- Email: [email protected]
Facial nerve hemangiomas: Vascular tumors or malformations?
Otolaryngology–Head and Neck Surgery (2010) 142, 108-114
ORIGINAL RESEARCH–OTOLOGY AND NEUROTOLOGY
Facial nerve hemangiomas: Vascular tumors or malformations? Margo McKenna Benoit, MD, Paula E. North, MD, PhD, Michael J. McKenna, MD, Martin C. Mihm, MD, Matthew M. Johnson, MD, and Michael J. Cunningham, MD, Boston, MA; and Milwaukee, WI No sponsorships or competing interests have been disclosed for this article. ABSTRACT OBJECTIVE: To reclassify facial nerve hemangiomas in the context of presently accepted vascular lesion nomenclature by examining histology and immunohistochemical markers. STUDY DESIGN: Cohort analysis of patients diagnosed with a facial nerve hemangioma between 1990 and 2008. SETTING: Collaborative analysis at a specialty hospital and a major academic hospital. SUBJECTS AND METHODS: Seven subjects were identified on composite review of office charts, a pathology database spanning both institutions, and an encrypted patient registry. Clinical data were compiled, and hematoxylin-eosin–stained specimens were reviewed. For six patients, archived pathological tissue was available for immunohistochemical evaluation of markers specific for infantile hemangioma (glucose transporter protein isoform 1 [GLUT1] and Lewis Y antigen) and for lymphatic endothelial cells (podoplanin). RESULTS: All patients clinically presented with slowly progressive facial weakness at a mean age of 45 years without prior symptomatology. Hemotoxylin-eosin–stained histopathological slides showed irregularly shaped, dilated lesional vessels with flattened endothelial cells, scant smooth muscle, and no internal elastic lamina. Both podoplanin staining for lymphatic endothelial cells and GLUT1 and LewisY antigen staining for infantile hemangioma endothelial cells were negative in lesional vessels in all specimens for which immunohistochemical analysis was performed. CONCLUSION: Lesions of the geniculate ganglion historically referred to as “hemangiomas” do not demonstrate clinical, histopathological, or immunohistochemical features consistent with a benign vascular tumor, but instead are consistent with venous malformation. We propose that these lesions be classified as “venous vascular malformations of the facial nerve.” This nomenclature should more accurately predict clinical behavior and guide therapeutic interventions. © 2010 American Academy of Otolaryngology–Head and Neck Surgery Foundation. All rights reserved.
H
emangiomas of the temporal bone are rare lesions that can occur either near the geniculate ganglion or within the internal auditory canal (IAC). Patients with IAC hemangiomas usually present with progressive sensorineural hearing loss, whereas those with hemangiomas arising from the geniculate ganglion often manifest slowly progressive facial nerve palsy, sometimes preceded by a facial tic.1 The latter patients are often initially misdiagnosed as having idiopathic facial nerve weakness (Bell’s palsy). Progressive or recurrent facial symptoms raise the suspicion of an alternate diagnosis, prompting CT or MRI to identify the etiological vascular lesion along the course of the facial nerve. Once a certain degree of facial nerve deficit is reached, surgical excision is recommended. Pathologists have traditionally diagnosed facial nerve “hemangiomas” rather generically on the basis of hematoxylin-eosin (H&E) preparations demonstrating the presence of a benign vascular lesion composed of collections of endothelium-lined vessels in a bed of scant connective tissue stroma. Although the term hemangioma is commonly applied to these facial nerve lesions, there are clinical indicators that suggest they are vascular malformations rather than vascular tumors. A classification system for vascular lesions based on clinical, histopathological, and cytological features was introduced by Mulliken and Glowacki in 1982.2 An international multidisciplinary effort has further refined this classification to enhance the recognition and proper treatment of these lesions3,4 (Table 1). Basic tenets of this new approach are that the term hemangioma should be reserved for benign vascular tumors that arise by cellular hyperplasia, whereas the term malformation should be used for errors of vascular morphogenesis that develop in utero and persist postnatally. Benign vascular tumors (hemangiomas) include a variety of clinically and histologically distinctive entities, the most common of which is the infantile hemangioma (IH), sometimes termed cellular tumor of infancy. Typically, IH presents within a few weeks of birth, grows rapidly for the first six to 18 months of life, and then slowly involutes over a
Received August 1, 2009; revised September 15, 2009; accepted October 7, 2009.
0194-5998/$36.00 © 2010 American Academy of Otolaryngology–Head and Neck Surgery Foundation. All rights reserved. doi:10.1016/j.otohns.2009.10.007
Benoit et al
Facial nerve hemangiomas: Vascular tumors or . . .
Table 1 Classification of the international society for the study of vascular anomalies Vascular tumors Hemangioma of infancy Superficial Deep Mixed Congenital hemangioma Rapidly involuting (RICH) Non-involuting (NICH) Kaposiform hemangioendothelioma Tufted angioma Lobular capillary hemangioma Hemangiopericytoma Vascular malformations Simple malformation Capillary (port wine stain) Venous Lymphatic Microcystic Macrocystic Arteriovenous malformation (AVM) Combined malformation Capillary-lymphatic-venous Capillary-venous Capillary-venous with AV shunting Adapted from Mulliken and Glowacki,2 Chang,3 and Enjolras et al.4
period of years. The endothelial cells of IH in the proliferative phase are plump and mitotically active, whereas those in the involuting and involuted phase are mitotically quiescent and more flattened.5 On immunohistochemical assessment, IHs have been shown to stain positively for glucose transporter protein isoform 1 (GLUT1) and Lewis Y antigen (LeY) in a sensitive and specific manner.6,7 In contrast to hemangiomas, vascular malformations are developmental anomalies, usually evident at birth, that do not undergo cellular proliferation or involution. Histopathologically they are composed of abnormal collections of vessels lined by mitotically quiescent, typically flattened, endothelial cells. Their clinical presentation is determined by physical location and constituent vessel subtype (capillary, venous, lymphatic, arterial, or combined) (Table 1). Proportional growth is the norm, although disproportionate growth can occur secondary to infection, trauma, hormonal influences, or progressive hemodynamic forces. All vascular malformations stain negatively for IH-associated immunohistochemical markers such as GLUT1 and LeY. Lymphatic and venous vascular malformations are low-flow lesions that can usually be distinguished by the appearance of their constituent vessels on routine H&E histological preparations, although this can sometimes be difficult. In such cases, lymphatic endothelial differentiation can be demonstrated by immunoreactivity for podoplanin with the D2-40 antibody. We obtained archived histopathological material from
109
patients diagnosed with facial nerve hemangiomas at the Massachusetts Eye and Ear Infirmary (MEEI) and Massachusetts General Hospital (MGH) over an 18-year period. The original H&E pathological slides were reevaluated. Additionally, immunohistochemical staining was performed for GLUT1, LeY, and podoplanin. We sought to reevaluate the classification of facial nerve vascular lesions according to current criteria. Relevant histological features and endothelial immunophenotype of the component vessels, coupled with clinical presentation and behavior, formed the basis of our evaluation.
