canine skin mast cell tumours
Grade
Breed
All breeds
Labrador Retriever
Golden Retriever
Boxer
Staffordshire Bull Terrier
Jack Russell Terrier

III

II

IIIII

III

Unknown

Total

24
3
5
2
3
0

11
5
2
0
1
0

152
46
16
18
9
5

10
5
2
1
0
0

20
5
3
0
3
1

3
0
1
1
0
0

220
64
29
22
16
6

the thoracic radiographs, and the other a suspected

histiocytic sarcoma. This latter dog, a bullmastiff,
presented for a recurrent MCT of the right medial
thigh, with satellite nodules in the right inguinum;
cytological analysis confirmed MCT metastasis to
the ipsilateral inguinal and popliteal LNs. On
presentation, the dog was also found to have a large
subcutaneous mass cranial to the right scapula
and a further mass in the right axilla; aspirates
from these masses were suggestive of a malignant
mesenchymal tumour, probably histiocytic in
origin. There was no ultrasonographic evidence
of metastasis to abdominal visceral organs, but
thoracic radiographs revealed multiple pulmonary
nodules. Unfortunately, neither dog was available
for postmortem. These animals were both excluded
from the analysis.
One dog had evidence of lymphadenopathy
(sternal LN) on thoracic radiographs.

Breed data
Boxers appeared to be at lower risk of developing LN
metastases (n = 1; OR: 0.095; 95% CI: 0.0120.718).
Golden Retrievers were at increased risk of multiple
tumours (n = 14; OR: 5.96; 95% CI: 2.5813.8)
but not of developing metastasis. No other breeds
showed statistically altered risk of metastasis or
multiple tumours.

Primary site
When site was considered, MCT on the head
(n = 36) appeared to be more likely to metastasize to
the local LN (n = 17; OR: 2.37; 95% CI: 1.144.94)
and distantly (n = 5; OR: 3.93; 95% CI: 1.3111.9).
Fifty-seven percent of dogs had tumours of limbs
(n = 126), but their site, including position on limb,

did not significantly affect behaviour. The other

tumour locations are shown in Table 3.

Other factors
MCT behaviour was not influenced by sex, age
or neutering status, or the presence or absence of
multiple tumours.
When analysed using a backward, stepwise logistic regression, grade (P = 0.018) and breed (Boxer)
(P = 0.033), were both significantly associated with
the presence or absence of metastatic spread.
Higher grade tumours were associated with a
higher risk of metastatic spread, and Boxer breed
with a lower risk of metastasis. Location (head)
appeared significantly associated with metastasis but did not provide any value beyond that
provided by the grade and breed, and so was
removed from the model at the penultimate step
(P = 0.035).

Treatment
A variety of different treatment modalities and combinations were instigated. The treatments given
before and after staging are outlined in Table 4.
Most dogs had undergone surgery at the referring practice. For the purposes of the study,
chemotherapy included corticosteroids, antihistamines or cytotoxic chemotherapeutic agents or
a combination of these. Fourteen dogs received
some form of systemic chemotherapy before their
referral, and this was continued in all these cases.
Radiotherapy used was coarsely fractionated (four
doses at 800850 cGy at weekly intervals; total
dose 3234 Gy), usually on incompletely resected
tumour scars with inclusion of cytologically positive LNs if present. Some dogs with inoperable

2012 Blackwell Publishing Ltd, Veterinary and Comparative Oncology, 12, 4, 287298

Canine mast cell tumour staging

291

Table 2. Location details of 15 dogs considered to have distant metastasis at the time of initial staging for mast cell tumour

Case

Primary tumour
site

Local lymph node

metastasis
(all cytologically
confirmed)

Distant
metastasis

Left elbow

Left prescapular

Spleen, skin (2)

2
3

Right ventral abdomen

Left parotid region

Right inguinal
Left submandibular

4
5
6

Right upper lip

Right thoracic wall
Left thoracic wall

Right submandibular
Right axillary
Left axillary

7
8
9
10

Left eyebrow
Right axilla
Nose
Scrotum, Lower jaw

Left submandibular
Right axillary
Submandibular
Inguinal, Submandibular

Skin (3)
Prescapular LN, skin
(multiple)
Right prescapular LN
Right prescapular LN
Left prescapular LN,
sternal LN
Left prescapular LN
Right prescapular LN
Prescapular LN
Medial iliac LN

