ORIGINAL STUDIES
Use of Contrast-Enhanced Ultrasound in the Differential Diagnosis of Adrenal Tumors in Dogs Paolo Bargellini, DVM, Riccardo Orlandi, DVM, PhD, Alfredo Dentini, DVM, PhD, Chiara Paloni, DVM, Giuseppe Rubini, DVM, Paolo Fonti, DVM, Alessia Diana, DVM, Mark E. Peterson, DVM, DACVIM, Cristiano Boiti, DVM
ABSTRACT We evaluated the diagnostic accuracy of the contrast-enhanced ultrasonography (CEUS), using a second-generation microbubble contrast agent, in differentiating the different types of adrenal mass lesions in 24 dogs. At B-mode ultrasound, 9 lesions involved the right adrenal gland, 14 the left, and 1 was bilateral. Each dog received a bolus of the contrast agent into the cephalic vein, immediately followed by a 5-mL saline flush. The first contrast enhancement of each adrenal lesion was evaluated qualitatively to assess the degree of enhancement and its distribution during the wash-in and wash-out phases, as well as the presence of non-vascularized areas and specific vascular patterns. Pathological diagnoses were determined in all dogs by histopathology or by cytology. Combining enhancement degree and vascularity resulted in the best predictive model, allowing CEUS to differentiate adrenocortical adenoma (n¼10), adenocarcinoma (n¼7), and pheochromocytoma (n¼7) with an accuracy of 91.7% (P , 0.001). Combining enhancement degree and vascularity, CEUS can discriminate malignant versus benign adrenal lesions with a sensitivity of 100.0%, a specificity of 80.0%, and an accuracy of 91.7% (P , 0.001). In conclusion, results of this study confirm that CEUS is useful for differentiating between the different types of adrenal tumors in dogs. (J Am Anim Hosp Assoc 2016; 52:132–143. DOI 10.5326/JAAHA-MS-6363)
Introduction
any obvious clinical signs at all.4,5 Often, these adrenal masses are
The differential list for adrenal pathologies that cause adrenal mass
incidental findings encountered during imaging screening for other
lesions encompasses several types of benign and malignant tumors,
diseases.6 All adrenal mass lesions, including incidentalomas,
including adrenocortical adenomas, adrenocortical adenocarcino-
require clinical evaluation to reach a definitive diagnosis. Pituitary
mas, and pheochromocytomas, as well as metastatic cancers and
and adrenal tumors may coexist in dogs with hyperadrenocorti-
other less frequent tumor types.1 In dogs, primary adrenal gland
cism, thus further complicating the diagnosis.7
tumors have a relatively low incidence, representing only 0.17 to
Imaging of adrenal glands by ultrasound is a valuable
0.76% of all the neoplasias, but account for 10 to 20% of dogs
technique for the detection of adrenal hyperplasia and adrenal
diagnosed with Cushing’s syndrome.2,3 The adrenal masses can
masses in dogs.8,9 Overall, bilateral and/or unilateral adrenal gland
cause a large array of clinical signs ranging from hyperadrenocor-
enlargement, with loss of the normal shape and parenchymal
ticism to catecholamine excess depending on the cell type involved
structure, are frequently diagnosed in dogs.10 However, differen-
by the pathology; in some dogs, the adrenal tumors do not cause
tiating between benign and malignant adrenal masses (i.e.,
From the Tyrus Clinica Veterinaria, Terni, Italy (P.B., R.O., A.D., C.P.);
AUC (area under the receiver-operating characteristic curve); CEUS
Ultravet Diagnostic, Bologna, Italy (G.R.); Centro Veterinario
(contrast-enhanced ultrasonography); CT (computed tomography);
Specialistico, Roma, Italy (P.F.); Dipartimento di Scienze Mediche
MRI (magnetic resonance imaging)
Veterinarie, Ozzano Emilia, Italy (A.D.); Animal Endocrine Clinic, New York, NY (M.E.P.); and Laboratorio di Biotecnologie Fisiologiche, Universita` degli Studi di Perugia, Perugia, Italy (C.B.). Correspondence:
[email protected] (R.O.)
132
JAAHA.ORG
Q 2016 by American Animal Hospital Association
CEUS in Diagnosis of Adrenal Tumors in Dogs
adrenocortical adenomas and carcinoma) or between tumors
examined using B-mode ultrasound with linear (4–13 MHz) and
arising from the adrenal cortex or medulla (i.e., pheochromocy-
microconvex transducers (6.6–8.0 MHz) to obtain scans of adrenal
toma) based only on their morphological appearance on ultra-
glands and other abdominal organs, including liver and ovaries of
sound remains inconclusive.11,12 In human medicine, computed
intact females. The adrenal mass lesions were evaluated for size,
tomography (CT) and magnetic resonance imaging (MRI) can
margins, echogenicity, echotexture, and extension to adjacent
differentiate benign and malignant adrenal masses in such a way
vessels (caudal vena cava and/or aorta). During the B-mode
that only a limited number of patients requires adrenal biopsy to
examination, adrenal glands were also measured in the longitudinal
confirm the diagnosis.13
plane: the greatest cranio-caudal dimension was defined as the
Recently, contrast-enhanced ultrasonography (CEUS) has
adrenal length, whereas the greatest dorso-ventral dimension
been employed to evaluate the perfusion patterns of the adrenal
perpendicular to the longitudinal axis was defined as adrenal
gland in normal dogs, as well as in dogs with pituitary-dependent
thickness. An adrenal gland was considered atrophic when its
hyperadrenocorticism.14–16 In humans, CEUS has been found to
thickness was less than 5 mm.18
provide useful clinical information, being accurate enough to discriminate adrenal adenomas and non-adenomatous masses with a sensitivity comparable to that of CT and MRI techniques.