Methods Following study approval by the Human Research Institutional Review Boards at both institutions, a retrospective review was conducted to identify patients with the diagnosis of a facial nerve or geniculate hemangioma at either hospital for the years 1990 through 2008. Three main sources were used in this search. First, one of the senior author’s (M.J.M.) office records was searched by International Classification of Diseases (ICD) code for “facial nerve primary lesion” and “hemangioma.” Second, a pathology database (PowerPath, EMC Corporation, Hopkinton, MA) was searched by diagnosis (key word search for “facial nerve” and “hemangioma” or “geniculate” and “hemangioma”) or ICD code. The joining word “and” was used to specify that both criteria must be met; the joining word “or” was used such that any of the listed criteria could be met. Third, the MGH Research Patient Data Registry (RPDR), a database containing more than 2.1 million encrypted records spanning the last 13 years, was searched as shown in Table 2. Exclusion criteria included hemangioma of the skin, external auditory canal, middle ear, mastoid, or orbit; intramuscular hemangioma in the temporal fossa; or failure to undergo surgery for presumed geniculate hemangioma diagnosed on imaging. Cases selected according to
Table 2 RPDR search criteria A
B
Hemangioma of Diseases of ear, nose, intracranial structures throat\surgical Hemangioma of Miscellaneous ear, nose, unspecified site mouth and throat procedures Hemangioma of other Biopsy of middle and inner site ear Hemangioma Other diagnostic procedures on middle and inner ear RPDR, Research Patient Data Registry (Massachusetts General Hospital). One item from column A plus one item from column B were sufficient for case selection from the database.
110
Otolaryngology–Head and Neck Surgery, Vol 142, No 1, January 2010
Table 3 Patient characteristics Patient
Age at diagnosis (yrs)
Presenting symptom
Time to diagnosis
1
43
3 yrs
2 3
50 39
4
38
Facial weakness, vertigo Facial weakness Facial twitching, weakness Facial weakness
5
48
6 7
41 55
HL, facial weakness Facial weakness HL, facial weakness
Surgical approach
Preop HB
Postop HB
CAP 11%
Middle fossa
VI
IV
3 yrs 10 mo
CAP 2% N/A
Middle fossa Middle fossa
VI V
IV III
15 mo
Complete degeneration N/A
Middle fossa
III
IV
Middle fossa
IV
IV
N/A CAP 8%
Middle fossa Middle fossa
VI VI
IV III
8 yrs 3 yrs 18 mo
ENoG
CAP, compound action potential as a percentage of unaffected side; ENoG, electroneuronography; HB, House Brackmann facial nerve score; HL, hearing loss; N/A, not available; postop, postoperative; preop, preoperative.
these criteria were further narrowed by manually reviewing clinical information to confirm a diagnosis of geniculate or facial nerve hemangioma. For each patient meeting all inclusion and exclusion criteria, names and medical record numbers were crosschecked and verified between hospitals. Further clinical information was accessed from electronic and paper medical records, including inpatient and outpatient clinic notes, discharge summaries, operative reports, and formal pathology reports. Archived histopathological slides were retrieved and reviewed blindly by two pathologists experienced with vascular anomalies (M.C.M., P.E.N.). Pathological specimen blocks were obtained, re-cut, and immunohistochemically stained for podoplanin, GLUT1, and LeY antigen. Immunohistochemical staining methods have been previously published.6 Immunoreactivities were independently evaluated and scored by both pathologists (M.C.M., P.E.N.).
Results Seven patients met all inclusion and exclusion criteria. Three of these patients were identified by searching ICD office diagnosis codes. One patient was returned from the RPDR search. Six patients were identified through the pathology database. Of these six, two overlapped with office charts and one overlapped with RPDR. All seven patients in this series presented with facial nerve weakness. A geniculate lesion was identified by CT imaging with intravenous contrast dye or MRI in all patients. Table 3 shows patient characteristics including age at diagnosis, presenting symptom(s), time to diagnosis, electroneurography results, surgical approach, preoperative House-Brackmann (HB) score, and postoperative HB score at the most recent follow-up examination.
Review of H&E Histopathology Seven patients had surgical specimens available for evaluation. All original H&E-stained slides were reviewed. All specimens consisted of a vascular lesion adjacent to ganglion and neural tissue. The vascular lesions were characterized by irregularly shaped dilated vessels with scant mural smooth muscle in all cases (Fig 1). No internal elastic laminae were present, a finding consistent with venous or lymphatic, rather than arterial, origin. Lesional endothelial cells were mitotically quiescent in all seven specimens. Each pathological report from the initial reading of the specimen was obtained; all carried a diagnosis of “hemangioma.”
Immunohistochemistry Archived tissue was immunohistochemically stained in seven patients. One specimen lacked antigenicity even in internal controls (possibly owing to specimen age) and, therefore, was excluded from further analysis. Podoplanin staining utilizing the D2-40 antibody was negative for endothelial cells in all lesional vessels, confirming lack of lymphatic differentiation. Endothelial cells in normal lymphatic vessels within the tissue were podoplanin-positive and served as internal positive controls (Fig 2). These same six specimens were successfully immunoreacted for two markers of IHs: GLUT1 and LeY antigen. The endothelial lining of the lesional vascular channels was negative for GLUT1 and LeY antigen in all six cases (Figs 3 and 4).
Discussion Vascular lesions of the geniculate ganglion are uncommon. It is often difficult to distinguish these lesions from other temporal bone lesions on the basis of current imaging meth-
Benoit et al
Facial nerve hemangiomas: Vascular tumors or . . .