11

Left scapula

Left prescapular

Spleen

12

Right stifle

Popliteal

Medial iliac LN, skin

(multiple), spleen

13

Right lower jaw

Right submandibular

Spleen

14

Right thigh

Right inguinal

Medial iliac LN, liver,

spleen, skin

15

Left ear base

Left submandibular

Left prescapular

gross disease were treated with radiotherapy in

conjunction with prednisolone chemotherapy.

Follow-up
From the original 220 dogs, follow-up data were
available for 185 dogs. Thirty dogs were considered
lost to follow-up and five dogs were excluded
from follow-up analysis. Two described earlier
had evidence of another incidental but metastatic
tumour found during work up for the MCT. Three
other dogs were excluded as the MCT was an
incidental finding after the dog was presented

Other details
Spleen on ultrasound appearance.
Not sampled because of other
cytologically confirmed
metastases
Cytologically confirmed
Cytologically confirmed
Cytologically confirmed
Cytologically confirmed
Prescapular cytologically
confirmed
Cytologically confirmed
Cytologically confirmed
Cytologically confirmed
Inguinal and submandibular
nodes cytologically confirmed.
Dog died at first consultation
because of mast cell
degranulation. Medial iliac node
significantly enlarged not
sampled because of death
Splenectomy 6 weeks prior to
referral because of rupture.
Considered to be metastatic
although histopathology not
performed
Spleen on ultrasound appearance.
Not sampled because of other
cytologically confirmed
metastases
Splenic nodule aspirated and
cytologically confirmed
Spleen and liver on ultrasound
appearance. Not sampled
because of other metastases
Cytologically confirmed

for management of another neoplasm: a nasal

carcinoma, a mediastinal tumour and a cutaneous
haemangiosarcoma.
Sixty-five dogs were alive at the time of last
follow-up, and for these animals the median followup time was 1160 days (range: 4652793 days). One
hundred and twenty dogs were dead (all causes) at
the time of last follow-up, with an overall median
survival time of 465 days (range: 03050 days).
Of the 147 dogs that had no LN or distant
metastases at presentation, at follow-up, 41 were
alive with no further MCT disease, 38 were dead
with no signs of MCT disease, 12 died because of

2012 Blackwell Publishing Ltd, Veterinary and Comparative Oncology, 12, 4, 287298

292

J. Warland et al.

Table 3. The metastatic behaviour of skin mast cell

tumours, at presentation, based on their anatomical

location, in 196 dogs seen at a UK veterinary referral hospital

Location

Total
number

Number
metastasized

36
20
54
51
16
7

17
4
20
13
7
1

11
1

2
0

Head
Trunk
Upper limb
Lower limb
Foot
Limb (not specified/multiple
same limb)
Inguinal/Scrotal
Tail

Where multiple tumours were present in the same location

or area, these were considered together, but where multiple
tumours of different locations were present, these are not
included (n = 24).
Table 4. Summary of the treatments received by 220 dogs

for their skin mast cell tumours before and after initial
staging
No
metastasis Metastasis
Prior to referral/staging
No previous surgical treatment
One surgery by referring practice
Multiple surgeries at referring
practice
Chemotherapy (including
steroids)
After referral/staging
No treatment at all (including
prestaging)
No further treatment
Surgical interventions
Further surgery at referral hospital
First surgery at referral hospital
Adjunctive to surgery
Chemotherapy
Radiotherapy
Chemotherapy and radiotherapy
Non-surgical treatments
Radiotherapy (alone)
Chemotherapy (alone)
Radiotherapy and chemotherapy
Cryosurgery