17
Contrast-Enhanced Ultrasonography For CEUS examination, the probe used was a linear transducer (3–
Our hypothesis was that the contrast enhancement patterns
8 MHz) with low-mechanical index (0.04–0.05) to avoid the
found with CEUS might be specific to each different type of adrenal
rupture of microbubbles; the presets of the machine were modified
pathology. Therefore, the purpose of this study was to evaluate
as follows: Dynamic Range, 7, and Persistence, between 6 and 12.
whether CEUS can reliably differentiate dogs with benign or
The ultrasound contrast agent employed was the sulphur-
malignant adrenal lesions, as well as between the different types of
hexafluoride echo-signal enhancerc. The average diameter of the
adrenal tumors.
microbubbles was 2.5 lm when dissolved in physiological solution. Microbubbles consist of phospholipid capsules that are stabilized
Materials and Methods
by surfactant substances and filled with sulphur hexafluoride. Once
Clinical Cases
prepared, the aqueous solution of the contrast agent remains stable
All dogs included were required to have informed owner consent,
for 6–8 hr before use, but its lifespan decreases to 5 min following
and procedures were performed in accordance with Italian laws on
IV injection. In each dog, a bolus dose (0.03 mL/kg of body weight)
animal care. In this study, we enrolled 24 dogs (13 females and 11
of the freshly prepared contrast agent was rapidly infused via an
males) over a period of 33 mo (from October 2010 to June 2013),
18G catheter, provided with a three-way-valve inserted in the
all of which were found to have either unilateral (23 dogs) or
cephalic vein, immediately followed by a 2.5- or 5-mL saline flush
bilateral (1 dog) adrenal mass lesions identified during an
(NaCl 0.9%), depending on dog size.
abdominal ultrasound examination. Once the adrenal mass lesions
The contrast examination was performed by two experienced
were identified, the dogs recruited in this study underwent a
operators: the first injected the contrast medium through the
standardized protocol that included B-mode sonography imaging
catheterized vein, while the second performed the ultrasound scans
followed by CEUS. Additional inclusion criteria required a
of the adrenal glands by holding the transducer as still as possible
cytologic or histologic confirmation of an adrenal gland tumor
on the selected position during the contrast study. The left and
and no treatment for hyperadrenocorticism (Cushing’s syndrome).
right adrenal glands were examined on right and left lateral
No other dogs with adrenal mass lesions were diagnosed during
recumbent position, respectively, following separate injection of
this 33-mo study period, other than two cases of suspected adrenal
ultrasound contrast agent bolus 10 min apart, starting from the
tumors without histopathologic confirmation and seventeen cases
gland with the adrenal lesions previously identified by B-mode
of bilateral adrenal hyperplastic nodules; none of these cases were
ultrasound. The adrenal target lesions were scanned continuously
included in this study.
for 5 min and the dynamic process of contrast enhancement was examined in real time. The adrenal glands were imaged in their
B-Mode Ultrasonography Examination
longitudinal axis. For the left adrenal gland, the phrenicoabdominal a
Following IV injection of 0.1 mg/kg butorphanol , all scanning
vessels, left renal artery, cranial mesenteric, and celiac arteries
procedures were performed by two operators on mildly sedated
(when possible) were maintained in the image to serve as an in vivo
dogs in lateral recumbent position using a machine equipped with
reference; for the right adrenal gland, the in vivo reference vessels
a contrast tuned imaging technologyb. The dogs were first
included in the image were caudal vena cava and right renal artery.
JAAHA.ORG
133
During each examination, a digital video clip of 1-min in duration
groups, characterized by marked anisokaryosis with severe nuclear
was taken during the injection of the contrast agent and stored in
dysmetria, aberrant chromatin pattern, and multiple atypical
the hard disk of the ultrasoundscanner.
nucleoli for the adrenal adenocarcinoma.19,20
Evaluation of B-Mode US and CEUS Imaging
Statistical Analysis
The images of each dog were analyzed by two radiologists. Neither
The statistical analysis was performed using a software packaged.
had performed the ultrasound examination and both were blinded
The Shapiro-Wilk test was used to evaluate normality of
to the clinical diagnosis and to the result of the histological
continuous data that were expressed as mean (6 standard
examination.
deviation)and median. The single dog with a diagnosis of
For evaluation of the B-mode ultrasound, echogenicity of the
adrenocortical lipoma was excluded from the statistical analysis.
adrenal mass lesions was compared to normal-surrounding adrenal
Due to nonparametric distribution, the significance of differences
or contralateral gland and categorized as hypoechoic, isoechoic,
in sizes between each adrenal tumor type was examined using the
hyperechoic, or mixed hypo/hyperechoic. Echotexture of the
Kruskal-Wallis test, while that of tumors diagnosed as benign and
adrenal lesion was defined as homogeneous or heterogeneous and
malignant was calculated with the Mann Whitney U tests. The
its margins were characterized as well- or poorly defined. Position
histological/cytological diagnosis of each adrenal lesion was used as
and number of the adrenal lesions, as well as the eventual atrophy
the gold standard of reference for all dogs. The association between
of the contralateral gland, were also evaluated. For the CEUS
different adrenal lesions and B mode/CEUS features was assessed
images, the first contrast enhancement within the adrenal mass
using Fisher’s exact test. Receiver-operating characteristic curves
lesion was evaluated qualitatively to assess the following: (1) the
and logistic regression analysis were used for evaluating the
degree of enhancement (hypo-, iso-, and hyper-enhancement)
diagnostic performance (i.e., sensitivity, specificity, and diagnostic
during the arterial phase; (2) the internal distribution of contrast
accuracy, defined as % of correct prediction with reference
enhancement (homogeneous, heterogeneous, nodular); (3) the
diagnosis) of qualitative B-mode and CEUS enhancement patterns
presence of non-vascularized areas; (4) specific vascularization
for differentiating malignant versus benign adrenal lesions and
patterns (centripetal, peripheral, disordered); and (5) the time of
between the tumor types, respectively.21 In the simple logistic
contrast enhancement (fast/synchronous/slow) during the early
regression analysis, all independent variables were included as
wash-in and wash-out phases. Signal intensity of the adrenal mass
covariates. In the multiple logistic regression analysis, correlated
lesion was compared with the surrounding normal adrenal tissue
variables were not included simultaneously in the model, and the
(when present) or with the contralateral adrenal gland.14,15,16
stepwise selection method was used. Differences were considered significant at P , 0.05.
