Role of imaging in the diagnosis of parotid infantile hemangiomas [623911]

Role of imaging in the diagnosis of parotid infantile hemangiomas
Fabiola C. Webera,b, Arin K. Greenea,c, Denise M. Adamsa,d, Marilyn G. Lianga,e,
Mohammed H. Alomaria,c, Stephan D. Vossf, Gulraiz Chaudrya,b,*
aVascular Anomalies Center, Boston Children's Hospital and Harvard Medical School, Boston, MA, USA
bDivision of Vascular and Interventional Radiology, Boston Children's Hospital and Harvard Medical School, Boston, MA, USA
cDepartment of Plastic and Oral Surgery, Boston Children's Hospital and Harvard Medical School, Boston, MA, USA
dDivision of Hematology/Oncology, Boston Children's Hospital and Harvard Medical School, Boston, MA, USA
eDepartment of Dermatology, Boston Children's Hospital and Harvard Medical School, Boston, MA, USA
fDepartment of Radiology, Boston Children's Hospital and Harvard Medical School, Boston, MA, USA
article info
Article history:
Received 4 July 2017Received in revised form
29 August 2017
Accepted 31 August 2017Available online 4 September 2017
Keywords:
HemangiomaImagingParotidabstract
Objectives: To review the clinical presentation, imaging and follow-up of parotid infantile hemangiomas
(IH).
Methods: Over a 15-year period, all patients with a clinical diagnosis of parotid IH were evaluated. Im-
aging was available in 35. The medical records, photographs, and radiology studies of these patients werereviewed.
Results: All patients presented at less than 4 months of age (M:F, 13:21). 19 (55)% of tumors were on the
left and were bilateral in 2 patients. The majority (29 patients) presented due to localized swelling orpalpable mass; the remainder had a cutaneous lesion, but no palpable mass at the time of presentation.
The referring diagnosis was incomplete or incorrect in 9 patients (26%).
The imaging studies all demonstrated a well-de fined homogeneous mass, with no abnormality of the
surrounding subcutaneous fat. Sonography showed a uniformly vascular lesion with pulsatile fast- flow
seen on Doppler. On MRI, the lesion was hyperintense on T2-weighted images, isointense on T1, with
intense enhancement post-contrast.Oral therapy (propranolol or corticosteroids) was prescribed in 15 (45%). Follow-up in 28 patientsdemonstrated stability of the lesion in 11, regression in size in 11 and complete involution in 6. After
involution 2 patients underwent resection of residual tissue and/or excess skin.
Conclusions: Typical clinical presentation alone may be adequate to establish a diagnosis of parotid in-
fantile hemangioma. However, in patients with no overlying cutaneous lesion, imaging can play a critical
role in con firming the diagnosis. The sonographic findings are suf ficiently characteristic to allow for a
definitive diagnosis, obviating the need for further investigations. If diagnostic uncertainty remains or
the full extent of the lesion cannot be appreciated, then MRI should be preferred over CT to avoid ionizing
radiation.
©2017 Elsevier B.V. All rights reserved.
1. Introduction
Infantile hemangioma is a benign growth of endothelial cells
and represents the most common tumor of infancy [1]. The rate of
incidence is approximately 4% with risk factors including female
gender, low-birth weight, and Caucasian race [2]. The majority of
hemangiomas are cutaneous and are most commonly found in thehead and neck [1]. Most cases present with a typical appearance of
a“strawberry ”hemangioma with expansion after birth; the diag-
nosis is therefore rarely in doubt in these cases.
Infantile hemangiomas are the most common tumor of the
parotid gland in children and ninety percent of all salivary hem-
angiomas occur in the parotid gland [3]. Parotid hemangiomas
usually appear within a few weeks of birth and are mostly
asymptomatic. Like all infantile hemangiomas they follow an active
proliferative phase during the first 6e9 months of life, followed by
an involution phase that is usually complete by 3 e5 years of age
[4e6].*Corresponding author. Division of Vascular and Interventional Radiology, Bos-
ton Children's Hospital, 300 Longwood Avenue, Boston, MA 02115, USA.
