This document describes a study that evaluated the computed tomography (CT) features of 19 canine thyroid tumors. Key findings include: - Most tumors (17/19) were carcinomas and appeared heterogeneous with mineralization and well-defined margins post-contrast. - CT identified invasion in 8 tumors that were histopathologically invasive, while 5 non-invasive tumors showed no invasion on CT. - Palpation assessing mobility had low sensitivity (71%) and specificity (0%) for detecting histological invasion compared to CT invasion which had 70% sensitivity and 100% specificity. - CT provided information on tumor characteristics, extent, and invasion that was useful for surgical planning and determining prognosis.

1. PAPER
Journal of Small Animal Practice • Vol 55 • June 2014 • © 2014 British Small Animal Veterinary Association 323
Journal of Small Animal Practice (2014) 55, 323–329
DOI: 10.1111/jsap.12219
Accepted: 19 February 2014; Published online: 7 April 2014
Computed tomographic appearance
of canine thyroid tumours
K. Deitz*, L. Gilmour*, V. Wilke† and E. Riedesel*
*Department of Veterinary Clinical Sciences, College of Veterinary Medicine, Iowa State University, Ames, IA 50011, USA
†Department of Veterinary Clinical Sciences, College of Veterinary Medicine, University of Minnesota, St. Paul, MN 55108, USA
Presented in oral abstract form at the American College of Veterinary Internal Medicine Forum, New Orleans, LA, May 2012.
OBJECTIVES: To describe the computed tomography features of canine thyroid tumours.
METHODS: Retrospective study of records of dogs with a thyroid tumour and neck computed tomography.
Neck computed tomographies were evaluated for tumour characteristics. Thoracic radiographs and
computed tomographies were evaluated for lung nodules.
RESULTS: Of 19 identified cases, 17 were carcinomas and 2 were adenomas; 12 had mineralisation, 16
had heterogeneous attenuation and 16 were unilateral. Tumours were located from the temporoman-
dibular joint to C5. Sixteen had well-defined margins postcontrast. Tumours were ovoid and mean vol-
ume was 57·4 cm3
. By computed tomography, eight had definitive or possible invasion into surrounding
structures; all eight were histopathologically invasive carcinomas. Five histopathologically non-inva-
sive tumours and two adenomas had no computed tomography invasion into surrounding structures.
Four had complete palpable mobility (two adenomas and two histopathologically invasive carcinomas);
one had computed tomography evidence of possible invasion. The sensitivity of palpable mass mobil-
ity to determine histopathological invasion was 71% with 0% specificity. The sensitivity of computed
tomography invasion to determine histopathological invasion was 70% with 100% specificity.
CLINICAL SIGNIFICANCE: Computed tomography scans revealed several common features. Palpable mass
mobility was not definitive for lack of histopathological invasion. Computed tomography invasion was
specific but not very sensitive for histopathological invasion.
INTRODUCTION
Thyroid tumours account for 1 to 2% of all canine cancers
(Harari et al. 1986, Wucherer & Wilke 2010). They can occur
anywhere from the base of the tongue (Lantz & Salisbury 1989)
to the base of the heart (Liptak et al. 2008, MacDonald et al.
2009). Most clinically apparent tumours are carcinomas (Leav
et al. 1976) with 16 to 60% of dogs having metastases at the time
of diagnosis (Leav et al. 1976, Birchard & Roesel 1981, Carver
et al. 1995).
Boxers, beagles, golden retrievers and Siberian huskies are
over-represented (Leav et al. 1976, Birchard & Roesel 1981,
Harari et al. 1986, Wucherer & Wilke 2010). Clinical signs
range from the finding of a neck mass to respiratory difficulty,
cough, vomiting, regurgitation, dysphagia, weight loss, anorexia
and bark change (Harari et al. 1986, Feldman & Nelson 2004).
Dogs with functional tumours present with signs relating to
hyperthyroidism such as weight loss, hyperactivity, tachypnoea,
polydipsia, polyuria, polyphagia, diarrhoea and vomiting (Feld-
man & Nelson 2004).
In humans, computed tomography (CT) is used to evaluate
the extent of large tumours and demonstrate invasion into sur-
rounding structures (King 2008). Invasion indicates that exten-
sive reconstructive surgery may be needed, or may render the
patient inoperable (King 2008). A thoracic CT can be performed
concurrently with the neck CT to evaluate for pulmonary metas-
tasis. The purpose of this study was to describe the CT appear-
ance of canine thyroid tumours. Histopathological invasion, CT
invasion into surrounding structures and mass mobility on pal-
pation were compared.
ttp://www.bsava.com

2. K. Deitz et al.
324 Journal of Small Animal Practice • Vol 55 • June 2014 • © 2014 British Small Animal Veterinary Association
the fixed category. In addition, masses with possible or defini-
tive invasion into surrounding structures on CT were consid-
ered to have CT invasion. Adenomas were not included in the
calculations.