Reference Diagnosis Definitive diagnosis of the adrenal mass lesions was made by
Results
cytopathologic or histopathologic evaluation reviewed by a
Clinical Findings and Reference Diagnosis
veterinary pathologist. The adrenal specimens were obtained by
The signalment (age, breed, sex) and clinical signs for the 24 dogs
fine-needle aspiration (n ¼ 3), at the time of elective surgery (n ¼
are summarized in Table 1. Affected dogs ranged in age from 6 to
6), or at necropsy (n ¼ 15) within 3 mo from CEUS examination.
15 yr (median, 10.5 yr) and weighed 7 to 36 kg (median, 12 kg).
The biopsy specimen, taken at surgery or at necropsy, were fixed in
Only six dogs had clinical signs attributable to hyperadrenocorti-
formalin, processed according to routine histological techniques,
cism, whereas three dogs reported episodes of collapse and one
and stained with hematoxylin and eosin, whereas the fine-needle
retinal detachment. Fourteen dogs did not have any clear clinical
aspirates were smeared onto slides, air dried, and stained with
signs related to adrenal pathology, and the adrenal lesions were
Romanovsky stain or fixed in 95% ethanol for formal cytopatho-
found incidentally during abdominal ultrasound examination for
logical review. The tumors were diagnosed histopathologically,
workup of unrelated diseases. The 25 adrenal lesions classified on
while the cytological diagnosis of the other three tumors was made
the basis of histopathological (n¼22) or cytological (n¼3)
using the following features: presence of cells loosely aggregated
evaluation (Table 1) included adrenocortical adenomas (10;
together with many other single cells, characterized by numerous
40.0%), adrenocortical adenocarcinomas (7; 28.0%), pheochromo-
naked nucleoli, high nuclear pleomorphism, and basophilic
cytomas (7; 28.0%), and lipoma (1; 4.0%). Dog #4 in Table 1 was
cytoplasm with sparse eosinophilic granules for the pheochromo-
found to have bilateral adrenocortical adenoma. All pheochromo-
cytoma; and the presence of cells predominantly arranged in
cytomas here examined were diagnosed as malignant.
134
JAAHA
|
52:3 May/Jun 2016
Mixed-breed dog
Mixed-breed dog
Mixed-breed dog
West Highland white terrier
20
21
22
23
Mixed-breed dog
Hound
19
24
Mixed-breed dog
18
Mixed-breed dog
German shepherd dog
16
17
Bulldog
15
10
13
14
10
8
12
8
10
10
7
13
F
F
M
F
M
M
F
M
F
F
F
F
M
F
M
M
M
F
F
F
M
F
M
F
Sex
*dog with bilateral adrenocortical adenoma The dogs are grouped according to the histological/cytological reference diagnosis. C, cytological; F, female; H, histological; M, male.
Lipoma
Pheochromocytoma
Mixed-breed dog
14
9
10
Mixed-breed dog
12
6
9
Yorkshire terrier
Bulldog
9
12
German shepherd dog
8
15
11
13
Mixed-breed dog
7
14
Mixed-breed dog
6
11
13
Mixed-breed dog
Rottweiler
5
Boxer
Mixed-breed dog*
4
10
11
Mixed-breed dog
3
12
13
Age (Yr)
10
Labrador Retriever
2
Adenocarcinoma
English setter
1
Adenoma
Breed
Case Number
Diagnosis
Signalment and Reference Diagnosis
H
C
C
H
H
H
H
H
C
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Reference Diagnosis
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
þ
-
þ -
-
þ
-
þ
-
þ -
-
þ
þ -
-
-
-
-
-
-
þ
-
Polyphagia
þ
Polyuria / Polydypsia
Signalment and Key Clinical Signs of 24 Dogs with Adrenal Tumor Lesions Included in the Study
TABLE 1
-
-
-
-
-
-
-
-
-
þ
-
-
þ
-
þ
-
þ
-
þ
-
-
-
þ
-
Abdominal Enlargement
-
-
-
-
-
-
-
-
-
-
-
-
þ
-
-
-
þ
-
þ
-
-
-
þ
-
Alopecia
-
-
-
-
-
-
-
-
-
-
þ
-
þ
-
þ -
þ
-
-
þ
-
þ
-
þ
þ
-
þ
-
þ
-
þ
-
þ
þ
þ
-
þ
Incidentaloma
þ
-
-
-
-
þ
-
þ -
þ
-
-
-
þ
-
-
-
-
-
-
-
-
-
-
-
-
-
Retinal Detachment
þ
-
-
-
-
-
-
-
-
þ
-
-
-
þ
-
-
-
þ
-
-
-
-
-
Episodes of Collapse
þ
-
-
-
Weakness
Clinical Signs
CEUS in Diagnosis of Adrenal Tumors in Dogs
JAAHA.ORG
135
TABLE 2 B-mode Ultrasonographic Appearance of 24 Adrenal Lesions Investigated in 23 Dogs and Grouped According to the Reference Diagnosis B-Mode US Qualitative Variable
Adenoma (n¼10)
Adenocarcinoma (n¼7)
Pheochromocytoma (n¼7)
P
Position Left and right
2 (20.0)
0 (0.0)
0 (0.0)
Left
4 (40.0)
5 (71.4)
4 (57.1)
ns
Right
4 (40.0)
2 (28.6)
3 (42.9)
Single
8 (80.0)
7 (100.0)
7 (100.0)
Multiple
2 (20.0)
0 (0.0)
0 (0.0)
Homogeneous
6 (60.0)
1 (14.3)
4 (57.1)
Inhomogeneous
4 (40.0)
6 (85.7)
3 (42.9)
Hypoechoic
3 (30.0)
1 (14.3)
0 (0.0)
Isoechoic
0 (0.0)
0 (0.0)
1 (14.3)
Hyperechoic
4 (40.0)
2 (28.6)
4 (57.1)
Mixed Hypo/Hyperechoic
3 (30.0)
4 (57.1)
2 (28.6)
Number ns
Echotexture ns
Echogenicity ns
Margin Irregular
2 (20.0)
a
4 (57.1)
a,b
6 (85.7)
b
Well defined
8 (80.0)
a
3 (42.9)
a,b
1 (14.3)
b
0.05
Atrophy of contralateral adrenal gland Absent
7 (70.0)
5 (71.4)
7 (100.0)
Present
3 (30.0)
2 (28.6)
0 (0.0)
ns
a,b Values with the same letter in the same row are not significantly different ns ¼ not significant Descriptive analysis and Fisher’s exact test: number of adrenal lesions and percentage (in parentheses). The adrenal lesions were compared to normal surrounding adrenal tissue or contralateral adrenal gland. The single dog with a diagnosis of lipoma was excluded from the statistical analysis.