32
108
12

26
38
4

31

5
8

4
6

11
69
10

17
15
7

0
6
14
1

0
9
11
0

unknown causes, 1 died of pancytopenia following

chlorambucil therapy and 23 were lost to followup. Thirty-two (22%) subsequently developed
further MCT disease; nine developed recurrence
at the primary site, of which four also developed
metastasis; nine other dogs developed metastasis,

which was considered to have been a result of

the presenting tumour, eight of these nine died
of their disease. Two dogs died of progressive
disease from the non-resected primary tumour.
No dogs developed distant metastasis without local
LN involvement. Additionally, 12 dogs developed
further de novo tumours, and two of these
metastasized; this could not have been predicted
from initial staging. Therefore, 15 dogs developed
metastases subsequent to the initial staging. The
majority of these were seen at the referral hospital
(n = 9), but some were assessed by the referring
veterinarian (n = 6). Similar criteria were used
to define metastasis at the initial diagnosis; nine
metastatic lesions were cytologically confirmed
and six were highly suspicious on the basis
of clinical signs and diagnostic investigations;
three dogs had rapidly growing, palpably very
enlarged LNs draining from the MCT location;
two had ultrasonographic and radiographic changes
consistent with metastatic spread splenomegaly
combined with multiple hypoechoic nodules in
the splenic parenchyma, as well as markedly
enlarged hypoechoic LNs and multiple cutaneous
nodules (not cytologically confirmed as MCTs); one
final dog had enlarged, bleeding sublumbar LNs,
within the expected lymphatic drainage pattern,
visualized at exploratory coeliotomy by the referring
veterinarian but not sampled.
Of the 53 dogs considered to have LN metastases
only at presentation, 15 were alive at the final
follow-up (median follow-up 1045 days; range:
6432421), and four were lost to follow-up. Eleven
of these animals were treated with a combination
of therapies that included radiotherapy of the
affected node. Four animals did not receive
radiotherapy; two received systemic cytotoxic
chemotherapy in combination with corticosteroids,
one systemic corticosteroids alone and one dog
received solely surgical intervention. Of the 34
that died, 27 were due to MCT disease, 5 of
other causes and 2 of an unknown cause. The
27 dogs that died because of MCT disease had
received radiotherapy alone (n = 6), radiotherapy
and chemotherapy (n = 7), chemotherapy alone
(n = 11) and surgery alone (n = 3). None of seven
dogs that died because of other or unknown causes
had evidence of further metastatic spread, all had

2012 Blackwell Publishing Ltd, Veterinary and Comparative Oncology, 12, 4, 287298

Canine mast cell tumour staging

received radiotherapy (including the LN), two with

adjunctive chemotherapy.
Out of the 15 dogs with LN and distant metastatic
disease at presentation, 12 died of MCT disease, and
the remainder were lost to follow-up.
Of the total 185 dogs with follow-up, 61 initially
presented with metastatic disease (nodal or nodal
and distant). Thirteen dogs subsequently developed
metastatic disease from the original tumour. Using
this data (excluding the two dogs that developed
metastasis from de novo MCT), the sensitivity of
correctly detecting or predicting metastatic disease
with initial LN assessment was 82.4% (95% CI:
71.590.0%), with a negative predictive value of
89.5% (95% CI: 82.494.1%). For the purpose
of analysis, all dogs with detected metastasis
were considered to have metastatic disease (i.e.
specificity = 100%).

Discussion
The main focus of this study was the utility of the
initial staging of MCTs. The key finding was that no
tumours spread distantly without first spreading to
the local LN.
Survival analysis was considered weak because
of the retrospective nature of the study, a lack
of definitive cause of death data (i.e. postmortem
examinations) as well as a huge variety of treatments
and combinations. These factors will invariably act
as confounding factors to the survival analysis,
and therefore in-depth analysis of survival was
considered likely to be flawed and not undertaken.
The follow-up data collected and presented were
used to validate the results of the initial staging. A
small proportion of dogs were lost to follow-up,
but for the majority the follow-up period was of
considerable duration (median 1160 days) and thus
supports the conclusions, although the possibility
of metastasis development beyond the follow-up
period cannot be ruled out.
Because of the limitations of the pathology
reports in this study, the authors did not distinguish
cutaneous and subcutaneous MCTs, and rather
considered them to be skin MCT. Of the 220
tumours, 217 had been assigned a grade by
the pathologist, which should only be used for
cutaneous tumours. However, it is the authors