Thoracic CT scans were evaluated for lung nodules and ecto-
pic thyroid tissue (Liptak et al. 2008, Kang et al. 2012). All
cervical and thoracic CT scans were independently evaluated
by three of the authors (KD, LG and ER), and then each CT
scan was evaluated by consensus of the three authors, similar to a
previously published study (Marolf et al. 2011). Evaluators were
a board-certified internist (KD), a first-year radiology resident
(LG) and a board-certified radiologist (ER).
RESULTS
Study population and clinical characteristics
Nineteen dogs met the inclusion criteria. Breeds included four
golden retrievers, three Labrador retrievers and one each of
the following: beagle, bichon frise, Border collie, bullmastiff,
Cardigan Welsh corgi, English springer spaniel, German shep-
herd dog, German shorthaired pointer, greyhound, miniature
poodle, mixed breed and rat terrier. The mean age was 10·7
years (range 6·4 to 15·5 years). There were 9 females (7 spayed, 2
entire) and 10 males (9 neutered, 1 entire). Tumours were diag-
nosed by either incisional (n=2) or excisional (n=17) biopsy. Of
the tumours diagnosed by excisional biopsy, 15 of 17 were carci-
nomas (5 NI and 10 invasive) and 2 of 17 were adenomas. Both
tumours diagnosed with incisional biopsy were carcinomas.
The presenting complaint varied with two dogs having unex-
pected neck masses serendipitously palpated at physical exami-
nation; both were adenomas. Ten dogs were presented because
the owner found a neck mass with no other clinical signs. The
remaining seven dogs were presented with at least two of the
following clinical signs: dysphagia (n=3), neck swelling (n=3),
cough (n=2), hyperactivity (n=2), panting (n=2), polydipsia
(n=2), polyuria (n=2), regurgitation (n=2), vomiting (n=2) and
one each of decreased appetite, exercise intolerance, front limb
lameness, heat intolerance, increased respiratory effort, neck pain
and weight loss. Palpable mass mobility was described in 13, of
these 4 were fixed, 4 were mobile and 5 were mobile with attach-
ment. Twelve of the 19 dogs had circulating total T4 measured;
7 had reference interval total T4 concentrations, 4 had elevated
and 1 had decreased total T4 concentration.
Imaging characteristics
Eighteen tumours were generally ovoid and well defined; the
remaining tumour was ovoid with irregular margins. Three
masses had attached nodular or cystic structures. Mean volume
calculated for 18 of 19 tumours was 57·4 cm3
±standard devia-
tion (sd) of 65·6 cm3
(range 5·2 to 266·3 cm3
). The volume of the
irregular tumour was not calculated.
Ten of the categorical CT tumour characteristics that were
evaluated are shown in Table 1, indicating the number of
scans demonstrating each characteristic. Only 3 of 19 tumours
had homogenous attenuation and 7 of 16 heterogeneously
MATERIALS AND METHODS
A 5-year medical records search (2006 to 2011) was undertaken
at Iowa State University and University of Minnesota. Dogs
with a histopathologically confirmed cervical thyroid tumour
and a precontrast and postcontrast CT scan of the tumour were
included.
The following data were recorded from medical records: breed,
age, sex, presenting complaint, palpable mass mobility (fixed,
mobile or mobile with attachment based on the clinician’s sub-
jective assessment), total thyroxine (T4) and histopathological
description and diagnosis. Tumours were defined as non-invasive
(NI; tumour cells confined within the capsule) or invasive (cap-
sular ± vascular invasion or invasion into surrounding structures)
based on histopathological description. Available thoracic radio-
graphs were reviewed for pulmonary metastasis.
CT scans of the neck, and in some cases, the thorax, were
acquired using either a single-slice spiral scanner (Ultra-Z Mar-
coni) or a 16-slice multidetector scanner (Aquilon-16 Toshiba).
Dogs were scanned in dorsal recumbency under general anaes-
thesia. Contiguous axial images were obtained using a soft-tissue
algorithm before and after intravenous (IV) administration of 0·5
mL/kg of iodinated contrast medium (Iohexol – Omnipaque 240
or Ioversol – Optiray 350). Image volumes were reconstructed in
three planes with variable slice thickness (1 to 6 mm) depending
on patient size. kVP and mAs varied from 120 to 135 and 60 to
250, respectively.