B-mode Ultrasound
(0.3 6 0.04 cm) was found in all three dogs with adrenocortical
Upon B-mode ultrasonography examination, the adrenal lesions
adenoma and in two out of three dogs with adenocarcinoma
involved the right adrenal gland in 9 dogs (37.5%), the left
(Table 1).
adrenal gland in 14 dogs (58.3%), and both adrenal glands in 1
The ultrasonographic appearances of the adrenal lesions are
dog (4.2%). In most dogs, the adrenal lesions were not precisely
provided in Table 2. Adrenocortical adenomas had well-defined
localizable within cortical or medullary region. The mean size of
margins (Figure 1A), while most adenocarcinomas showed non-
adrenocortical adenomas was 1.5 6 0.8 cm (median, 1.2 cm;
homogeneous parenchyma characterized by anechoic areas, hyper-
range, 0.7–3.1 cm), whereas that of adenocarcinomas was 2.5 6
echoic spots, septa, and areas of different echogenicity (Figure 1B).
1.4 cm (median, 2.9 cm; range 0.8–4.6 cm) and that of
The pheochromocytomas had variable echotexture and poorly
pheochromocytomas was 2.4 6 1.3 cm (median, 2.8 cm; range
defined margins (Figure 1C). In two of the seven dogs with
0.9–4.0 cm). The size of the adrenocortical lipoma was 1.7 cm.
adenocarcinoma, the tumors invaded the surrounding tissue, and
There was no statistical difference in the mean size of the adrenal
in four cases these altered the profile and surface of the adrenal
mass lesions among the diagnosis groups or between malignant
gland. Metastases were detected in surrounding renal and lumbar
and benign adrenal tumors. Of the six dogs with clinical signs of
aortic lymph nodes of one dog and in the lung of another. In four
Cushing’s syndrome, atrophy of the contralateral adrenal gland
out of seven dogs, the pheochromocytoma invaded the caudal vena
136
JAAHA
|
52:3 May/Jun 2016
CEUS in Diagnosis of Adrenal Tumors in Dogs
FIGURE 1
Representative B-
mode images of different adrenal mass lesions. (A) Adrenal adenoma (arrows). (B) Adrenal adenocarcinoma (arrows), which appeared as an heterogeneous nodular lesion with irregular margins. (C) Pheochromocytoma (white arrows) infiltrating tissues (black arrows) and invading the caudal vena cava (red arrow). (D) Adrenal lipoma (arrows) showing a well defined mass in the caudal pole of left adrenal gland with no invasion of surrounding tissues.
cava (Figure 1C) and, in these cases, the diagnosis could be made
Contrast-Enhanced Ultrasound
by B-mode ultrasound. In the dog with adrenal lipoma, ultrasound
In none of our cases did we observe adverse effects associated with
scanning characterized a well-defined mass in the caudal pole of left
the IV injection of the contrast agent. The enhancement patterns of
adrenal gland with hyperechoic margins and hypoechoic core; the adrenal surface was deformed by this mass, but there was no apparent invasion of surrounding tissues (Figure 1D). Adrenal masses with irregular margins were found more frequently (P , 0.05) in malignant than in benign lesions (Table 2). By using irregular adrenal margins of the lesions as the criteria, the B-mode
the adrenal lesions after contrast injection are shown in Table 3. The degree of contrast enhancement greatly differed (P , 0.001) among the diagnostic groups, with all adenomas and adenocarcinomas having a reduced enhancement and all pheochromocytomas having an increased enhancement (Figure 2). The distribution of
accuracy in differentiating malignant versus benign tumors was
the ultrasound contrast agent within the adrenal lesions was also
75%, with a sensitivity of 71.4% and a specificity of 80%. There
different (P , 0.05), showing either homogenous, heterogeneous,
were no statistical differences in the other qualitative ultrasound
or nodular patterns, depending on tumor type. In two adrenal
variables among the diagnosis groups or between benign and
adenomas, local perfusion defects were evidenced by the absence of
malignant adrenal lesions (Table 2).
contrast enhancement. Non-vascularized areas were identified in all
JAAHA.ORG
137
TABLE 3 Enhancement Patterns in 24 Adrenal Lesions Grouped According to the Reference Diagnosis Examined in 23 Dogs on ContrastEnhanced Ultrasonography* CEUS Qualitative Variable
Adenoma (n¼10)
Adenocarcinoma (n¼7)
Pheochromocytoma (n¼7)
P
Enhancement degree Hypo-enhancement
10 (100.0)
Iso-enhancement
0 (0.0)
§
Hyper-enhancement
0 (0.0)
§
§
7 (100.0)
§
0 (0.0) **
0 (0.0)
§
0 (0.0)
0 (0.0)
§
7 (100.0) **
0.001
§
Enhancement distribution 6 (60.0)
§
0 (0.0) **
Heterogeneous
2 (20.0)
§
4 (57.1)
§
4 (57.1)
Nodular
2 (20.0)
§
3 (42.9)
§
0 (0.0)
8 (80.0)
§
0 (0.0) **
3 (42.9) §,**
2 (20.0)
§
7 (100.0) **
4 (57.1) §,**
Centripetal
7 (70.0)
§
1 (14.3) **
0 (0.0) **
Peripheral
1 (10.0)
§
2 (28.6)
Disordered
2 (20.0)
§
4 (57.1) §,**
Homogeneous
3 (42.9) §,**
0.05
§
§
Vascularity Presence Absence
0.01
Enhancement pattern §
0 (0.0)
0.01
§
7 (100.0) **
Wash-in Slow
10 (100.0)
§
7 (100.0)
§
0 (0.0) **
Synchronous
0 (0.0)
§
0 (0.0)
§
6 (85.7) **
Fast
0 (0.0)
§
0 (0.0)
§
1 (14.3)
0 (0.0)
§
0 (0.0)
§
7 (100.0) **
0 (0.0)
§
0 (0.0)
§
0 (0.0)
0.001
§
Wash-out Slow Synchronous Fast
10 (100.0)
§
7 (100.0)
§
0.001
§
0 (0.0) **
*Descriptive analysis and Fisher’s exact test: number of adrenal lesions and percentage (in parentheses). The enhancement characteristics of the adrenal lesions were compared to normal surrounding adrenal tissue or contralateral adrenal gland. The single dog with a diagnosis of lipoma was excluded from the statistical analysis. §Values in the same row are not significantly different. **Values in the same row are not significantly different.