293

experience that some pathologists will assign an

intermediate grade to subcutaneous tumours. These
tumours appear to have a similar behaviour to
intermediate-grade cutaneous tumours, possibly
with less propensity to metastasize,15 and therefore
it should not weaken this study that they are not
distinguished. Unfortunately, it was not feasible
to review the histopathology for all the tumours
as many of the submissions are held in external
laboratories, having been excised in practice by the
referring veterinary surgeon.
Conjunctival tumours were excluded from the
study as they were not truly of skin origin. Recent
literature suggests that these tumours have a less
metastatic behaviour16 and a favourable long-term
prognosis. Of the 36 tumours sited on the head,
eight were in the lip but there was insufficient
information in the records to determine if these
were of mucosal origin: if this was the case this
might explain the higher incidence of metastasis
for head tumours as mucosal MCTs have been
reported to be more malignant.17
The logistic regression demonstrated that the
tumour grade was significantly associated with
metastatic behaviour, consistent with previous
studies. The Boxer breed was significantly associated
with low metastatic potential, and the regression
analysis showed that this breed was an independent
predictor of behaviour beyond the grade alone.
The majority of Boxers presented with grade II
tumours, so it would seem that they less commonly
develop metastatic grade II tumours than other
breeds.
In this dataset, 30.9% of tumours had metastasized at presentation, which is similar to a previous
study of another referral population.18 The population in this study is likely to be biased to contain
a different spectrum of disease behaviour to that
seen in general practice. Compared to the overall
population, it is likely that higher grade or more
metastatic tumours are represented. The referral
centre is one of a limited number offering radiotherapy in the UK, so it is likely that a higher
number of incompletely resected tumours, or those
from surgically challenging sites, are seen. This
may explain the large proportion of animals seen
with limb tumours. This propensity of the data to
contain more metastatic tumours than a general

2012 Blackwell Publishing Ltd, Veterinary and Comparative Oncology, 12, 4, 287298

294

J. Warland et al.

practice environment should not be a weakness; it

strengthens the findings about the utility of staging
methods, as metastasis is more likely.
The analysis has shown that no dogs presented
for staging and treatment of their MCT had
evidence of distant metastasis without first showing
local LN metastasis. For this reason, the authors
would suggest that the local LN is sentinel
to the evaluation of metastasis of MCT. This
finding was further consolidated because none of
the dogs that subsequently developed metastatic
disease developed distant disease without local LN
involvement. As has been found in previous studies,
the presence of multiple MCTs did not alter the risk
of metastasis.3
In this study, abdominal visceral organs (liver,
spleen, etc.) were not routinely aspirated unless
ultrasonographic abnormalities were found. This
owes to the potentially invasive nature of the
procedure, which was considered of little value.
Two recent publications have suggested that there
is value in aspirating ultrasonographically normal
appearing liver and spleen for the detection of
MCT metastases19,20 ; however, a larger study
demonstrated no benefit to aspirating normal
appearing visceral organs.21 The clinical staging of
the animals in this study was carried out before the
publication of these later studies. Although the lack
of this FNA data could have been a weakness, the
follow-up data mean that the authors are confident
that dogs were not incorrectly diagnosed as free
of metastatic disease because of undetected visceral
lesions; only a small number of dogs developed
metastatic disease subsequent to being assessed
as being free of secondary spread, and none of
these animals developed visceral spread without
LN metastasis. There remains a possibility that the
disease was more extensive in some dogs with LN
spread than was detected, but this does not weaken
the conclusions of the study.
Although two dogs had radiographic evidence
of pulmonary metastases, in neither case was this
considered to be due to the MCT as both dogs
had another tumour more consistent with the
metastases seen. The finding of no convincing
pulmonary metastasis in this study strongly suggests
that thoracic imaging is of little or no value in the
staging of MCTs. However, it could be argued that