111
odologies. A relatively small vascular lesion can produce profound functional impairment because of focused pressure or, possibly, arterial “vascular steal,” a phenomenon that compromises blood flow to the facial nerve and results in ischemic insult.8 High-resolution CT imaging of the temporal bone can often identify discreet lesions or diffuse enhancement of the facial nerve. So-called cavernous hemangiomas of the IAC are iso- or hyperintense on T1weighted MRI images, are hyperintense on T2-weighted images, and enhance with gadolinium contrast.9 Vascular malformations are typically classified as high-flow or lowflow on MRI on the basis of the presence of signal voids. Lymphatic and venous malformations are low-flow lesions that are typically hypointense on T1-weighted and hyperintense on T2-weighted images, and both may enhance with intravenous contrast.10 The use of high-flow Doppler ultrasound can often distinguish hemangiomas and the different types of vascular malformations within soft tissue structures;11 however, this technique is not applicable to temporal bone lesions of the geniculate or IAC. Histopathological evaluation provides more definitive information to aid in the differential diagnosis of vascular lesions, regardless of location. A diagnosis of “facial nerve hemangioma” has historically been given to geniculate lesions characterized by multiple dilated vascular channels on H&E stains. Further examination of such H&E slides, however, reveals certain features that support the alternative
Figure 1 (A) Representative image from case 5 showing dilated thin-walled vascular spaces (arrow) with intervening hypocellular collagenous stroma. (Hematoxylin-eosin stain; original magnification: ⫻40.) (B) Representative higher-power image of case 5 emphasizing irregular dilated vascular spaces lacking elastic lamina. (Hematoxylin-eosin stain; original magnification: ⫻100.)
Figure 2 Immunohistochemistry for podoplanin showing absent staining in the cells lining the vascular spaces (asterisk) and positive staining surrounding ganglion cells at right (arrow). (Original magnification: ⫻400.)
112
Otolaryngology–Head and Neck Surgery, Vol 142, No 1, January 2010
Figure 3 Immunohistochemistry for GLUT1 showing absence of staining in the cells lining the vascular spaces (asterisk), with positive staining of scattered erythrocytes (arrow). (Original magnification: ⫻400.)
classification of these lesions as vascular malformations rather than vascular tumors (i.e., hemangiomas). IH (i.e., the classic cellular hemangioma of infancy) would be expected to show proliferating endothelial cells with frequent mitoses in the proliferative phase or, alternatively, expanded fibrofatty tissue with occasional residual capillaries in the involutional phase.6,12 Instead, the facial nerve vascular lesions in our study display widely ectatic vascular channels rimmed by thin smooth muscle coats without evident elastic laminae, as well as flattened and mitotically quiescent endothelial cells. These features are consistent with either lymphatic or venous vascular malformation. Lymphatic malformations may have proteinaceous fluid and lymphocytes within the dilated vascular channels, whereas venous malformations can show multiple thrombi in various stages of organization.13 In the absence of relevant clinical information, the distinction between hemangiomas and vascular malformations, and the differentiation among the various types of vascular malformations, is difficult on the basis of histopathological features alone. Immunohistochemical profiling can provide additional valuable diagnostic information. In the past decade, important advancements have been made in the characterization of vascular lesions using immunohistochemical markers. In a large retrospective immunohistochemical study of a wide variety of vascular tumors and malformations, North and colleagues7 reported that GLUT1 is universally and strongly expressed by endothelial
cells in all IH specimens but is not expressed at immunohistochemically detectable levels in other types of benign vascular tumors, nor in vascular malformations, either venous or lymphatic. GLUT1 is an erythrocyte-type glucose transporter protein that is expressed by endothelial cells at sites of blood-tissue barriers such as the CNS, placenta, retina, and in peripheral nerves, but not in the vasculature of skin and soft tissue.7 IHs express GLUT1 in all phases, and its glucose transport function may confer a growth or survival advantage in the proliferative phase. In addition, IHs have intense immunoreactivity for a number of antigens also shared by the placental vasculature, including LeY,6 an oligosaccharide that is also present in epithelia and activated T cells. In contrast, the D2-40 antibody, a monoclonal antibody directed against the antigen podoplanin that is expressed by lymphatic endothelial cells,14 does not stain the endothelial cells of nonlymphatic vascular malformation subtypes nor those of hemangiomas of any variety. Reassessment of the H&E histological preparations in this study revealed findings that were inconsistent with the original diagnosis of “hemangioma.” The lesional vessels were not capillaries; rather, they were grossly dilated vessels with thin smooth muscle coats and a lining of flattened endothelial cells without evident endothelial mitotic activity. Immunohistochemical staining was used to further guide proper classification. GLUT1 and LeY were negative in all lesions, confirming that these lesions were not residual
Figure 4 Immunohistochemistry for Lewis Y antigen showing absent staining in the cells lining the vascular spaces (asterisk), with positive epithelial staining in a ductal structure at the bottom of the field (arrow). (Original magnification: ⫻400.)
Benoit et al
Facial nerve hemangiomas: Vascular tumors or . . .