adrenal adenocarcinomas and also in four pheochromocytomas
the caudal vena cava. In these latter cases, the microbubbles were
(Figure 3).
clearly visualized in the arterial vessels of the pheochromocytoma
The enhancement patterns varied (P , 0.01) among adrenal
inside the caudal vena cava during the early wash-in phase; only
lesions (Table 3). In most adenomas, the enhancement arose
later did they appear within the venous system, allowing the
simultaneously from different small arteries, which maintained a ramified but centripetal direction (Figure 2A). The vascularization showed a disordered pattern in most adrenocortical adenocarcinomas and in all pheochromocytomas. In one dog with adenocarcinomas, the enhancement originated from the periphery of the lesion with small arteries directed centripetally, whereas, in
illumination of the entire caudal vena cava. Compared to the normal adrenal parenchyma, the perfusion pattern of all adrenal adenomas and adenocarcinomas was characterized by slow wash-in and fast wash-out phases, whereas that of pheochromocytomas had a slow wash-out phase, regardless of their size (Figure 2). The
two dogs with adenocarcinomas, only periphery vessels were seen
perfusion pattern of the lipoma was characterized by greatly
causing an annular enhancement pattern (Table 3 and Figure 2B).
reduced enhancement of the contrast agent during the different
In addition, CEUS examination allowed a detailed view of
phases of its vascular distribution. During the arterial phase, two
those adrenal lesions that invaded the surrounding tissues and/or
138
JAAHA
|
52:3 May/Jun 2016
vessels entering the lesions from the periphery were visualized.
CEUS in Diagnosis of Adrenal Tumors in Dogs
Representative sagittal images of the adrenal glands of three dogs showing different vascular patterns associated with adenoma (A), adenocarcinoma (B), and pheochromocytoma (C) adrenal lesions. Each image shows the contrast-enhanced ultrasound (right) and B-mode aspects (left) of the adrenal gland acquired at comparable times after contrast agent injection. (A1) The initial wash-in (17sec) shows a slow distribution of contrast enhancement within the adrenal adenoma (white arrows) compared to that of adjacent adrenal parenchyma (black arrow). (A2) Recorded
FIGURE 2
28 sec after the appearance of the ultrasound contrast agent, shows the typical vascular hypo-enhancement pattern of the adenoma also during the peak of enhancement. (A3) (36 sec) shows the wash-out period of the contrast, which, although progressive, remains hypo-enhanced compared to adjacent normal parenchyma. (B1–B3) Shows the blood flow patterns of an adenocarcinoma. In the early stages of wash-in (19 sec), large arterial vessels with disorderly distributions of contrast agent (B1 arrows) are visualized within the adrenal mass lesion. (B2) During the distribution phase (23 sec), the adenocarcinoma is constantly hypo-enhanced as a result of heterogeneous vascular areas and large necrotic areas that are not vascularized (in the dorsal part of mass lesion). (B3) Recorded 29 sec after ultrasound contrast injection; shows an adenocarcinoma characterized by a marked hypoenhancement pattern with only the margins of the mass lesion being highlighted by the contrast. (C) Shows representative blood flow patterns of a pheochromocytoma. (C1) In the early stages of wash-in (14 sec), the pheochromocytoma (white arrows) has a very pronounced vascular enhancement. During the arterial phase of contrast distribution, vascularized tissue inside the central vena cava confirms the infiltration and diffusion of the lesion within the cava; it remains without contrast medium (blue arrow). (C2) During the venous phase (29 sec), the contrast is present in the central vena cava (blue arrow) that, at this stage, appears nearly iso-enhanced with the adrenal lesion. (C3) the wash-out of the contrast from the mass lesion is homogeneous, but slow compared to normal adrenal parenchyma (53 sec).
JAAHA.ORG
139
TABLE 4 Independent Factors for Predicting the Nature of the Adrenal Lesions (Malignant Versus Benign) at Binary Logistic Regression Analysis Using Different Qualitative Enhancement Variables on Contrast-Enhanced Ultrasonography*
FIGURE 3
Representative contrast-enhanced ultrasonography
Qualitative Variable
Sensitivity
Specificity
Accuracy
P†
Enhancement degree
50.0
100.0
70.8
,0.01
Vascularity
78.6
80.0
79.2
,0.01
Enhancement Pattern
92.9
70.0
83.3
,0.01
Wash-in
50.0
100.0
70.8
,0.01
Wash-out
50.0
100.0
70.8
,0.01
*Accuracy expressed as % for correct prediction using histologic/cytologic examination as reference diagnosis. Data from the dog with a diagnosis of lipoma were not included in the statistical analysis. †Chi-Square model (goodness of fit test) for prediction of malignancy.
image of a pheochromocytoma recorded 17 sec after the injection of contrast medium showing hyper-enhancement due to high perfusion
The area under the receiver-operating characteristic curve
pattern. Small, non-vascularized areas (arrows) are present within
(AUC) for the enhancement pattern was higher (AUC ¼ 0.84, P ,
the mass, a characteristic frequently found in malignant lesions.