Each CT scan was evaluated for the following subjective
tumour characteristics: shape, precontrast attenuation unifor-
mity (heterogeneous/homogeneous), cavitation (hypodense
areas precontrast; present/absent), mineralisation on precontrast
images (present/absent), contrast enhancement pattern (rim,
heterogeneous/homogeneous), appearance of margins post-
contrast (well- or ill-defined), invasion into adjacent structures
(local vasculature, bone and muscle), appearance of surround-
ing lymph nodes (LNs) and presence of peritumoural aberrant
blood vessels (none, mild, moderate and marked). Each CT scan
was evaluated for the following objective characteristics: volume
(using the formula for an ellipse), location (cranial/caudal extent;
unilateral/bilateral/ventral and left/right), presence and location
of normal-appearing thyroid tissue, attenuation of tumour and
normal thyroid lobes precontrast and postcontrast [measured in
Hounsfield units (HU) using regions of interest drawn on recon-
structed images to accommodate the largest amount of tissue in
any plane] and intratumoural vascularisation (the difference in
postcontrast and precontrast attenuation in HUs; <50: mild, 50
to 100: moderate and >100: marked) (Taeymans et al. 2013).
Palpable mass mobility and evidence of CT invasion into
surrounding structures were compared to histopathological
invasion. Sensitivity, specificity, positive predictive values and
negative predictive values were calculated for invasion on CT
and mass mobility on palpation, respectively, to determine
histopathological invasion, using histopathological description as
the gold standard. For the purpose of this calculation, masses that
were completely fixed or mobile with attachment were placed in

3. Journal of Small Animal Practice • Vol 55 • June 2014 • © 2014 British Small Animal Veterinary Association 325
CT appearance of canine thyroid tumours
located tumours had normal thyroid lobes visible (Fig 4B).
These two tumours, both carcinomas, also invaded the hyoid
bones. The bilateral tumour was predominately right-sided but
extended ventrally to the left. It was very large extending from
mid-bulla to C5 and obscured or invaded normal thyroid tissue
as no normal thyroid gland was visible.
Visible and objectively normal-appearing thyroid tissue was
found in 14 of 19 dogs. Mean ±sd of precontrast attenuation of
the normal thyroid lobes was 94 ±19 (range 53 to 127 HU); mean
±sd of postcontrast attenuation of these lobes was 154 ±22 (range
105 to 186 HU). Eleven of 16 dogs with unilateral tumours had
visible contralateral thyroid lobes, whereas 4 of 16 with unilateral
tumours had no visible thyroid tissue. The one dog with bilat-
eral tumours had no visible thyroid tissue. Three of 19 dogs had
bilaterally visible normal-appearing thyroid lobes. The tumours
in each of these dogs were located cranially (3) and ventrally (2).
The cranial margins were either rostral to the TMJ or in the plane
of the bulla. This group of three dogs includes the two dogs with
the ventral tumours invading bone.
The combined mean ±sd attenuation of the tumours precon-
trast was 44 ±13 HU (range 20 to 80 HU) with a postcontrast
mean ±sd attenuation of 105 ±33 HU (range 36 to 159 HU).
This resulted in 8 of 19 tumours being classified as having mild
intratumoural vascularisation (simple mean difference of 32
HU), 10 of 19 having moderate intratumoural vascularisation
(simple mean difference of 76 HU) and 1 of 19 having marked
intratumoural vascularisation (117 HU difference). One carci-
noma had the lowest contrast enhancement of −1 HU owing to
a large hypoattenuating central region that contained gas density
foci. Tumours with mild vascularisation tended to be heteroge-
neous with cavitations (5 of 8), whereas tumours with moderate
vascularisation tended to be heterogeneous with no cavitations
(5 of 10). Only one moderately vascularised tumour had cavita-
tion. The single strongly vascularised tumour was heterogeneous
with cavitation. Combining the attenuation pattern and degree
of contrast enhancement, the two adenomas had a heterogeneous
and cavitation pattern with mild intratumoural vascularisation.
The 16 carcinomas were more diverse; however, 7 of 16 had
heterogeneous cavitation patterns with moderate intratumoural
vascularisation. Sixteen tumours had well-defined margins post-
contrast (Figs 1 and 2) and seven had distinct peripheral rim
enhancement. Mean mass attenuation was compared to mean
left epaxial muscle attenuation at the level of C2. Epaxial muscle
precontrast mean ±sd attenuation was 62 ±4 HU (range 51 to
70 HU) and postcontrast mean ±sd attenuation was 71 ±6 HU
(range 54 to 78 HU).