IH. The podoplanin stains performed on all specimens were also negative, arguing against a lymphatic origin. Although no specific stain for venous vascular malformations has been identified and validated in the literature to date, the histological appearance and immunohistochemical profile outlined above validate the conclusion that all of the lesions evaluated in this study are most accurately classified as venous malformations, and that they are inconsistent with vascular tumors (hemangiomas) and lymphatic or arteriovenous malformations. A variant temporal bone lesion known as “osseous” or “ossifying” hemangioma is also reported in the literature. Such lesions have a characteristic honeycomb or sunburst radiographic appearance attributed to intralesional lamellar bony trabeculae documented on histopathological examination.15 Osseous features were not present in any of our specimens, so it is unknown whether such ossifying hemangiomas would likewise demonstrate the vascular morphology and immmunohistochemical staining pattern seen in the specimens we have evaluated. Correct nosological classification of geniculate vascular lesions will provide clinicians and researchers with a better understanding of the biological origin and pathophysiology of these lesions and help direct rational therapies. Multiple factors have been identified that play a role in abnormal vessel growth and development, including vascular endothelial growth receptors, tumor growth factor-b1, and Tie2 signaling pathways, and the Notch3 gene, which mediates interactions between endothelial and smooth muscle cells.12,16 Clinical trials are underway to investigate the effects of agents that modify angiogenesis pathways (www. cancer.gov/clinicaltrials). Some of these agents may be effective in treating mitotically active vascular tumors (hemangiomas), but not in treating long-standing vascular developmental anomalies (malformations) that cause symptoms in adolescence and adulthood owing to slow vascular expansion in response to hemodynamic or other factors. Distinguishing between hemangiomas and vascular malformations on a histological and molecular level is critical both in guiding therapy for individual patients and in designing therapeutic trials. Correctly identifying facial nerve lesions as venous malformations rather than hemangiomas has important clinical management implications. Patients with lesions of this type can be correctly counseled about the origin and expected natural history of their disorder. This is particularly important in this era when information from nonphysician sources such as the Internet plays such a large role in patient education. Patients carrying a diagnosis of facial nerve hemangioma may come to have false expectations about the possible spontaneous involution of their lesion, or they may be confused by the onset of symptoms relatively late in life. Providing patients with an accurate diagnosis allows them to pursue factual information relevant to their disease entity. An accurate diagnosis also facilitates communication among clinicians. It is interesting to note that two previous
113
reports did suggest that these IAC and geniculate lesions are actually vascular malformations.15,17 The traditional nomenclature referring to the lesions as “hemangiomas” has, however, persisted. More recent literature reclassifying hemangiomas as vascular malformations at various anatomical sites demonstrates increasing effort among all disciplines to correctly identify and classify these lesions.18,19 There are two main limitations to the present study. One is the small sample size. Given the rarity of this clinical entity, we could identify only seven cases of facial nerve or geniculate hemangioma requiring surgical intervention and, hence, providing pathological specimens over the past 18 years. There were additional patients with a suspected vascular lesion of the geniculate ganglion who did not warrant or refused surgery. Given the relatively long time interval employed and the incorporation of data from two tertiary referral institutions with a busy joint neuro-otological service, we think it is unlikely that a significantly larger series of patients could be obtained. The second potential limitation of this study is that the immunohistochemical evidence supporting the diagnosis of venous malformation is based on negative results—in other words, on absence of staining. It could be argued that the negative results are due to inadequate processing or staining techniques. However, our evaluation ruled out this possibility by requiring that the expected positive staining of appropriate internal positive controls for all immunoreactions be strongly present. No specific immunohistochemical stain for venous vessels, at least in routinely fixed, paraffinembedded tissue, is available. The characteristic morphology of the lesional vascular channels, coupled with demonstration of lack of expression of the lymphatic marker podoplanin, is sufficient for the diagnosis of venous malformation in this location, just as it is in the more common cutaneous and subcutaneous lesions.13 In summary, lesions of the facial nerve and geniculate ganglion historically referred to as “hemangiomas” do not share any of the clinical, histopathological, or immunohistochemical features characteristic of IHs or other vascular tumors. Such lesions should be more appropriately classified as venous vascular malformations of the facial nerve consistent with current terminology. This nomenclature carries prognostic and therapeutic implications for patients as well as for clinicians in the fields of both otolaryngology and pathology. All research was supported by departmental funding.
Author Information From the Department of Otology and Laryngology, Harvard Medical School (Drs Benoit, McKenna, and Cunningham), Boston, MA; the Department of Otolaryngology, Massachusetts Eye and Ear Infirmary (Drs Benoit, McKenna, and Cunningham), Boston, MA; the Department of Pathology, Medical College of Wisconsin (Dr North), Milwaukee, WI; and the Department of Pathology, Massachusetts General Hospital (Drs Mihm and Johnson), Boston, MA.
114
Otolaryngology–Head and Neck Surgery, Vol 142, No 1, January 2010
Corresponding author: Margo McKenna Benoit, Massachusetts Eye and Ear Infirmary, Department of Otolaryngology, 243 Charles St, Boston, MA 02114. E-mail address: [email protected]. This article was accepted for presentation at the 2009 AAO-HNSF Annual Meeting & OTO EXPO, San Diego, CA, October 4-7, 2009.
Author Contributions Margo McKenna Benoit, conception and design, acquisition of data, analysis and interpretation, drafting of manuscript, final approval; Michael J. McKenna, conception and design, acquisition of data, analysis and interpretation, drafting of manuscript, final approval; Martin C. Mihm, conception and design, acquisition of data, analysis and interpretation, drafting of manuscript, final approval; Paula E. North, acquisition of data, analysis and interpretation, drafting of manuscript, final approval; Matthew M. Johnson, acquisition of data, analysis and interpretation, drafting of manuscript, final approval; Michael J. Cunningham, conception and design, analysis and interpretation, drafting of manuscript, final approval.
Disclosures Competing interests: None. Sponsorships: None.
References 1. Friedman O, Neff BA, Willcox TO, et al. Temporal bone hemangiomas involving the facial nerve. Otol Neurotol 2002;23:760 – 6. 2. Mulliken JB, Glowacki J. Hemangiomas and vascular malformations in infants and children: a classification based on endothelial characteristics. Plast Reconstr Surg 1982;69:412–22. 3. Chang MW. Updated classification of hemangiomas and other vascular anomalies. Lymph Res Biol 2003;1:259 – 65. 4. Enjolras O, Wassef M, Chapot R, editors. Color atlas of vascular tumors and vascular malformations. Cambridge: Cambridge University Press; 2007.
5. North PE, Waner M, Buckmiller L, et al. Vascular tumors of infancy and childhood: beyond capillary hemangioma. Cardiovasc Pathol 2006;15:303–17. 6. North PE, Waner M, Mizeracki A, et al. A unique microvascular phenotype shared by juvenile hemangiomas and human placenta. Arch Dermatol 2001;137:559 –70. 7. North PE, Milton W, Mizeracki A, et al. GLUT1: a newly discovered immunohistochemical marker for juvenile hemangiomas. Hum Pathol 2000;31:11–22. 8. Jackler RK, Brackmann DE, editors. Neurotology. Saint Louis: Mosby; 1994. p. 1323– 4. 9. Omojola MF, Al Hawashim NS, Al Zuwayed M, et al. CT and MRI features of cavernous haemangioma of the internal auditory canal. Br J Radiol 1997;70:1184 –7. 10. Baker LL, Dillon WP, Hieshima GB, et al. Hemangiomas and vascular malformations of the head and neck: MR characterization. Am J Neuroradiol 1993;14:307–14. 11. Dubois J, Garel L. Imaging and therapeutic approach of hemangiomas and vascular malformations in the pediatric age group. Pediatr Radiol 1999;29:879 –93. 12. Tille JC, Pepper MS. Hereditary vascular anomalies: new insights into their pathogenesis. Arterioscl Thromb Vasc Biol 2004;24:1578 –90. 13. North PE, Mihm MC. Histopathological diagnosis of infantile hemangiomas and vascular malformations. Facial Plast Surg Clin North Am 2001;9:505–24. 14. Kahn HJ, Bailey D, Marks A. Monoclonal antibody D2-40, a new marker of lymphatic endothelium, reacts with Kaposi’s sarcoma and a subset of angiosarcomas. Mod Pathol 2002;15:434 – 40. 15. Eby TL, Fisch U, Makek MS. Facial nerve management in temporal bone hemangiomas. Am J Otol 1992;13:223–32. 16. Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000;407:249 –57. 17. Mangham CA, Carberry JN, Brackmann DE. Management of intratemporal vascular tumors. Laryngoscope 1981;91:867–76. 18. Yeo A, Majithia A, Kalan A. Haemangioma or vascular malformation of the tympanic membrane? A case report and review of the literature. Indian J Otolaryngol Head Neck Surg 2008;60:59 – 61. 19. Greene AK, Rogers GF, Mulliken JB. Intraosseus “hemangiomas” are malformations and not tumors. Plast Reconstr Surg 2007;119: 1949 –50.