0.01) than those for the other categories (vascularity, enhancement degree, wash-in, wash-out, AUC ¼ 0.75, P , 0.05) and significantly different from the true area (AUC ¼ 0.5) of the null hypothesis of the reference line in discriminating malignant vs. benign adrenocortical tumors (Figure 4). Of all the CEUS variables here studied by simple logistic analysis, the enhancement pattern had the highest sensitivity (92.9%, P , 0.01) and accuracy (83.3%, P , 0.01) in predicting the nature of the adrenal lesions (malignant versus benign), while the highest specificities (100%, P , 0.01) were achieved by enhancement degree and both the wash-in and washout phases (Table 4). Including both enhancement degree and vascularity in the model, the predictive power of CEUS in discriminating malignant versus benign adrenal lesions increased (P , 0.001) achieving a sensitivity of 100, a specificity of 80, and an accuracy of 91.7%. In a multinomial logistic regression analysis, use of either contrast enhancement, wash-in, or wash out alone had an accuracy of 70.8% (P , 0.001) in differentiating adrenocortical adenomas, adrenocortical adenocarcinomas, and pheochromocytomas. When several CEUS parameters were simultaneously included (stepwise selection), the best model (P , 0.001) resulted by combining two
Graph of receiver operating characteristic curve of
enhancement categories, such as contrast enhancement degree and
different qualitative enhancement variables found on contrast-
vascularity (or vascularity and either wash-in or wash-out
enhanced ultrasonography examination of 24 adrenal mass lesions
characteristics). Using either one of these combinations, all
in 23 dogs. Positive state was malignant lesions (i.e., adenocarcinomas
adrenocortical adenocarcinomas and pheochromocytomas, as well
and pheochromocytomas); negative state was benign lesions (i.e.,
as 80% of adrenocortical adenomas, were correctly diagnosed; the
adenoma). Data from the dog with a diagnosis of lipoma were not
overall predictive accuracy of CEUS was 91.7% congruent with the
included in the statistical analysis.
reference diagnosis (P , 0.001).
FIGURE 4
140
JAAHA
|
52:3 May/Jun 2016
CEUS in Diagnosis of Adrenal Tumors in Dogs
Discussion
adrenocortical hormones.9,31 Thus, for all these reasons, a non-
As far we know, this is the first report that characterizes the
invasive and safe diagnostic tool such as CEUS, capable of
vascular perfusion patterns of different adrenal tumor types by use of CEUS in dogs. Our results indicate that this technique is useful in distinguishing between benign and malignant adrenal mass lesions, allowing the differential diagnosis between adrenocortical adenoma, adenocarcinoma, and pheochromocytoma with accuracy comparable to that found in humans using CEUS or reference CT scan and MRI.22,23 With the widespread use of ultrasound screening in dogs, the number of unexpected discoveries of adrenal lesions is greatly increasing, as has also been reported in human medicine.6 Once detected, the management of dogs with either adrenal mass or incidentalomas, which are often not showing any or only vague clinical signs, requires an appropriate diagnostic approach to differentiate between benign and malignant tumors for directing proper treatment. Although most incidental adrenal masses in humans are benign, their characterization (i.e., benign versus malignant) remains a challenging task.6 This is true also in veterinary medicine, where the clinical and economical resources are often limited. In the present study, adrenal mass lesions included adenoma, adenocarcinoma, pheochromocytoma, and lipoma, but more than half were indeed incidentalomas and not related to any specific
discriminating between the long differential list of primary versus secondary and benign versus malignant adrenal lesions, would be of great value for the clinician. In dogs, CEUS has already been used to characterize the vascular patterns of hyperplastic adrenal glands and focal lesions in the liver and spleen.15,16,32,33 In humans, the potential validity of CEUS using a second-generation contrast agent as a tool for differentiating between benign and malignant adrenal mass lesions has been evaluated only in a few published studies, but with conflicting results.22,34 Some studies have failed to find any useful vascular pattern for diagnosing the different types of adrenal mass lesions.34 In contrast, other studies, by using time-intensity curves, were able to identify malignant adrenal masses with a sensitivity of 100% and specificity of 82%, on the basis of early arterial contrast enhancement followed by fast wash-out.22 By CEUS, all of the qualitative enhancement patterns examined in our dogs differed significantly between benign and malignant adrenal lesions as well as from that of normal adrenal glands.14,15 Adrenal adenomas commonly showed hypo-enhancement, a homogenous distribution, and a centripetal pattern with an abnormal flow rate during the wash-in and wash-out phases. Most adenocarcinoma were characterized by small arterial vessels entering the adrenal lesion during the wash-in phase, disordered
clinical signs of adrenal disease. No difference in the prevalence
distribution of the enhancement pattern, and the presence of non-
between left- or right-sided adrenal masses was evident, in
vascularized areas likely due to necrotic and/or calcified regions.
agreement with previous findings.10
The presence of non-vascularized areas, however, do not represent
Although large adrenal masses associated with clinical signs of
typical landmarks that differentiate malignant from benign adrenal
adrenal dysfunction are more often indicative of malignancy, no
tumors because they can also occur in adrenal adenomas. In
definitive differentiation between benign and malignant adrenal
addition, adrenal malignant lesions might be too small to have
lesions can be made using B-mode ultrasonographic criteria
hypovascularized areas. Pheochromocytomas were mainly charac-
alone.1,3,9,10,24 Our results partly confirm this limitation of B-
terized by increased degree of enhancement, disordered distribu-
mode ultrasonography. In fact, only the findings of irregular
tion pattern, synchronous wash-in phase, and slow wash-out phase.
margins due to invasion of surrounding tissue allowed the
By B-mode, adrenal lipoma appears as a well-defined lesion with
differentiation of malignant versus benign adrenal tumors with
hyperechoic margins and hypoechoic core that can be easily
an accuracy of 75%. Other ultrasound morphological features such
misdiagnosed with nonfunctional adenoma or adenocarcinoma,
as vascular invasion and/or metastasis may predict the nature and/
based on size and echotexture. Although it may seem difficult to
or type of the adrenal lesions.25,26 The use of CT and/or MRI can
differentiate lipoma from other adrenal tumors by B-mode criteria
better characterize adrenal masses to help differentiate benign and
alone, CEUS imaging may help in diagnosing this type of lesion. In
malignant lesions in humans as well as dogs.27,28,29 Fine-needle
our case, the lipoma was characterized by a marked hypo-
aspiration of adrenal masses do not always lead to a definite
enhancement pattern during all the distribution phases compared
diagnosis, and this procedure is certainly not without risk.30 In
to normal glandular parenchyma. In addition, the lesion appeared
addition, the adrenocorticotropic hormone stimulation and/or
well-defined, without infiltration of surrounding tissues.