One hyperthyroid dog had a rostroventral tumour with nor-
mal thyroid lobes bilaterally (Fig 4). This was interpreted as ecto-
pic neoplastic hyperactive thyroid tissue. One hyperthyroid dog
had a bilateral mass (more right-sided) with no normal thyroid
lobes noted, indicating that normal thyroid tissue was either
invaded or obscured by neoplastic tissue. The remaining two
hyperthyroid dogs had unilateral masses with normal thyroid
lobes noted on the contralateral side. The one dog with a low
total T4 concentration had the largest tumour (left-sided) and a
normal-appearing thyroid lobe on the right.
Table 1. CT imaging characteristics of histopathologically
confirmed thyroid tumours in 19 dogs
Imaging
characteristic
No. of
tumours
Precontrast attenu-
ation uniformity
Heterogeneous 16
Homogeneous 3
Cavitation Yes 7
No 12
Mineralisation Yes 12
No 7
LN involvement Unilaterally enlarged retroph. 8
Unilaterally enlarged mandibular 2
Unilaterally enlarged sup. cervical 1
Retroph. attached to/invaded by tumour 5
Normal LNs 6
LNs not seen 2
Aberrant blood
vessels
None 5
Mild 5
Moderate 7
Marked 2
Cranial extent Cranial to the TMJ 1
Cranial aspect of bulla 2
Mid bulla 2
Cranial C2 4
Mid C2 4
Caudal C2 5
C2/C3 junction 1
Caudal extent Mid C1 1
Mid C2 1
Mid C3 1
Caudal C3 3
C3/C4 junction 1
Mid C4 4
Caudal C4 2
Cranial C5 1
Mid C5 1
Caudal C5 4
Location Unilateral 16
Bilateral 1
Ventral 2
Normal thyroid
tissue
Opposite side of tumour 10
None seen 5
Bilateral 3
Two normal lobes, mass attached to one 1
Intratumoural
vascularisation
Mild 8
Moderate 10
Marked 1
CT Computed tomography, LN Lymph node, retroph. Retropharyngeal, sup. Superficial, TMJ
Temporomandibular joint
attenuating tumours had regions of hypodensity categorised as
cavitations (Figs 1 and 2). Twelve of the tumours had precontrast
mineralisation. The two adenomas had heterogeneous attenua-
tion with cavitation and mineralisation.
Tumours were variably located. Thirteen had the cranial mar-
gin somewhere along the plane of C2. One had its cranial margin
in the plane of the C2/C3 junction. Five had the cranial margin
cranial to the plane of C2; one with its cranial extent at the tem-
poromandibular joint (TMJ) and four at the bulla. The centers of
all but four were in the plane of C3 through cranial C4 (Fig 3).
Six extended into or caudal to the plane of C5.
No dogs had discrete tumours in both lobes. Sixteen were uni-
laterally located (11 left-sided and 5 right-sided); one was bilat-
eral and the other two were ventral (Fig 4A). The two ventrally

4. K. Deitz et al.
326 Journal of Small Animal Practice • Vol 55 • June 2014 • © 2014 British Small Animal Veterinary Association
A B C
FIG 1. A 10-year-old entire female border collie. Postcontrast images highlighting common characteristics of a well-defined, unilateral, left-sided thy-
roid tumour centred at C3/C4. Histopathologically, this was a well-differentiated, non-invasive thyroid follicular cell carcinoma. (A) Axial image of an
ovoid-shaped tumour (black star) at the level of caudal C3. The mass has discrete, rounded margins and heterogeneous contrast enhancement, with a
central hypodense region. Tumour margins abut the oesophagus (arrow) containing an oesophageal stethoscope, trachea containing an endotracheal
tube, the left sternothyrohyoideus (S), longus colli (LCo) and longus capitis (LCa) muscles and the left internal carotid artery (open arrow). (B and C)
Sagittal and dorsal (C) images highlighting engorged vessels cranial and caudal to the mass (open arrows) and a discrete hypodense region in the
center of the mass (arrow). Remnants of normal thyroid tissue were present (asterisk)
FIG 2. An eight-year-old entire female German shepherd dog. Postcontrast
axial image of an ovoid, unilateral left-sided thyroid carcinoma (thick
arrow) at the level of C3. The mass has discrete, rounded margins and
heterogeneous contrast enhancement with moderate vascularisation. The
tumour margins abut the oesophagus (E), left sternothyrohyoideus and
longus capitis muscles and left internal carotid artery. A normal-appear-
ing right thyroid is present (thin arrow)
FIG 3. Location of thyroid tumours. The line indicates the most cranial
and caudal extent of the range of tumours. The dots indicate the approxi-
mate center location of each tumour. There is mild mineralisation of the
cricoid cartilage. Gas in the cranial oesophagus marks the caudal end of
the upper oesophageal sphincter
The four hyperthyroid dogs had a precontrast mean ±sd atten-
uation of 59 ±15 HU (range 48 to 80 HU) and a postcontrast
mean ±sd attenuation of 123 ±34 HU (range 73 to 150 HU).