E-mail address: gulraiz.chaudry@childrens.harvard.edu (G. Chaudry).
Contents lists available at ScienceDirect
International Journal of Pediatric Otorhinolaryngology
journal homepage: http://www.ijporlonline.com/
http://dx.doi.org/10.1016/j.ijporl.2017.08.035
0165-5876/ ©2017 Elsevier B.V. All rights reserved.International Journal of Pediatric Otorhinolaryngology 102 (2017) 61 e66

The majority of parotid hemangiomas are diagnosed based on
clinical history and physical examination. However, in the absence
of an overlying cutaneous hemangioma, these present as a palpable
mass or a focal area of swelling, requiring imaging to establish a
diagnosis.
The purpose of this study was to identify the clinical presenta-
tion, imaging characteristics, and the management of parotid
hemangiomas.
2. Materials and methods
The study was compliant with the Health Insurance Portability
and Accountability Act and was approved by the institutional re-
view board. The database from the Vascular Anomalies Center was
searched for patients from with the diagnosis of parotid heman-
giomas. Over a 13 year period (2004 e2017) the database search
yielded a total of 35 patients that had imaging studies available for
review. A retrospective review of the medical records, photographs,
and radiology studies was performed. Demographic information,
clinical presentation and physical examination findings were
recorded. The imaging was reviewed independently by two pedi-atric radiologists (FW and GC). The findings documented were tu-
mor size, site and type of imaging. The lesions were evaluated for
echogenicity, calci fication, fat content, morphology, vascularity,
enhancement and involvement of local structures. On MRI, the
signal intensity on T1-weighted and T2-weighted images was
assessed relative to that of skeletal muscle. The same imaging
criteria were evaluated for follow-up imaging.
Sonography had been performed in 21 patients. The ultrasound
scans were performed on a variety of ultrasound machines with
color flow images and spectral Doppler traces obtained in all. Three
patients had a contrast enhanced CT scan completed, with sagittal
and coronal reformatting. An MRI was obtained in 23 infants, with
T1 and T2 weighted sequences performed in at least two planes,
and intravenous contrast administered in all patients.
Initial imaging evaluation of the lesion was performed in our
institution in 12 of the patients. Nine of these had an ultrasound
scan performed as the primary investigation. An MRI was per-
formed as the first imaging modality in the remainder, all of whom
presented without an associated cutaneous lesion.
3. Results3.1. Clinical presentation
All of the patients presented at 4 months of age or less, with 29%
(n¼10) presenting by 6 weeks of age. The female-to-male ratio was
approximately 2.2:1 (24 females and 11 males). Twenty (57%) of the
tumors were on the left and bilateral hemangiomas were seen in 2.
The majority (82%, n ¼29) presented due to localized swelling or
palpable mass ( Fig. 1 ). Of these only 7 infants had an overlying
cutaneous lesion ( Fig. 2 ). The remainder had a cutaneous lesion of
the cheek but no palpable mass at the time of presentation ( Fig. 2 ).
The cutaneous lesion extended to the lip in 4, neck in 3 and pinna in
2. There were no remote cutaneous lesions in any of the patients.
There was no clinical or radiological evidence of PHACE syndrome.
Twenty-six patients were referred with a correct diagnosis of
parotid hemangioma. In the remainder the referring diagnosis was
lymphadenopathy (2), lymphatic/vascular malformation (4), cheek
mass (2) and mumps (1). The findings are summarized in Table 1 .
3.2. Imaging characteristics
The sonographic images all demonstrated a well-circumscribed
mass replacing most of the visualized parotid gland ( Fig. 3 a). The
Fig. 1. 2 month old infant with right parotid swelling, but no cutaneous findings.
Fig. 2. 3 month old infant with a large cutaneous hemangioma overlying the parotid
lesion.
Table 1Demographics and results.