dexamethasone suppression tests may be inconclusive, as both
In the present study, the most specific qualitative trait for
benign and malignant adrenocortical masses can be nonfunctional
differentiating between benign and malignant adrenal lesions was
or functional, secreting excess amounts of glucocorticoids or other
the contrast enhancement pattern with a sensitivity, specificity, and
JAAHA.ORG
141
accuracy of 92.9, 70, and 83.3%, respectively. By combining the
different tumor types, especially in those cases in which the adrenal
degree of contrast enhancement with vascularity, adrenal malignant
lesion has overlapping features at B-mode imaging. There is little
lesions are evidenced with greater sensitivity, specificity, and
doubt that the differential diagnosis of the adrenal lesions with
accuracy (100, 80, and 91.7%, respectively). Contrast enhancement
CEUS in dogs can rely also on the analysis of quantitative
degree, together with vascularity (or vascularity together with either
parameters derived from time-intensity curves. However, in our
wash-in or wash-out characteristics), were also the most effective
experience, qualitative evaluation of perfusion patterns is more
qualitative variables for differentiating between adrenal adenomas,
reliable than quantification of blood flow parameters in arriving at
adrenal adenocarcinomas, and pheochromocytomas with an
a functional diagnosis sufficient to serve clinical needs. In fact,
overall accuracy in predicting the correct diagnosis of 91.7%.
actual ultrasound enhancement measurements are subject to
Interestingly, using either one of these two combinations, all
several physiological conditions (i.e., injection protocol, cardiac
adrenocortical adenocarcinomas and pheochromocytomas were
output, pharmacological sedation) and individualized ultrasound
correctly diagnosed, being congruent with the reference diagnosis.
settings (i.e., machine gains, dynamic range, gray scale, actual
The pathophysiologic bases of these findings are not clear, but
dimension of the region of interest), all of which are potential
likely reflect the different angiogenic activity of these tumors, as
sources of variability.
well as their vascular architecture and blood flow velocity. In addition, the blood perfusion patterns characterizing the different
Conclusion
adrenal tumor types were also much different from the nodular
In conclusion, results of this study confirm that CEUS is a useful,
vascular pattern found in dogs with pituitary-dependent hyperad-
non-invasive, and cost-effective tool for differentiating the types of
renocorticism.
15
benign and malignant adrenal tumors in dogs. The diagnostic
One limitation of the present study was that our dogs with
accuracy of CEUS relies on specific vascularization patterns that
adrenal mass lesions did not have a complete endocrine workup for
characterize and distinguish between nodular hyperplasia, adreno-
hyperadrenocorticism or pheochromocytoma (e.g., low- or high-
cortical adenoma, adenocarcinoma, and other types of adrenal
dose dexamethasone suppression testing, plasma adrenocortico-
tumors, such as pheochromocytoma and lipoma. Wider applica-
tropic hormone concentration, serum sex hormone values,
tion of CEUS in the diagnostic approach for the differentiation of
aldosterone, or plasma or urine catecholamine levels). Such
adrenal lesions will likely reduce the number of dogs requiring
endocrine testing likely would have helped better define the
unnecessary diagnostic procedures such as fine-needle aspiration
secretory nature of the adrenal masses identified in our dogs.
and core biopsy and/or other imaging techniques such as CT and
However, the main purpose of this study was not to determine the
MRI without compromising accurate diagnosis.
secretory nature of these adrenal masses, but, rather, to differentiate between their different types (e.g., adrenocortical adenomas,
FOOTNOTES
adrenocortical carcinoma, pheochromocytoma, and other lesions).
a b
One must remember that some adrenal mass lesions will be
c
nonsecretory, whereas others will secrete one or more adrenocor-
d
Butorphanol; Dolorex, Intervet Srl, Milan, Italy MyLab 70 X Vision Gold; Esaote, Florence, Italy Contrast agent: SonoVue; Bracco Imaging, Milan, Italy SPSS Inc., version 20.0 for Windows; Chicago, IL
tical sex hormones not routinely analyzed in the workup for dogs with hyperadrenocorticism or pheochromocytoma. This may be especially true for dogs with incidentalomas, which made up most of the cases in this study. The relatively small number of dogs enrolled in this study necessarily limits the estimation on the diagnostic accuracy of CEUS from a statistical point of view, but all our CEUS diagnosis were confirmed histopathologically. However, the diagnostic performance of qualitative CEUS analysis was similar to that reported in much larger studies obtained in human patients, suggesting that CEUS characteristics of adrenal lesions of dogs are quite similar to those in man.22,35 Therefore, we believe that CEUS should be considered in the diagnostic workup for the detection of adrenal lesions, as an aid in differentiation and evaluation of
142
JAAHA
|
52:3 May/Jun 2016
REFERENCES 1. Besso JG, Penninck DG, Gliatto JM. Retrospective ultrasonographic evaluation of adrenal lesions in 26 dogs. Vet Radiol Ultrasound 1997;38: 448–455. 2. Myers NC. Adrenal incidentalomas. Vet Clin North Am Small Animal Pract 1997;27:381–399. 3. Reusch CE, Feldman EC. Canine hyperadrenocorticism due to adrenocortical neoplasia. Pretreatment evaluation of 41 dogs. J Vet Intern Med 1991;5:3–10. 4. Tidwell AS, Penninck DG, Besso JG. Imaging of adrenal gland disorders. Vet Clin North Am Small AnimPract 1997;27:237–254. 5. Kook PH, Grest P, Raute-Kreinsen U, et al. Addison’s disease due to bilateral adrenal malignancy in a dog. J Small AnimPract 2010;51:333– 336. 6. Arnold DT, Reed JB, Burt K. Evaluation and management of the incidental adrenal mass. BUMC Proceedings 2003;16:7–12.