thoracic radiographs may be indicated to screen for

comorbidities; in this series, two dogs appeared to
have unrelated metastatic disease.
Whilst positive LN seems to be prerequisite
to distant metastasis, it is noteworthy that 22 of
the 53 dogs (42%) with cytologically positive LNs
did not subsequently develop distant metastases
or indeed die from MCT-related disease. Of the
animals that were alive despite LN metastatic
spread, a large proportion (11 of 15) had received
radiotherapy of the LN, and a further three had
received systemic medical therapy. All the seven
animals that died of another or an unknown
cause had also received radiotherapy to the LN.
Whilst it is possible that their survival was the
result of successful treatment, it should also be
acknowledged that interpretation of LN cytology
in these cases can be challenging because mast
cells can be found in FNAs of LNs in normal
animals; at least one dog survived with apparent LN
metastasis despite receiving no therapy targeting
the LN. We have previously shown that LN status
had no influence on progression-free survival time
in dogs with MCTs treated with prednisolone and
radiotherapy.22 Other studies have demonstrated
that cytological evidence of LN metastasis from
MCT is associated with a poorer prognosis.14 Krick
et al.18 used various criteria to define possible,
probable and certain metastasis, including the
presence and number/size of aggregates of mast
cells, the degree of pleomorphism, anisocytosis,
anisokaryosis and the level of granulation. They
found that animals with probable or certain LN
metastases lived significantly shorter times than
animals with possible or no LNs metastases. As
with this study, a variety of treatments and lack
of gold standard diagnosis of metastatic spread
meant that assessing the accuracy of their system for
diagnosing metastasis is difficult. Current textbook
recommendations are that metastasis of a MCT to
a LN may be diagnosed on cytology if mast cells
represent greater than 3% of the cell population.23
However, using this criterion, up to 25% of normal
dogs would be diagnosed with a metastatic MCT,24
and therefore FNA results from LNs draining a
MCT must be treated with caution. In this study,
the clinical pathologists used previously published
values for normal numbers of mast cells, combined

2012 Blackwell Publishing Ltd, Veterinary and Comparative Oncology, 12, 4, 287298

Canine mast cell tumour staging

with cytological features of malignant cells in order

to define LN metastasis. As discussed above, there is
significant discrepancy within the literature about
defining LN metastasis; future work needs to focus
on improving the sensitivity and specificity of LN
evaluation.
The concept of the sentinel lymph node (SLN),
the first LN to reliably receive metastatic spread
from a primary tumour, is well established in
human oncology. Histological examination of
these LNs provides information on the status
of lymphatic metastasis for the entire lymphatic
system. In human breast cancer25 and melanoma,26
assessment of SLNs is part of the standard of
care, providing vital prognostic information and
guiding treatment; the finding of a negative LN
saves the patient from aggressive surgery with
high morbidity and limited chance of improving
prognosis, whereas those patients with SLN
metastasis undergo these procedures to improve
prognosis, if the increased morbidity is considered
worthwhile. Tuohy et al.27 reviewed the state of SLN
assessment in veterinary oncology; there have been
limited studies on the exact lymphatic drainage of
most anatomical sites in veterinary species, except
for the mammary gland, meaning that the SLN
identification and assessment is still challenging.
As has been shown in MCT,14 a recent study of
canine mammary carcinomas showed prognostic
value in assessment of the regional LN.28 In
this study, the local LN appeared to be sentinel.
Currently, it is not always known which LNs
are the drainage sites from a particular location,
and so better identification of defined draining
basins would aid the process of identifying SLNs.
Given the difficulties of interpreting LN cytology,
better methods for assessing the SLN, including
routine histopathology to more accurately stage
the disease, are likely to improve management of
MCT. Future studies involving contrast-assisted
ultrasound, lymphoscintigraphy or methylene blue
dye are likely to improve our current understanding
of SLN biology in animals.27
The study has a number of limitations that must
be considered. As a result of the long period of time
over which these animals were referred, a number
of different clinicians, clinical pathologists and radiologists were involved in their care and assessment.