Parameter Value
Mean age at diagnosis (range) 2.4 months (1 day e4 months)
Female to male ratio 2.2:1Localized swelling 29 (82%)Overlying cutaneous lesion 13 (36%)Correct referring diagnosis 26 (74%)F.C. Weber et al. / International Journal of Pediatric Otorhinolaryngology 102 (2017) 61 e66 62

mass had a lobular internal structure with interspersed fine linear
echogenic septations. On color Doppler and power Doppler imag-
ing the parotid masses were hypervascular ( Fig. 3 b), with fast
(arterial) flow con firmed on spectral Doppler tracings ( Fig. 3 c).
There was no evidence of increased echogenicity or flow in the
surrounding soft tissues.
The CT studies demonstrated a lobulated mass replacing the
normal parotid tissue ( Fig. 4 ). The borders of the tumor were again
clearly delineated with no evidence of stranding in the surroundingfat. The lesion enhanced homogeneously, with prominent
enhancing vessels seen within and surrounding the hemangioma.
The MRI studies we reviewed demonstrated similar imaging
features. On the T1-weighted images the lesion was clearly delin-
eated and isointense to the adjacent skeletal muscle ( Fig. 5 a). On
fat-suppressed T2-weighted images there was enlargement of the
parotid gland, which was almost entirely replaced by the uniformly
hyperintense lesion ( Fig. 5 b). On administration of contrast, there
was intense homogeneous enhancement of the lesion ( Fig. 5 c).
Intralesional flow voids, consistent with fast flow, were seen in all
patients, although these were less prominent if the study was
performed within the first month of life. No perilesional edema, fat
stranding or abnormal enhancement was present on the studies
reviewed.
3.3. Management
Twenty (57%) of the patients were observed, with no docu-
mented medical or surgical therapy. Twelve patients with large or
enlarging mass received oral corticosteroids; additional proprano-
lol was prescribed in one. Three other patients received propranolol
only; all 3 were treated within the last 2 years of the study when
propranolol was adopted as the primary method of treatment. Early
in the included time period, one child received interferon therapy.
Core biopsies were performed in 3 children; the histopathological
findings were consistent with infantile hemangioma in 2 and
inconclusive in 1.
3.4. Follow-up
Ten patients were either lost to follow-up or were only reviewed
Fig. 3. (a) Grayscale sonographic image demonstrating well de fined mass in the right parotid gland. (b) Marked hypervascularity of the mass is seen on Color Doppler. (c) Spectral
Doppler waveforms con firm arterial flow. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 4. Post contrast axial CT image. Homogeneous enhancement of an enlarged right
parotid gland with normal appearance of the surrounding subcutaneous fat.F.C. Weber et al. / International Journal of Pediatric Otorhinolaryngology 102 (2017) 61 e66 63

in our institution to establish diagnosis or for treatment recom-
mendations. The mean and median follow-up period was 29
months and 13 months respectively (3 e120 months). Nineteen
children were followed with clinical examination alone. Follow-up
ultrasound was performed in 6 and MRI in 4. The lesion remained
stable in 11 children, of which 3 received oral corticosteroids (with
additional propranolol in 1). Regression of the lesion was seen in 12,
of which 5 received oral corticosteroids. There was complete
involution of the tumor in 6 patients, of whom only 2 received
medical treatment. Two children required surgical intervention
during the involuted phase; one due to ulceration and scarring and
the other to excise residual excess skin. Both of the surgical pro-
cedures were performed on the components of the lesion extend-
ing to the lower lip. There were no long-term complications in any
of the patients. During the follow up period, there was no clinical orradiological evidence to suggest that the initial diagnosis was
incorrect.
4. Discussion
Unfortunately, the term “hemangioma ”has been incorrectly
applied to a multitude of lesions and continues to be so. In 1982
Mulliken and Glowacki devised a biological classi fication of
vascular anomalies, differentiating tumors such as hemangiomas,
with increased cellularity and endothelial cell turnover, from
vascular malformations which arise from abnormal vascular
channels [5]. This classi fication was adopted by the International
Society for the Study of Vascular Anomalies (ISSVA) in 1996 and
biological scheme was retained in the most recent update in 2015[6]. Malformations, particularly slow flow lesions, are often evident
at birth and grow over time [5]. In contrast, infantile hemangiomas
are typically not seen at birth, but the diagnosis is usually made
within the first month due to postnatal proliferation. Cutaneous
hemangiomas are readily diagnosed by physical examination.