CEUS in Diagnosis of Adrenal Tumors in Dogs
7. Greco DS, Peterson ME, Davidson AP, et al. Concurrent pituitary and adrenal tumors in dogs with hyperadrenocorticism: 17 cases (19781995). J Am Vet Med Assoc 1999;214:1349–1353. 8. Barthez PY, Nyland TG, Feldman EC. Ultrasonographic evaluation of the adrenal glands in dogs. J Am Vet Med Assoc 1995;207:1180–1183. 9. Grooters AM, Biller DS, Theisen SK, et al. Ultrasonographic characteristics of the adrenal glands in dogs with pituitary-dependent hyperadrenocorticism: comparison with normal dogs. J Vet Intern Med 1996;10:110–115. 10. Hoerauf A, Reusch C. Ultrasonographic characteristics of both adrenal glands in 15 dogs with functional adrenocortical tumors. J Am Anim Hosp Assoc. 1999;35:193–199. 11. Schelling CG. Ultrasonography of the adrenal gland. Probl Vet Med 1991;3:604–617. 12. Meyer S, von Mach MA, Ivan D, Scha¨fer S, et al. Color-coded duplex endoscopic ultrasound of the adrenals. J Endocrinol Invest 2008;31:882– 887. 13. Gross MD, Korobkin M, Bou-Assaly W, et al. Incidentally-discovered adrenal masses. Discov Med 2010;9:24–33. 14. Pey P, Vignoli M, Haers H, et al. Contrast-enhanced ultrasonography of the normal canine adrenal gland. Vet Radiol Ultrasound 2011;52:560– 567. 15. Bargellini P, Orlandi R, Paloni C, et al. Contrast-enhanced ultrasonography of adrenal glands in dogs with pituitary-dependent Cushing’s syndrome. Vet Radiol Ultrasound 2013;54:283–292. 16. Pey P, Daminet S, Smets PM, et al. Contrast-enhanced ultrasonographic evaluation of adrenal glands in dogs with pituitary-dependent hyperadrenocorticism. Am J Vet Res 2013;74:417–425. 17. Friedrich-Rust M, Schneider G, Bohle RM, et al. Contrast-enhanced sonography of adrenal masses: differentiation of adenomas and nonadenomatous lesions. Am J Roentgenol 2008;191:1852–1860. 18. Benchekroun G, de Fornel-Thibaud P, Rodrı´guez Pi˜neiro MI, et al. Ultrasonography criteria for differentiating ACTH dependency from ACTH independency in 47 dogs with hyperadrenocorticism and equivocal adrenal asymmetry. J Vet Intern Med 2010;24(5):1077–1085. 19. Raskin RE, Meyer D. Canine and Feline Cytology: A Color Atlas and Interpretation Guide. 2nd ed. St. Louis: Saunders Elsevier; 2009. 20. Labelle, P., Kyles, A.E., Farver, T.B., et al. Indicators of malignancy of canine adrenocortical tumors: histopathology and proliferation index. Vet. Pathol 2004;41,490–497.
21. Akobeng AK. Understanding diagnostic tests 1: sensitivity, specificity and predictive values. Acta Paediatr 2007;96:338–341. 22. Friedrich-Rust M, Glasemann T, Polta A, et al. Differentiation between benign and malignant adrenal mass using contrast-enhanced ultrasound. Ultraschall Med 2011;32:460–471. 23. Taffel M, Haji-Momenian S, Nikolaidis P, et al. Adrenal imaging: a comprehensive review. Radiol Clin North Am 2012;50(2):219–243. 24. Medeiros LJ, Weiss LM. New developments in the pathologic diagnosis of adrenal cortical neoplasms. A review. Am J Clin Pathol 1992;97:73– 83. 25. Gilson SD, Withrow SF, Wheeler SL, et al. Pheochromocytoma in 50 dogs. J Vet Intern Med 1994;8:228–232. 26. Barthez PY, Marks SL, Woo J, et al. Pheochromocytoma in dogs: 61 cases (1984–1995). J Vet Intern Med 1997;11:272–278. 27. Lumachi F, Borsato S, Tregnaghi A, et al. CT-scan, MRI and imageguided FNA cytology of incidental adrenal masses. Eur J Surg Oncol 2003; 29: 689–692. 28. Schultz RM, Wisner ER, Johnson EG, et al. Contrast-enhanced computed tomography as a preoperative indicator of vascular invasion from adrenal masses in dogs. Vet Radiol Ultrasound 2009;50:625–629. 29. Spall B, Chen AV, Tucker RL, et al. Imaging diagnosis-metastatic adrenal pheochromocytoma in a dog. Vet Radiol Ultrasound 2011;52: 534–537. 30. Paulsen SD, Nghiem HV, Korobkin M, et al. Changing role of imagingguided percutaneous biopsy of adrenal masses: evaluation of 50 adrenal biopsies. Am J Roentgenol 2004;182:1033–1037. 31. Peterson ME, Gilbertson SR, Drucker WD. Plasma cortisol response to exogenous ACTH in 22 dogs with hyperadrenocorticism caused by adrenocortical neoplasia. J Am Vet Med Assoc 1982;180:542–544. 32. Ivancic M, Long F, Seiler GS. Contrast harmonic ultrasonography of splenic masses and associated liver nodules in dogs. J Am Vet Med Assoc 2009;234:88–94. 33. Nakamura K, Sasaki N, Murakami M, et al. Contrast-enhanced ultrasonography for characterization of focal splenic lesions in dogs. J Vet Intern Med 2010;24:1290–1297. 34. Dietrich CF, Ignee A, Barreiros AP, et al. Contrast-enhanced ultrasound for imaging of adrenal masses. Ultraschall Med 2010; 31: 163–168. 35. Kooistra HS, Galac S, Buijtels JJ, et al. Endocrine diseases in animals. Horm Res 2009;71 Suppl 1:144–147.
JAAHA.ORG
143