The dog with a low total T4 concentration had attenuation val-
ues of 20 and 59 HU precontrast and postcontrast, respectively.
Five tumours had definitive or possible vascular invasion (all
five were carcinomas; one NI, three invasive and one could not
be determined as diagnosis was based on incisional biopsy).
Two had definitive vessel invasion (one right jugular and one
left carotid), two had suspected vessel invasion (one left internal
carotid and one right carotid) and in one scan, the right jugu-
lar vein was poorly demarcated from adjacent tumour tissue.
Two tumours invaded the basihyoid bone (Fig 4A); one of these
had possible ceratohyoid invasion. Both tumours had definitive
invasion of the attaching musculature. Three tumours had pos-
sible or suspected laryngeal, pharyngeal or oesophageal muscula-
ture invasion.
Five CT scans had no increase in number of aberrant peritu-
moural blood vessels associated with the tumour, five had mild
increase in blood vessels, seven had moderate increase in blood
vessels and two scans had marked increase in blood vessels (Fig 5).
Two scans had no discernible LNs and six had normal-appear-
ing LNs. Eleven had LN abnormalities, with some having more
than one abnormality and/or more than one LN involved. LN
changes included asymmetry in size; asymmetry in degree and/
or homogeneity of contrast enhancement; indistinct margination
or LNs adhered to, replaced or invaded by the tumour. Ten scans
had changes involving the medial retropharyngeal LNs; in 7 of
10 the affected node was ipsilateral to the tumour. Other LNs
with changes included the mandibular (n=3; 1 on the right in

5. Journal of Small Animal Practice • Vol 55 • June 2014 • © 2014 British Small Animal Veterinary Association 327
CT appearance of canine thyroid tumours
A B
FIG 4. A six-year-old neutered male Labrador retriever. Precontrast axial images of a ventrally located thyroid carcinoma. The tumour was histologically
invasive with venous tumour emboli. (A) Thyroid tumour (black star) at the level of the craniocervical junction. The mass causes regional destruction
of the basihyoid and ceratohyoid bones. Invasion of the sternohyoid and geniohyoid muscles that attach to the basihyoid bone is highly suspected.
Thyrohyoid bones are marked with arrows. (B) Cervical tissues at the level of C3 with two normal-appearing thyroid glands (black arrows)
FIG 5. A seven-year-old spayed female golden retriever. Postcontrast images of a thyroid carcinoma centred at C4 with atypical computed tomography
(CT) characteristics. Histologic evidence of invasion into the oesophagus and connective tissue adjacent to the trachea was present. (A) Axial image
at the level of the craniocervical junction. There is marked increase in number and size of regional vasculature. The right linguofacial vein (asterisk)
is markedly distended and tortuous, with a large, heterogeneous filling defect. Differentials for the filling defect include neoplastic invasion, vascular
thrombus or a combination of these. Ventral to the vessel is a fluid-dense pocket that most likely represents a seroma associated with recent surgical
biopsy in this patient. (B) Axial image at the level of C3. The tumour is amorphous in shape and highly vascular. It is difficult to delineate tumour from
blood vessels containing filling defects. Separate areas of mineralisation are present (arrows), which were also identified on precontrast images. Both
external jugular veins are visible and subjectively distended. (C) Sagittal image lateral to the vertebral column at the level of the right external ear
canal (asterisk). The distended right linguofacial vein (black arrows) with hypodense filling defects is visible draining into the external jugular vein
along with the right maxillary artery (open arrow). (D) Dorsal image in the plane of the trachea. An endotracheal tube is visible (arrow). The distended
right linguofacial vein leading into the right external jugular vein (open arrow), and with mineralisation in the right side of the tumor are also visible
(asterisk). (E) Ventral-dorsal 3D shaded volume rendering showing relationship of vascular structures in and around the thyroid mass on each side of
the neck
A B
D E
C

6. K. Deitz et al.
328 Journal of Small Animal Practice • Vol 55 • June 2014 • © 2014 British Small Animal Veterinary Association
a recent study on CT appearance of canine thyroid tumours
(Taeymans et al. 2013). In this study and the study by Taeymans
et al. (2013), tumour attenuation values were lower than normal
thyroid attenuation values (Taeymans et al. 2008). This is likely
due to differences in metabolism and storage of iodine between
tumour and normal tissue (Silverman et al. 1984, Arger et al.