295

Tumour histopathology was also performed at a

number of different laboratories. Interpretation of
ultrasound, cytology and histopathology involves
some unavoidable subjectivity; in particular, it has
been shown that different pathologists may grade
the same MCT differently,10 and the same is true
for interpretation of LN cytology. The retrospective
analysis relied on accurate medical records, both at
the referral hospital and the primary practice, where
follow-up was requested. Recall bias can become
a more significant problem in studies undertaken
over long periods of time; however, the study was
undertaken using the oncology database and clinical records produced at the time, and no data were
gathered through mental recall of clinical information, either by veterinarians or by owners. The
majority of important follow-up data was gathered
from the contemporaneously produced oncology
database rather than from referring veterinarians,
again reducing the risk of bias being introduced. It
is important to recognize that not all local LNs were
sampled by FNA; in many cases, where the local LN
was the prescapular (forelimb lesions) or inguinal
(upper hind limb, trunk), it was not possible to
locate or palpate the LN and thus not possible
to collect FNAs from this site; palpation is very
insensitive for detecting enlargement of the medial
iliac nodes, and potential morbidity meant that
medial iliac nodes that were ultrasonographically
normal were also not sampled. If we had included
only those dogs whose LNs were palpable and thus
aspirated in this study, this would have produced a
very biased population less relevant to those seen in
practice.
The lack of cytological information on LN status
at the time of presentation could be viewed as a
weakness in this study, in that it is not certain
that occult metastases were not present in some
of the nodes classified as negative. However, we
strongly believe that, when considering the followup information as well, this does not detract from
the main finding in this series of cases that no MCT
metastasized distantly without first involving the
local LN.
To address this potential weakness, a prospective
study should be designed with the intention of
sampling every local LN, with imaging guidance or
other techniques if necessary.

2012 Blackwell Publishing Ltd, Veterinary and Comparative Oncology, 12, 4, 287298

296

J. Warland et al.

Fine Needle Aspirate Mast Cell

Tumour

Is the Local
Lymph Node
Subcutaneous
No

Yes

Is Local
Lymph
Node
Palpable
?

Ultrasound
assessment of
Local LN +/- FNA
Yes
FNA LN

Is Lymph
Node
Metastasis
present?

No

No
Can
tumour be
excised
with 2cm
margins?

Yes/Suspicious
No

Further Staging
-Abdominal ultrasound
-Further lymph node evaluation
-+/- Blood sample (if appropriate)

Is Distant
Metastasis
Present?

Yes

Yes
Biopsy Primary
Tumour

Excise with 2cm

margins

No
Excise Primary
Tumour (2cm margins)
& Lymph Node for
Histopathology

Further Treatment &

Prognosis Dictated by
Grade, Prognostic
Markers (e.g. MCM,
Ki67, Mitotic Index) &
Metastatic Status

Figure 1. Summary flow diagram of the authors suggested approach to a newly diagnosed canine skin MCT. Modified
from a previously suggested textbook algorithm.29

In the authors opinion, the results of this study

suggest that staging of MCTs should begin with
assessment of the local LN. If this fails to show
evidence of metastatic disease, then the utility of
further staging is likely to be limited. It is likely that
combining information from the primary tumour

grade and evaluation and SLN staging will provide the best information on the extent and likely
behaviour of MCT. We propose the use of the flow
diagram (Fig. 1) to best stage a newly diagnosed
MCT. Further evaluation of the Kiupel et al.10 twotier grading system, immunohistochemical markers

2012 Blackwell Publishing Ltd, Veterinary and Comparative Oncology, 12, 4, 287298

Canine mast cell tumour staging

and further studies into the best method of assessment of the SLN are vital to improve our management of MCT. A prospective study using more accurate markers of metastatic MCT spread should be
undertaken to confirm these preliminary findings.
Further to this, investigations into the optimal management strategies, including medical and surgical
treatments, given the SLN status, will drive forward
our treatment of these challenging tumours.

Acknowledgements
The authors would like to thank all the staff at
the Queens Veterinary School Hospital for their
help with the diagnosis and treatment of these dogs
and the referring veterinary practices for providing
vital follow-up information. The authors would
like to thank Tess Hoather for her management
of the Oncology Database. Dr Mark Holmes is
acknowledged for his assistance with the statistical
analyses.