ORIGINAL RESEARCH–OTOLOGY AND NEUROTOLOGY
Facial nerve hemangiomas: Vascular tumors or malformations? Margo McKenna Benoit, MD, Paula E. North, MD, PhD, Michael J. McKenna, MD, Martin C. Mihm, MD, Matthew M. Johnson, MD, and Michael J. Cunningham, MD, Boston, MA; and Milwaukee, WI No sponsorships or competing interests have been disclosed for this article. ABSTRACT OBJECTIVE: To reclassify facial nerve hemangiomas in the context of presently accepted vascular lesion nomenclature by examining histology and immunohistochemical markers. STUDY DESIGN: Cohort analysis of patients diagnosed with a facial nerve hemangioma between 1990 and 2008. SETTING: Collaborative analysis at a specialty hospital and a major academic hospital. SUBJECTS AND METHODS: Seven subjects were identified on composite review of office charts, a pathology database spanning both institutions, and an encrypted patient registry. Clinical data were compiled, and hematoxylin-eosin–stained specimens were reviewed. For six patients, archived pathological tissue was available for immunohistochemical evaluation of markers specific for infantile hemangioma (glucose transporter protein isoform 1 [GLUT1] and Lewis Y antigen) and for lymphatic endothelial cells (podoplanin). RESULTS: All patients clinically presented with slowly progressive facial weakness at a mean age of 45 years without prior symptomatology. Hemotoxylin-eosin–stained histopathological slides showed irregularly shaped, dilated lesional vessels with flattened endothelial cells, scant smooth muscle, and no internal elastic lamina. Both podoplanin staining for lymphatic endothelial cells and GLUT1 and LewisY antigen staining for infantile hemangioma endothelial cells were negative in lesional vessels in all specimens for which immunohistochemical analysis was performed. CONCLUSION: Lesions of the geniculate ganglion historically referred to as “hemangiomas” do not demonstrate clinical, histopathological, or immunohistochemical features consistent with a benign vascular tumor, but instead are consistent with venous malformation. We propose that these lesions be classified as “venous vascular malformations of the facial nerve.” This nomenclature should more accurately predict clinical behavior and guide therapeutic interventions. © 2010 American Academy of Otolaryngology–Head and Neck Surgery Foundation. All rights reserved.
H
emangiomas of the temporal bone are rare lesions that can occur either near the geniculate ganglion or within the internal auditory canal (IAC). Patients with IAC hemangiomas usually present with progressive sensorineural hearing loss, whereas those with hemangiomas arising from the geniculate ganglion often manifest slowly progressive facial nerve palsy, sometimes preceded by a facial tic.1 The latter patients are often initially misdiagnosed as having idiopathic facial nerve weakness (Bell’s palsy). Progressive or recurrent facial symptoms raise the suspicion of an alternate diagnosis, prompting CT or MRI to identify the etiological vascular lesion along the course of the facial nerve. Once a certain degree of facial nerve deficit is reached, surgical excision is recommended. Pathologists have traditionally diagnosed facial nerve “hemangiomas” rather generically on the basis of hematoxylin-eosin (H&E) preparations demonstrating the presence of a benign vascular lesion composed of collections of endothelium-lined vessels in a bed of scant connective tissue stroma. Although the term hemangioma is commonly applied to these facial nerve lesions, there are clinical indicators that suggest they are vascular malformations rather than vascular tumors. A classification system for vascular lesions based on clinical, histopathological, and cytological features was introduced by Mulliken and Glowacki in 1982.2 An international multidisciplinary effort has further refined this classification to enhance the recognition and proper treatment of these lesions3,4 (Table 1). Basic tenets of this new approach are that the term hemangioma should be reserved for benign vascular tumors that arise by cellular hyperplasia, whereas the term malformation should be used for errors of vascular morphogenesis that develop in utero and persist postnatally. Benign vascular tumors (hemangiomas) include a variety of clinically and histologically distinctive entities, the most common of which is the infantile hemangioma (IH), sometimes termed cellular tumor of infancy. Typically, IH presents within a few weeks of birth, grows rapidly for the first six to 18 months of life, and then slowly involutes over a
Received August 1, 2009; revised September 15, 2009; accepted October 7, 2009.
0194-5998/$36.00 © 2010 American Academy of Otolaryngology–Head and Neck Surgery Foundation. All rights reserved. doi:10.1016/j.otohns.2009.10.007
Benoit et al
Facial nerve hemangiomas: Vascular tumors or . . .