1985, Verschueren et al. 1992, Imanishi et al. 2000). In addi-
tion, all four hyperthyroid dogs had hypoattenuating tumours
compared with normal thyroid (Taeymans et al. 2008). Pre-
contrast attenuation of masses was lower than that of the sur-
rounding muscle tissue. This is true of diseased human thyroid
tissue, which is isoattenuating or hypoattenuating to surround-
ing musculature (Sekiya et al. 1979, Silverman et al. 1984, Arger
et al. 1985, Imanishi et al. 1991, 2000). Thyroid attenuation in
humans relates to iodine content (Imanishi et al. 2000). There is
decreased iodine in human papillary thyroid carcinoma (Hansson
et al. 2008); the same may be true in canine thyroid carcinoma.
Criteria for a more favourable outcome by surgical treatment
of thyroid gland carcinoma have been stated to be complete
mobility of the tumour, absence of local tumour invasion and
absence of distant metastasis (Klein et al. 1995, Liptak 2007,
Radlinsky 2007). Complete mobility is logically associated with
lack of tumour invasion and good potential for complete resec-
tion, whereas fixed mobility implies local tumour invasion and
poor success of adequate resection. Mobility on palpation was
somewhat sensitive for determining histopathological invasion
(71%) but was not specific (0%). The sensitivity of CT inva-
sion for determining histopathological invasion was similar to
mobility on palpation, but was much more specific at 100%.
This suggests that palpation alone should not be used to indicate
histopathological invasion and that evidence of invasion on CT
is more useful than palpation.
Three of the histopathologically invasive tumours had ill-
defined margins postcontrast, while none of the NI tumours had
ill-defined postcontrast margins, suggesting that ill-defined mar-
gins indicate histopathological invasion. Unfortunately, the find-
ing of well-defined margins postcontrast is not as useful, as most
histopathologically invasive tumours had well-defined postcon-
trast margins. The two adenomas had very similar CT appear-
ance to the carcinomas, including mineralisation and cavitation;
however, the adenomas had the smallest tumour volumes.
Five scans lacked normal thyroid tissue bilaterally and 11 scans
lacked normal thyroid tissue unilaterally. Such findings on CT
scans should raise suspicion for a thyroid tumour as the origin
of a neck mass. It can be difficult to identify a neoplastic thy-
roid mass when the tumour arises from ectopic tissue and normal
thyroids are seen bilaterally, as was the finding in three scans in
this study.
This study, the study by Taeymans et al. (2013) and a further
study by Rossi et al. (2013) described ectopic thyroid carcinomas
located at midline, ventral to the hyoid apparatus. This appears
to be a location where neoplastic transformation of ectopic thy-
roid can occur. This area may be more prone to development of
ectopic tissue because this is the location of the thyroglossal duct
and is a location for ectopic thyroid in humans (McHenry et al.
1993, Shah et al. 2007).
the scan with the bilateral but more right-sided mass, 1 ipsilateral
to the mass and 1 contralateral) and superficial cervical (n=1;
ipsilateral to the mass).
Histopathological invasion, CT postcontrast margins, CT
invasion into surrounding structures and palpable mass mobil-
ity were evaluated. There were five histopathologically NI
carcinomas and two adenomas; all seven of these masses had well-
defined margins postcontrast with no invasion into surrounding
structures on CT. Palpable mass mobility was described for six of
these. The two adenomas were mobile. Three of the NI carcino-
mas were mobile with attachment and one was fixed. There were
10 histopathologically invasive carcinomas. Three had ill-defined
margins postcontrast and eight had possible or definitive inva-
sion into surrounding structures on CT. Palpable mass mobility
was described for seven of these; two were mobile, three were
fixed and two were mobile with attachment.
The sensitivity of palpable mass mobility to determine his-
topathological invasion was 71% while the specificity was 0%,
with a positive predicative value of 55% and a negative predictive
value of 0%. The sensitivity of CT invasion to determine histo-
pathological invasion was 70% while the specificity was 100%,
with a positive predictive value of 100% and a negative predictive
value of 63%.