However, hemangiomas deep to the dermis may present as an
enlarging mass as late as 4 e5 months after birth.
In our study, we found that all parotid hemangiomas appeared
at 5 months of age or less, with half of the patients presenting with
clinical symptoms of swelling and a quarter presenting as a
palpable mass/lump. Overlying cutaneous hemangioma was seen
in only one-fourth of the patients. Previous studies have suggested
that cutaneous hemangiomas are seen in greater than fifty percent
of cases, overlying the lesion or elsewhere [4]. The relatively low
percentage of children in our study with accompanying cutaneoushemangiomas may be due to referral bias. Larger and deeper le-
sions, that require imaging, are more likely to be referred to a
specialist tertiary center. The remainder of the clinical picture is in
agreement with the published literature. It is known that parotid
hemangiomas present at birth or shortly thereafter [5]. Hemangi-
omas have a predictable pattern of growth with most appearing
during the first weeks of life and growing rapidly for 9 months
(proliferative phase). Afterwards they begin a process of regression,
better known as the involuting phase, which is usually complete in
4 years [7].
Ultrasound is the primary imaging modality in the investigation
of parotid masses and the diagnosis of infantile hemangioma can
usually be con fidently made. The normal parotid gland is homo-
geneous with echogenicity similar to that of the thyroid gland [8].
Fig. 5. (a) Pre-contrast axial T1-weighted image. A well-delineated mass, with flow voids, is seen in the left parotid gland. (b) Axial STIR image. The mass is homogeneous and
hyperintense. (c) Post contrast T1-weighted images with fat suppression. There is intense homogeneous enhancement of the mass without perilesiona li nflammatory changes.F.C. Weber et al. / International Journal of Pediatric Otorhinolaryngology 102 (2017) 61 e66 64

Intraparotid lymph nodes may be seen as multiple well de fined
hypoechoic areas [8]. In infantile hemangiomas, the gland is typi-
cally enlarged, but homogeneous in echogenicity with preservation
of its overall shape and maintenance of lobular architecture. Color
flow and spectral Doppler images con firm marked internal vascu-
larity and fast flow. If the sonographic findings include a high
resistance index on spectral Doppler, perilesional edema, or few
arteries/veins, further imaging should be obtained. Additional im-
aging may also be required to assess the full extent of the lesion or
its intricate relationship with adjacent structures.
The main differential diagnosis for infantile hemangiomas dur-
ing infancy is acute parotitis. However, clinically these infants
present with fever and swelling, with or without erythema in the
pre-auricular area [9]. On sonography, the parotid gland is typically
enlarged and heterogeneous, with enlarged intraparotid lymph
nodes. Hypoechoic foci may indicate areas of suppuration [8].
Venous and lymphatic malformations can also involve the pa-
rotid gland and present in the neonatal period or infancy. However,
the slow flow nature of these lesions is readily demonstrated by
sonography. Venous malformations are seen as multiple hypoehoic
channels that compress readily and re fill with blood on removal of
compression [10]. Phleboliths may or may not be seen. Lymphatic
malformations (LM) involving the parotid gland can be either
macrocystic or microcystic. Macrocystic LM is seen as a well-
defined, multilocular anechoic or hypoechoic lesion without in-
ternal flow. Fluid/ fluid levels may also be identi fied if there is
intralesional hemorrhage. On sonography, microcystic LM has the
appearance of an echogenic mass, with small channels, but no
discernible cysts [11].
Other benign and malignant tumors of the parotid gland are rare
in infancy. Pleomorphic adenomas are the most common non-
vascular tumor of the parotid glands, but these typically occur in
early adolescence, with a mean age of 12 years [3]. Malignant
neoplasms are rare in children and also present around the age of
10 years [8]. On sonography, malignant tumors are heterogeneous
and hypoechoic with indistinct borders [8].