Table 1 Classification of the international society for the study of vascular anomalies Vascular tumors Hemangioma of infancy Superficial Deep Mixed Congenital hemangioma Rapidly involuting (RICH) Non-involuting (NICH) Kaposiform hemangioendothelioma Tufted angioma Lobular capillary hemangioma Hemangiopericytoma Vascular malformations Simple malformation Capillary (port wine stain) Venous Lymphatic Microcystic Macrocystic Arteriovenous malformation (AVM) Combined malformation Capillary-lymphatic-venous Capillary-venous Capillary-venous with AV shunting Adapted from Mulliken and Glowacki,2 Chang,3 and Enjolras et al.4
period of years. The endothelial cells of IH in the proliferative phase are plump and mitotically active, whereas those in the involuting and involuted phase are mitotically quiescent and more flattened.5 On immunohistochemical assessment, IHs have been shown to stain positively for glucose transporter protein isoform 1 (GLUT1) and Lewis Y antigen (LeY) in a sensitive and specific manner.6,7 In contrast to hemangiomas, vascular malformations are developmental anomalies, usually evident at birth, that do not undergo cellular proliferation or involution. Histopathologically they are composed of abnormal collections of vessels lined by mitotically quiescent, typically flattened, endothelial cells. Their clinical presentation is determined by physical location and constituent vessel subtype (capillary, venous, lymphatic, arterial, or combined) (Table 1). Proportional growth is the norm, although disproportionate growth can occur secondary to infection, trauma, hormonal influences, or progressive hemodynamic forces. All vascular malformations stain negatively for IH-associated immunohistochemical markers such as GLUT1 and LeY. Lymphatic and venous vascular malformations are low-flow lesions that can usually be distinguished by the appearance of their constituent vessels on routine H&E histological preparations, although this can sometimes be difficult. In such cases, lymphatic endothelial differentiation can be demonstrated by immunoreactivity for podoplanin with the D2-40 antibody. We obtained archived histopathological material from
109
patients diagnosed with facial nerve hemangiomas at the Massachusetts Eye and Ear Infirmary (MEEI) and Massachusetts General Hospital (MGH) over an 18-year period. The original H&E pathological slides were reevaluated. Additionally, immunohistochemical staining was performed for GLUT1, LeY, and podoplanin. We sought to reevaluate the classification of facial nerve vascular lesions according to current criteria. Relevant histological features and endothelial immunophenotype of the component vessels, coupled with clinical presentation and behavior, formed the basis of our evaluation.
Methods Following study approval by the Human Research Institutional Review Boards at both institutions, a retrospective review was conducted to identify patients with the diagnosis of a facial nerve or geniculate hemangioma at either hospital for the years 1990 through 2008. Three main sources were used in this search. First, one of the senior author’s (M.J.M.) office records was searched by International Classification of Diseases (ICD) code for “facial nerve primary lesion” and “hemangioma.” Second, a pathology database (PowerPath, EMC Corporation, Hopkinton, MA) was searched by diagnosis (key word search for “facial nerve” and “hemangioma” or “geniculate” and “hemangioma”) or ICD code. The joining word “and” was used to specify that both criteria must be met; the joining word “or” was used such that any of the listed criteria could be met. Third, the MGH Research Patient Data Registry (RPDR), a database containing more than 2.1 million encrypted records spanning the last 13 years, was searched as shown in Table 2. Exclusion criteria included hemangioma of the skin, external auditory canal, middle ear, mastoid, or orbit; intramuscular hemangioma in the temporal fossa; or failure to undergo surgery for presumed geniculate hemangioma diagnosed on imaging. Cases selected according to
Table 2 RPDR search criteria A
B
Hemangioma of Diseases of ear, nose, intracranial structures throat\surgical Hemangioma of Miscellaneous ear, nose, unspecified site mouth and throat procedures Hemangioma of other Biopsy of middle and inner site ear Hemangioma Other diagnostic procedures on middle and inner ear RPDR, Research Patient Data Registry (Massachusetts General Hospital). One item from column A plus one item from column B were sufficient for case selection from the database.
110
Otolaryngology–Head and Neck Surgery, Vol 142, No 1, January 2010
Table 3 Patient characteristics Patient
Age at diagnosis (yrs)
Presenting symptom
Time to diagnosis
1
43
3 yrs
2 3
50 39
4
38
Facial weakness, vertigo Facial weakness Facial twitching, weakness Facial weakness
5
48
6 7
41 55
HL, facial weakness Facial weakness HL, facial weakness
Surgical approach
Preop HB
Postop HB
CAP 11%
Middle fossa
VI
IV
3 yrs 10 mo
CAP 2% N/A
Middle fossa Middle fossa
VI V
IV III
15 mo
Complete degeneration N/A
Middle fossa
III
IV
Middle fossa
IV
IV
N/A CAP 8%
Middle fossa Middle fossa
VI VI
IV III
8 yrs 3 yrs 18 mo
ENoG
CAP, compound action potential as a percentage of unaffected side; ENoG, electroneuronography; HB, House Brackmann facial nerve score; HL, hearing loss; N/A, not available; postop, postoperative; preop, preoperative.
these criteria were further narrowed by manually reviewing clinical information to confirm a diagnosis of geniculate or facial nerve hemangioma. For each patient meeting all inclusion and exclusion criteria, names and medical record numbers were crosschecked and verified between hospitals. Further clinical information was accessed from electronic and paper medical records, including inpatient and outpatient clinic notes, discharge summaries, operative reports, and formal pathology reports. Archived histopathological slides were retrieved and reviewed blindly by two pathologists experienced with vascular anomalies (M.C.M., P.E.N.). Pathological specimen blocks were obtained, re-cut, and immunohistochemically stained for podoplanin, GLUT1, and LeY antigen. Immunohistochemical staining methods have been previously published.6 Immunoreactivities were independently evaluated and scored by both pathologists (M.C.M., P.E.N.).
Results Seven patients met all inclusion and exclusion criteria. Three of these patients were identified by searching ICD office diagnosis codes. One patient was returned from the RPDR search. Six patients were identified through the pathology database. Of these six, two overlapped with office charts and one overlapped with RPDR. All seven patients in this series presented with facial nerve weakness. A geniculate lesion was identified by CT imaging with intravenous contrast dye or MRI in all patients. Table 3 shows patient characteristics including age at diagnosis, presenting symptom(s), time to diagnosis, electroneurography results, surgical approach, preoperative House-Brackmann (HB) score, and postoperative HB score at the most recent follow-up examination.
Review of H&E Histopathology Seven patients had surgical specimens available for evaluation. All original H&E-stained slides were reviewed. All specimens consisted of a vascular lesion adjacent to ganglion and neural tissue. The vascular lesions were characterized by irregularly shaped dilated vessels with scant mural smooth muscle in all cases (Fig 1). No internal elastic laminae were present, a finding consistent with venous or lymphatic, rather than arterial, origin. Lesional endothelial cells were mitotically quiescent in all seven specimens. Each pathological report from the initial reading of the specimen was obtained; all carried a diagnosis of “hemangioma.”