Seventeen dogs had thoracic radiographs; 4 had two projec-
tions (single lateral and ventrodorsal or dorsoventral) and 13 had
three projections (both lateral projections and ventrodorsal or
dorsoventral). No thoracic radiographs had lung nodules. Five
dogs had thoracic CT scans, including the two that did not have
radiographs; one had evidence of metastasis on CT and the other
did not. Three dogs had both a thoracic CT scan and radio-
graphs. In two dogs, there were lung nodules on CT but not on
radiographs. None of the thoracic CT scans showed abnormal
tissue in the cranial mediastinum and cardiac regions where ecto-
pic thyroid tissue has been most commonly described (Liptak
et al. 2008, Kang et al. 2012).
DISCUSSION
Common thyroid tumour CT findings in this study include het-
erogeneous attenuation characteristics, cavitation, ovoid shape,
mineralisation, well-defined margins postcontrast and mild
to moderate vascularisation. Aberrant vessels associated with
tumours were common. Masses were most frequently unilat-
eral. Most masses had their cranial margin within the plane of
the length of C2, with the caudal margin within the plane of
the length of C4 to C5. Average tumour volume was 57·4 cm3
,
which is similar to a recently published study where the average
tumour volume on CT was 65 cm3
(Taeymans et al. 2013) and,
as noted in that study, approximately 50 times the volume of the
normal canine thyroid (Taeymans et al. 2008).
Precontrast and postcontrast attenuation values for the nor-
mal thyroid glands in this study were similar to a previous study
reporting CT appearance of normal thyroid glands (Taeymans
et al. 2008). Mean precontrast and postcontrast attenuation
values for the tumours were also similar between this study and

7. Journal of Small Animal Practice • Vol 55 • June 2014 • © 2014 British Small Animal Veterinary Association 329
CT appearance of canine thyroid tumours
Feldman, E. C. & Nelson, R. W. (2004) Canine thyroid tumors and hyperthyroid-
ism. In: Canine and Feline Endocrinology and Reproduction. 3rd edn. Eds E. C.
Feldman and R. W. Nelson. Saunders Elsevier, St. Louis, MO, USA. pp 219-249
Hansson, M., Grunditz, T., Isaksson, M., et al. (2008) Iodine content and distribu-
tion in extratumoral and tumor thyroid tissue analyzed with X-ray fluorescence
and time-of-flight secondary ion mass spectrometry. Thyroid 18, 1215-1220
Harari, J., Patterson, J. S., Rosenthal, R. C., et al. (1986) Clinical and patho-
logic features of thyroid tumors in 26 dogs. Journal of the American Veterinary
Medical Association 188, 1160-1164
Imanishi, Y., Ehara, N., Mori, J., et al. (1991) Measurement of thyroid iodine by CT.
Journal of Computer Assisted Tomography 15, 287-290
Imanishi, Y., Ehara, N., Shinagawa, T., et al. (2000) Correlation of CT values, iodine
concentration, and histologic changes in the thyroid. Journal of Computer
Assisted Tomography 24, 322-326
Kang, M. H., Kim, D. Y. & Park H. M. (2012) Ectopic thyroid carcinoma infiltrating
the right atrium of the heart in a dog. Canadian Veterinary Journal 53, 177-181
King, A. D. (2008) Imaging for staging and management of thyroid cancer. Cancer
Imaging 8, 57-69
Klein, M. K., Powers, B. E., Withrow, S. J., et al. (1995) Treatment of thyroid car-
cinoma in dogs by surgical resection alone: 20 cases (1981-1989). Journal of
the American Veterinary Medical Association 206, 1007-1009
Lantz, G. C. & Salisbury, S. K. (1989) Surgical excision of ectopic thyroid car-
cinoma involving the base of the tongue in dogs: three cases (1980-1987).