In the computed tomographic (CT) imaging reviewed for this
study, we found that the hemangiomas appeared as well circum-
scribed tumors with distinct lobular architecture. There was
intense homogeneous enhancement of the tumor and large vessels
could be seen within and surrounding the parotid gland. However,
there was no evidence of perilesional edema, stranding or
enhancement. While CT is useful in demonstrating the full extent of
the tumor, its relationship to adjacent structures and any skeletal
changes, the major drawback of CT is exposure to ionizing radiation
[12].
MRI is therefore the secondary investigation of choice, as there
is no exposure to ionizing radiation and the imaging findings are
consistent. Hemangiomas are isointense to muscle on the T-1
weighted images, hyperintense on T2-weighted images and
demonstrate intense, homogeneous enhancement. Prominent flow
voids are usually present within the mass and can be accentuated
with gradient echo sequences. However, it is important to note that
flow voids may not be easily identi fied if the study is performed in
the neonatal period. Similar to prior published data we found that
this modality is the most informative given that it demonstrates
both the extent of the involvement within tissue planes and blood
flow[13]. The main drawback of MRI is that it is expensive and
sedation is required in the majority of cases.
Although the predictable regression of parotid hemangiomas
generally allows for a conservative approach with close observation
and parental reassurance, parotid hemangiomas may require
intervention to prevent a signi ficant facial deformity or to prevent
airway or feeding dif ficulties. The majority of the patients in this
study were treated before the advent of propranolol therapy, whenoral corticosteroids were prescribed as primary treatment [14,15] .
However, multiple studies have demonstrated an improved
response rate with propranolol when compared to corticosteroids
[16,17] . Propranolol is now considered first line therapy for these
tumors and the only FDA approved treatment for IH [18]. Overall it
is effective and well-tolerated and most serious side-effects can be
avoided by appropriate screening and exclusion or minimized by
appropriate monitoring [18]. The 37.5% (n ¼12) of patients
requiring drug therapy in our cohort is much higher than the 10%
requiring drug therapy in previously published studies. We attri-
bute the higher incidence to referral bias, as large tumors with a
greater potential risk of complications are more likely to be referred
to a tertiary care center.
5. Conclusion
Although age and clinical findings can be suf ficient in estab-
lishing a diagnosis of parotid hemangioma, imaging can be critical
in patients with no overlying cutaneous hemangiomas. The sono-
graphic findings and clinical history allow for a de finitive diagnosis
in the vast majority of these patients, obviating the need for further
investigation. For larger, deeper lesions cross-sectional imaging can
be performed to con firm diagnosis and identify the full extent of
the tumor, with MRI favored over CT.
Financial disclosure
The authors have no relevant con flict of interest to disclose. No
funding was provided for this work.
References
[1]S.J. Fishman, J.B. Mulliken, Hemangiomas and vascular malformations of in-
fancy and childhood, Pediatr. Surg. 40 (1993) 1177 e1200 .
[2]P. Dickison, E. Christou, O. Wargon, A prospective study of infantile heman-
giomas with a focus on incidence and risk factors, Pediatr. Dermatol. 28 (2011)
663e669.
[3]P. Ritwik, R. Brannon, A clinical analysis of nine new pediatric and adolescent
cases of benign minor salivary gland neoplasms and a review of the literature,
J. Med. Case Rep. 6 (2012) 287 .
[4]A.K. Greene, G.F. Rogers, J.B. Mulliken, Management of parotid hemangioma in
100 children, Plast. Reconstr. Surg. 113 (2004) 53 e60.
[5]J.B. Mulliken, J. Glowacki, Hemangiomas and vascular malformations in in-
fants and children: a classi fication based on endothelial characteristics, Plast.
Reconstr. Surg. 69 (1982) 412 e420.