Immunohistochemistry Archived tissue was immunohistochemically stained in seven patients. One specimen lacked antigenicity even in internal controls (possibly owing to specimen age) and, therefore, was excluded from further analysis. Podoplanin staining utilizing the D2-40 antibody was negative for endothelial cells in all lesional vessels, confirming lack of lymphatic differentiation. Endothelial cells in normal lymphatic vessels within the tissue were podoplanin-positive and served as internal positive controls (Fig 2). These same six specimens were successfully immunoreacted for two markers of IHs: GLUT1 and LeY antigen. The endothelial lining of the lesional vascular channels was negative for GLUT1 and LeY antigen in all six cases (Figs 3 and 4).
Discussion Vascular lesions of the geniculate ganglion are uncommon. It is often difficult to distinguish these lesions from other temporal bone lesions on the basis of current imaging meth-
Benoit et al
Facial nerve hemangiomas: Vascular tumors or . . .
111
odologies. A relatively small vascular lesion can produce profound functional impairment because of focused pressure or, possibly, arterial “vascular steal,” a phenomenon that compromises blood flow to the facial nerve and results in ischemic insult.8 High-resolution CT imaging of the temporal bone can often identify discreet lesions or diffuse enhancement of the facial nerve. So-called cavernous hemangiomas of the IAC are iso- or hyperintense on T1weighted MRI images, are hyperintense on T2-weighted images, and enhance with gadolinium contrast.9 Vascular malformations are typically classified as high-flow or lowflow on MRI on the basis of the presence of signal voids. Lymphatic and venous malformations are low-flow lesions that are typically hypointense on T1-weighted and hyperintense on T2-weighted images, and both may enhance with intravenous contrast.10 The use of high-flow Doppler ultrasound can often distinguish hemangiomas and the different types of vascular malformations within soft tissue structures;11 however, this technique is not applicable to temporal bone lesions of the geniculate or IAC. Histopathological evaluation provides more definitive information to aid in the differential diagnosis of vascular lesions, regardless of location. A diagnosis of “facial nerve hemangioma” has historically been given to geniculate lesions characterized by multiple dilated vascular channels on H&E stains. Further examination of such H&E slides, however, reveals certain features that support the alternative
Figure 1 (A) Representative image from case 5 showing dilated thin-walled vascular spaces (arrow) with intervening hypocellular collagenous stroma. (Hematoxylin-eosin stain; original magnification: ⫻40.) (B) Representative higher-power image of case 5 emphasizing irregular dilated vascular spaces lacking elastic lamina. (Hematoxylin-eosin stain; original magnification: ⫻100.)
Figure 2 Immunohistochemistry for podoplanin showing absent staining in the cells lining the vascular spaces (asterisk) and positive staining surrounding ganglion cells at right (arrow). (Original magnification: ⫻400.)
112
Otolaryngology–Head and Neck Surgery, Vol 142, No 1, January 2010
Figure 3 Immunohistochemistry for GLUT1 showing absence of staining in the cells lining the vascular spaces (asterisk), with positive staining of scattered erythrocytes (arrow). (Original magnification: ⫻400.)
classification of these lesions as vascular malformations rather than vascular tumors (i.e., hemangiomas). IH (i.e., the classic cellular hemangioma of infancy) would be expected to show proliferating endothelial cells with frequent mitoses in the proliferative phase or, alternatively, expanded fibrofatty tissue with occasional residual capillaries in the involutional phase.6,12 Instead, the facial nerve vascular lesions in our study display widely ectatic vascular channels rimmed by thin smooth muscle coats without evident elastic laminae, as well as flattened and mitotically quiescent endothelial cells. These features are consistent with either lymphatic or venous vascular malformation. Lymphatic malformations may have proteinaceous fluid and lymphocytes within the dilated vascular channels, whereas venous malformations can show multiple thrombi in various stages of organization.13 In the absence of relevant clinical information, the distinction between hemangiomas and vascular malformations, and the differentiation among the various types of vascular malformations, is difficult on the basis of histopathological features alone. Immunohistochemical profiling can provide additional valuable diagnostic information. In the past decade, important advancements have been made in the characterization of vascular lesions using immunohistochemical markers. In a large retrospective immunohistochemical study of a wide variety of vascular tumors and malformations, North and colleagues7 reported that GLUT1 is universally and strongly expressed by endothelial
cells in all IH specimens but is not expressed at immunohistochemically detectable levels in other types of benign vascular tumors, nor in vascular malformations, either venous or lymphatic. GLUT1 is an erythrocyte-type glucose transporter protein that is expressed by endothelial cells at sites of blood-tissue barriers such as the CNS, placenta, retina, and in peripheral nerves, but not in the vasculature of skin and soft tissue.7 IHs express GLUT1 in all phases, and its glucose transport function may confer a growth or survival advantage in the proliferative phase. In addition, IHs have intense immunoreactivity for a number of antigens also shared by the placental vasculature, including LeY,6 an oligosaccharide that is also present in epithelia and activated T cells. In contrast, the D2-40 antibody, a monoclonal antibody directed against the antigen podoplanin that is expressed by lymphatic endothelial cells,14 does not stain the endothelial cells of nonlymphatic vascular malformation subtypes nor those of hemangiomas of any variety. Reassessment of the H&E histological preparations in this study revealed findings that were inconsistent with the original diagnosis of “hemangioma.” The lesional vessels were not capillaries; rather, they were grossly dilated vessels with thin smooth muscle coats and a lining of flattened endothelial cells without evident endothelial mitotic activity. Immunohistochemical staining was used to further guide proper classification. GLUT1 and LeY were negative in all lesions, confirming that these lesions were not residual
Figure 4 Immunohistochemistry for Lewis Y antigen showing absent staining in the cells lining the vascular spaces (asterisk), with positive epithelial staining in a ductal structure at the bottom of the field (arrow). (Original magnification: ⫻400.)
Benoit et al
Facial nerve hemangiomas: Vascular tumors or . . .