References
1. Dorn CR, Taylor DO, Schneider R, Hubbard HH
and Klauber MR. Survey of animal neoplasms in
Alameda and Contra Costa counties, California II,
Cancer morbidity in dogs and cats from Alameda
County. Journal of the National Cancer Institute
1968; 40: 307318.
2. Dobson JM, Samuel S, Milstein H, Rogers K and
Wood JLN. Canine neoplasia in the UK: estimates
of incidence rates from a population of insured
dogs. The Journal of Small Animal Practice 2002; 43:
240246.
3. Murphy S, Sparkes AH, Blunden AS, Brearley MJ
and Smith KC. Effects of stage and number of
tumours on prognosis of dogs with cutaneous mast
cell tumours. The Veterinary Record 2006; 158:
287291.
4. Patnaik AK, Ehler W and MacEwan G. Canine
cutaneous mast cell tumor: morphologic grading
and survival time in 83 dogs. Veterinary Pathology
1984; 21: 469474.
5. Bostock D. The prognosis following surgical
removal of mastocytomas in dogs. The Journal of
Small Animal Practice 1973; 14: 2740.
6. Murphy S, Sparkes AH, Brearley MJ, Smith KC and
Blunden AS. Relationships between the histological
grade of cutaneous mast cell tumours in dogs, their
survival and the efficacy of surgical resection. The
Veterinary Record 2004; 154: 743746.

297

7. Romansik EM, Reilly CM, Kass PH, Moore PF and

London CA. Mitotic index is predictive for survival
for canine cutaneous mast cell tumors. Veterinary
Pathology 2007; 44: 335341.
8. Scase TJ, Edwards D, Miller J, Henley W, Smith K,
Blunden A and Murphy S. Canine mast cell tumors:
correlation of apoptosis and proliferation markers
with prognosis. Journal of Veterinary Internal
Medicine 2006; 20: 151158.
9. Berlato D, Stewart J, Newton R, Maglennon GA,
Monti P, Flindall A and Murphy S. Evaluation of
minichromosome maintenance protein 7 as a
prognostic marker in canine cutaneous mast cell
tumours. Veterinary and Comparative Oncology
2012; 10: 135142.
10. Kiupel M, Webster JD, Bailey KL, Best S, DeLay J,
Detrisac CJ, Fitzgerald SD, Gamble D, Ginn PE,
Goldschmidt MH, Hendrick MJ, Howerth EW,
Janovitz EB, Langohr I, Lenz SD, Lipscomb TP,
Miller MA, Misdorp W, Moroff S, Mullaney TP,
Neyens I, OToole D, Ramos-Vara J, Scase TJ,
Schulman FY, Sledge D, Smedley RC, Smith K,
Snyder PW, Southorn E, Stedman NL, Steficek BA,
Stromberg PC, Valli VE, Weisbrode SE, Yager J,
Heller J and Miller R. Proposal of a 2-tier
histological grading system for canine cutaneous
mast cell tumors to more accurately predict
biological behavior. Veterinary Pathology 2011; 48:
147155.
11. Madewell BR and Theilen GH. Tumours of the skin
and subcutaneous tissues; Part IV. Mast cell and
melanocytic neoplasms. In: Veterinary Cancer
Medicine. 2nd Chapter 14 edn., GH Theilen and BR
Madewell eds., Philadelphia, Lea & Febiger, 1987:
310325.
12. McManus PM. Frequency and severity of
mastocytemia in dogs with and without mast cell
tumors: 120 cases (19951997). Journal of the
American Veterinary Medical Association 1999; 215:
355357.
13. Thamm DH and Vail DM. Mast cell tumors. In:
Withrow & MacEwans Small Animal Clinical
Oncology. Chapter 19, SJ Withrow and DM Vail
eds., St Louis, Saunders Elsevier, 2007: 402424.
14. Blackwood L. Tumours of the skin and
subcutaneous tissues. In: BSAVA Manual of Canine
and Feline Oncology. 3rd Chapter 12 edn., JM
Dobson and BDX Lascelles eds., Gloucester,
BSAVA, 2011: 130158.
15. Newmana SJ, Mrkonjich L, Walker KK and
Rohrbach BW. Canine subcutaneous mast cell
tumour: diagnosis and prognosis. Journal of
Comparative Pathology 2007; 136: 231239.