Journal of the American Veterinary Medical Association 195, 1606-1608
Leav, I., Schiller, A. L., Rijnberk, A., et al. (1976) Adenomas and carcinomas of the
canine and feline thyroid. American Journal of Pathology 83, 61-93
Liptak, J. M. (2007) Canine thyroid carcinoma. Clinical Techniques in Small Animal
Practice 22, 75-81
Liptak, J. M., Kamstock, D. A., Dernell, W. S., et al. (2008) Cranial mediastinal
carcinomas in nine dogs. Veterinary and Comparative Oncology 6, 19-30
MacDonald, K. A., Cagney, O. & Magne, M. L. (2009) Echocardiographic and clini-
copathologic characterization of pericardial effusion in dogs: 107 cases (1985-
2006). Journal of the American Veterinary Medical Association 235, 1456-1461
Marolf, A. J., Gibbons, D. S. & Podell, B. K. (2011) Computed tomographic appear-
ance of primary lung tumors in dogs. Veterinary Radiology and Ultrasound 52,
168-172
McHenry, C., Danish, R. & Murphy, T. (1993) Atypical thyroglossal duct cyst: a rare
cause for a solitary cold thyroid nodule in childhood. The American Journal of
Surgery 59, 223-228
Radlinsky, M. G. (2007) Thyroid surgery in dogs and cats. The Veterinary Clinics of
North America Small Animal Practice 37, 789-798
Rossi, F., Caleri, E., Bacci, B., et al. (2013) Computed tomographic features
of basihyoid ectopic thyroid carcinoma in dogs. Veterinary Radiology and
Ultrasound 54, 575-581
Sekiya, T., Tada, S. & Kawakami, K. (1979) Clinical application of computed tomog-
raphy to thyroid disease. Journal of Computed Tomography 3, 185-193
Shah, B., Ravichand, C. & Juluri, S. (2007) Ectopic thyroid cancer. Annals of
Thoracic and Cardiovascular Surgery 13, 122-124
Silverman, P. M., Newman, G. E., Korobkin, M., et al. (1984) Computed tomography
in the evaluation of thyroid disease. American Journal of Roentgenology 142,
897-902
Taeymans, O., Schwarz, T. & Duchateau, L. (2008) Computed tomographic features
of the normal canine thyroid gland. Veterinary Radiology and Ultrasound 49, 13-19
Taeymans, O., Pennick, D. G. & Peters, R. (2013) Comparison between clinical,
ultrasound, CT, MRI, and pathology findings in dogs presented for suspected
thyroid carcinoma. Veterinary Radiology and Ultrasound 54, 61-70
Verschueren, C., Ruttleman, G. & Vos, J. (1992) Thyrotropin receptors in nor-
mal and neoplastic (primary and metastatic) canine thyroid tissue. Journal of
Endocrinology 132, 461-468
Wucherer, K. L. & Wilke, V. (2010) Thyroid cancer in dogs: an update based on
638 cases (1995-2005). Journal of the American Animal Hospital Association
46, 249-254
Two cases had lung nodules on thoracic CT, where none were
noted on thoracic radiographs. The increased sensitivity of tho-
racic CT in finding metastatic disease has been described for
other tumours (Alexander et al. 2012, Armbrust et al. 2012).
Thoracic CT should be considered in cases of suspected thyroid
cancer, and can be done with the cervical CT. The combina-
tion provides optimal information on which to base treatment
recommendations.
Limitations of the study are owing to its retrospective nature;
there was heavy reliance on medical records for descriptions of
mobility. Also, histopathology of tumours may not have been
representative; all margins of tumours may not have been submit-
ted or examined for invasiveness, which would affect histopath-
ological categorisation. Lastly, the small sample size precluded
additional statistical analysis.
These findings support the use of CT for staging and surgi-
cal planning of suspected thyroid tumours. There are several CT
characteristics common to thyroid tumours, including minerali-
sation and cavitation. The finding of bilaterally normal thyroid
glands on CT should not remove thyroid tumour from the dif-
ferentials for an unidentified neck mass. Lastly, palpable mass
mobility is a poor indicator of histopathological invasion. Inva-
sion into surrounding structures on CT is somewhat sensitive
and very specific for histopathological invasion.
Conflict of interest
None of the authors of this article has a financial or personal
relationship with other people or organisations that could inap-
propriately influence or bias the content of the paper.
References
Alexander, K., Joly, H., Blond, L., et al. (2012) A comparison of computed tomog-
raphy, computed radiography, and film-screen radiography for the detection of
canine pulmonary nodules. Veterinary Radiology and Ultrasound 53, 258-265
Arger, P. H., Jennings, A. S., Gordon, L. F., et al. (1985) Computed tomography find-
ings in clinically normal and abnormal thyroid patients. Journal of Computed
Tomography 9, 111-117
Armbrust, L. J., Biller, D. S., Bamford, A., et al. (2012) Comparison of three-view
thoracic radiography and computed tomography for detection of pulmonary
nodules in dogs with neoplasia. Journal of the American Veterinary Medical
Association 240, 1088-1094
Birchard, S. J. & Roesel, O. F. (1981) Neoplasia of the thyroid gland in the dog:
a retrospective study of 16 cases. Journal of the American Animal Hospital
Association 17, 369-372
Carver, J. R., Kapatkin, A. & Patnaik, A. K. (1995) A comparison of medullary thy-
roid carcinoma and thyroid adenocarcinoma in dogs: a retrospective study of 38
cases. Veterinary Surgery 24, 315-319