[6]M. Wassef, F. Blei, D. Adams, A. Alomari, E. Baselga, A. Berenstein, P. Burrows,
I.J. Frieden, M.C. Garzon, J.C. Lopez-Gutierrez, D.J. Lord, S. Mitchel, J. Powell,
J. Prendiville, M. Vikkula, ISSVA Board and Scienti fic Committee, Vascular
anomalies classi fication: recommendations from the international society for
the study of vascular anomalies, Pediatrics 136 (1) (2015 Jul) e203 ee214 .
[7]R.A. Couto, R.A. Maclellan, D. Zurakowski, A.K. Greene, Infantile hemangioma:
clinical assessment of the involuting phase and implications for management,
Plast. Reconstr. Surg. 130 (2012) 619 e624.
[8]K.S. Sodhi, M. Bartlett, N.K. Prabhu, Role of high resolution ultrasound in pa-
rotid lesions in children, Inter J. Pediatr. Otorhinolaryngol. 75 (2011)1353e1358 .
[9]H. Ozdemir, Adem Karbuz, E. Ciftci, S. Fitoz, E. Ince, U. Dogru, Acute neonatal
suppurative parotitis: a case report and review of the literature, J. Infect. Dis.
15 (2011) 500 e502.
[10] J. Dubois, M. Alison, Vascular anomalies: what a radiologist needs to know,
Pediatr. Radiol. 40 (2010) 895 e905.
[11] A.M. Cahill, E.L.F. Nijs, Pediatric vascular malformations: pathophysiology,
diagnosis, and the role of interventional radiology, Cardiovasc Interv. Radiol.
34 (2011) 691 e704.
[12] D.J. Roebuck, A.T. Ahuja, Hemangioendothelioma of the parotid gland in in-
fants: sonography and correlative MR imaging, Am. J. Neuroradiol. 21 (2000)
219e223.
[13] C.D. George, Y.Y. Ng, M.A. Hall-Craggs, B.M. Jones, Parotid haemangioma in
infants: MR imaging at 1.5T, Pediatr. Radiol. 21 (1991) 483 e485.
[14] A.K. Greene, R.A. Couto, Oral prednisolone for infantile hemangioma: ef ficacy
and safety using a standardized treatment protocol, Plast. Reconstr. Surg. 128
(3) (2011 Sep) 743 e752.
[15] A.K. Greene, Current concepts of vascular anomalies, J. Craniofac. Surg. 23
(2012) 220 e224.
[16] R.R. Sadykov, F. Podmelle, R.A. Sadykov, K.R. Kasimova, H.R. Metellmann, UseF.C. Weber et al. / International Journal of Pediatric Otorhinolaryngology 102 (2017) 61 e66 65

of propranolol for the treatment infantile hemangiomas in the maxillofacial
region, Int. J. Oral Maxillofac. Surg. 42 (7) (2013 Jul) 863 e867.
[17] B.A. Drolet, P.C. Frommelt, S.L. Chamlin, A. Haggstrom, N.M. Bauman, Y.E. Chiu,
R.H. Chun, M.C. Garzon, K.E. Holland, L. Liberman, S. MacLellan-Tobert,
A.J. Mancini, D. Metry, K.B. Puttgen, M. Seefeldt, R. Sidbury, K.M. Ward, F. Blei,
E. Baselga, L. Cassidy, D.H. Darrow, S. Joachim, E.K. Kwon, K. Martin, J. Perkins,D.H. Siegel, R.J. Boucek, I.J. Frieden, Initiation and use of propranolol forinfantile hemangioma: report of a consensus conference, Pediatrics 131 (1)
(2013 Jan) 128 e140.
[18] C. L/C19eaute-Labr /C18eze, O. Boccara, C. Degrugillier-Chopinet, J. Mazereeuw-Hautier,
S. Prey, G. Lebb /C19e, S. Gautier, V. Ortis, M. Lafon, A. Montagne, A. Delarue,
J.J. Voisard, Safety of oral propranolol for the treatment of infantile heman-
gioma: a systematic review, Pediatrics (4) (2016 Oct) 138 .F.C. Weber et al. / International Journal of Pediatric Otorhinolaryngology 102 (2017) 61 e66 66