2012 Blackwell Publishing Ltd, Veterinary and Comparative Oncology, 12, 4, 287298

298

J. Warland et al.

16. Fife M, Blocker T, Fife T, Dubielzig RR and Dunn

K. Canine conjunctival mast cell tumors: a
retrospective study. Veterinary Ophthalmology 2011;
14: 153160.
17. Hillman LA, Garrett LD, de Lorimier LP, Charney
SC, Borst LB and Fan TM. Biological behavior of
oral and perioral mast cell tumors in dogs: 44 cases
(19962006). Journal of the American Veterinary
Medical Association 2010; 237: 936942.
18. Krick EL, Billings AP, Shofer FS, Watanabe S and
Sorenmo KU. Cytological lymph node evaluation in
dogs with cutaneous mast cell tumors: association
with grade and survival. Veterinary and
Comparative Oncology 2009; 7: 130138.
19. Book AP, Fidel J, Wills T, Bryan J, Sellon R and
Mattoon J. Correlation of ultrasound findings, liver
and spleen cytology, and prognosis in the clinical
staging of high metastatic risk canine mast cell
tumors. Veterinary Radiology and Ultrasound 2011;
52: 548554.
20. Stefanello D, Valenti P, Faverzani S, Bronzo V,
Fiorbianco V, Pinto da Cunha N, Romussi S,
Cantatore M and Caniatti M. Ultrasound-guided
cytology of spleen and liver: a prognostic tool in
canine cutaneous mast cell tumor. Journal of
Veterinary Internal Medicine 2009; 23:
10511057.
21. Finora K, Leibman NF, Fettman MJ, Powers BE,
Hackett TA and Withrow SJ. Cytological
comparison of fine-needle aspirates of liver and
spleen of normal dogs and of dogs with cutaneous
mast cell tumours and an ultrasonographically
normal appearing liver and spleen. Veterinary and
Comparative Oncology 2006; 4: 178183.

22. Dobson J, Cohen S and Gould S. Treatment of

canine mast cell tumours with prednisolone and
radiotherapy. Veterinary and Comparative Oncology
2004; 2: 132141.
23. Duncan J. The lymph nodes. In: Diagnostic Cytology
and Haematology of the Dog and Cat. 2nd edn.,
RL Cowell, RD Tyler and JH Meinkoth, eds., St
Louis, Mosby, 1999: 97103.
24. Bookbinder PF, Butt MT and Harvey HJ.
Determination of the number of mast cells in
lymph node, bone marrow and buffy coat
cytological specimens in dogs. Journal of the
American Veterinary Medical Association 1992; 200:
16481650.
25. Goyal A and Mansel RE. Recent advances in
sentinel lymph node biopsy for breast cancer.
Current Opinion in Oncology 2008; 20: 621626.
26. Cochran AJ, Ohsie SJ and Binder SW. Pathobiology
of the sentinel lymph node. Current Opinion in
Oncology 2008; 20: 190195.
27. Tuohy JL, Milgram J, Worley DR and Dernell WS.
A review of sentinel lymph node evaluation and its
need for its incorporation into veterinary oncology.
Veterinary and Comparative Oncology 2009; 7:
8191.
28. Szczubial M and Lopuszynski W. Prognostic value
of regional lymph node status in canine mammary
carcinomas. Veterinary and Comparative Oncology
2011; 9: 296303.
29. Brearley MJ and Murphy S. Mast cell tumors. In:
Decision Making in Small Animal Oncology.
Chapter 7, DJ Argyle, MJ Brearley and MM Turek
eds., Ames, Blackwell Publishing, 2008: 147159.

2012 Blackwell Publishing Ltd, Veterinary and Comparative Oncology, 12, 4, 287298