IH. The podoplanin stains performed on all specimens were also negative, arguing against a lymphatic origin. Although no specific stain for venous vascular malformations has been identified and validated in the literature to date, the histological appearance and immunohistochemical profile outlined above validate the conclusion that all of the lesions evaluated in this study are most accurately classified as venous malformations, and that they are inconsistent with vascular tumors (hemangiomas) and lymphatic or arteriovenous malformations. A variant temporal bone lesion known as “osseous” or “ossifying” hemangioma is also reported in the literature. Such lesions have a characteristic honeycomb or sunburst radiographic appearance attributed to intralesional lamellar bony trabeculae documented on histopathological examination.15 Osseous features were not present in any of our specimens, so it is unknown whether such ossifying hemangiomas would likewise demonstrate the vascular morphology and immmunohistochemical staining pattern seen in the specimens we have evaluated. Correct nosological classification of geniculate vascular lesions will provide clinicians and researchers with a better understanding of the biological origin and pathophysiology of these lesions and help direct rational therapies. Multiple factors have been identified that play a role in abnormal vessel growth and development, including vascular endothelial growth receptors, tumor growth factor-b1, and Tie2 signaling pathways, and the Notch3 gene, which mediates interactions between endothelial and smooth muscle cells.12,16 Clinical trials are underway to investigate the effects of agents that modify angiogenesis pathways (www. cancer.gov/clinicaltrials). Some of these agents may be effective in treating mitotically active vascular tumors (hemangiomas), but not in treating long-standing vascular developmental anomalies (malformations) that cause symptoms in adolescence and adulthood owing to slow vascular expansion in response to hemodynamic or other factors. Distinguishing between hemangiomas and vascular malformations on a histological and molecular level is critical both in guiding therapy for individual patients and in designing therapeutic trials. Correctly identifying facial nerve lesions as venous malformations rather than hemangiomas has important clinical management implications. Patients with lesions of this type can be correctly counseled about the origin and expected natural history of their disorder. This is particularly important in this era when information from nonphysician sources such as the Internet plays such a large role in patient education. Patients carrying a diagnosis of facial nerve hemangioma may come to have false expectations about the possible spontaneous involution of their lesion, or they may be confused by the onset of symptoms relatively late in life. Providing patients with an accurate diagnosis allows them to pursue factual information relevant to their disease entity. An accurate diagnosis also facilitates communication among clinicians. It is interesting to note that two previous
113
reports did suggest that these IAC and geniculate lesions are actually vascular malformations.15,17 The traditional nomenclature referring to the lesions as “hemangiomas” has, however, persisted. More recent literature reclassifying hemangiomas as vascular malformations at various anatomical sites demonstrates increasing effort among all disciplines to correctly identify and classify these lesions.18,19 There are two main limitations to the present study. One is the small sample size. Given the rarity of this clinical entity, we could identify only seven cases of facial nerve or geniculate hemangioma requiring surgical intervention and, hence, providing pathological specimens over the past 18 years. There were additional patients with a suspected vascular lesion of the geniculate ganglion who did not warrant or refused surgery. Given the relatively long time interval employed and the incorporation of data from two tertiary referral institutions with a busy joint neuro-otological service, we think it is unlikely that a significantly larger series of patients could be obtained. The second potential limitation of this study is that the immunohistochemical evidence supporting the diagnosis of venous malformation is based on negative results—in other words, on absence of staining. It could be argued that the negative results are due to inadequate processing or staining techniques. However, our evaluation ruled out this possibility by requiring that the expected positive staining of appropriate internal positive controls for all immunoreactions be strongly present. No specific immunohistochemical stain for venous vessels, at least in routinely fixed, paraffinembedded tissue, is available. The characteristic morphology of the lesional vascular channels, coupled with demonstration of lack of expression of the lymphatic marker podoplanin, is sufficient for the diagnosis of venous malformation in this location, just as it is in the more common cutaneous and subcutaneous lesions.13 In summary, lesions of the facial nerve and geniculate ganglion historically referred to as “hemangiomas” do not share any of the clinical, histopathological, or immunohistochemical features characteristic of IHs or other vascular tumors. Such lesions should be more appropriately classified as venous vascular malformations of the facial nerve consistent with current terminology. This nomenclature carries prognostic and therapeutic implications for patients as well as for clinicians in the fields of both otolaryngology and pathology. All research was supported by departmental funding.
Author Information From the Department of Otology and Laryngology, Harvard Medical School (Drs Benoit, McKenna, and Cunningham), Boston, MA; the Department of Otolaryngology, Massachusetts Eye and Ear Infirmary (Drs Benoit, McKenna, and Cunningham), Boston, MA; the Department of Pathology, Medical College of Wisconsin (Dr North), Milwaukee, WI; and the Department of Pathology, Massachusetts General Hospital (Drs Mihm and Johnson), Boston, MA.
114
Otolaryngology–Head and Neck Surgery, Vol 142, No 1, January 2010
Corresponding author: Margo McKenna Benoit, Massachusetts Eye and Ear Infirmary, Department of Otolaryngology, 243 Charles St, Boston, MA 02114. E-mail address: [email protected]. This article was accepted for presentation at the 2009 AAO-HNSF Annual Meeting & OTO EXPO, San Diego, CA, October 4-7, 2009.
Author Contributions Margo McKenna Benoit, conception and design, acquisition of data, analysis and interpretation, drafting of manuscript, final approval; Michael J. McKenna, conception and design, acquisition of data, analysis and interpretation, drafting of manuscript, final approval; Martin C. Mihm, conception and design, acquisition of data, analysis and interpretation, drafting of manuscript, final approval; Paula E. North, acquisition of data, analysis and interpretation, drafting of manuscript, final approval; Matthew M. Johnson, acquisition of data, analysis and interpretation, drafting of manuscript, final approval; Michael J. Cunningham, conception and design, analysis and interpretation, drafting of manuscript, final approval.
Disclosures Competing interests: None. Sponsorships: None.
References 1. Friedman O, Neff BA, Willcox TO, et al. Temporal bone hemangiomas involving the facial nerve. Otol Neurotol 2002;23:760 – 6. 2. Mulliken JB, Glowacki J. Hemangiomas and vascular malformations in infants and children: a classification based on endothelial characteristics. Plast Reconstr Surg 1982;69:412–22. 3. Chang MW. Updated classification of hemangiomas and other vascular anomalies. Lymph Res Biol 2003;1:259 – 65. 4. Enjolras O, Wassef M, Chapot R, editors. Color atlas of vascular tumors and vascular malformations. Cambridge: Cambridge University Press; 2007.
5. North PE, Waner M, Buckmiller L, et al. Vascular tumors of infancy and childhood: beyond capillary hemangioma. Cardiovasc Pathol 2006;15:303–17. 6. North PE, Waner M, Mizeracki A, et al. A unique microvascular phenotype shared by juvenile hemangiomas and human placenta. Arch Dermatol 2001;137:559 –70. 7. North PE, Milton W, Mizeracki A, et al. GLUT1: a newly discovered immunohistochemical marker for juvenile hemangiomas. Hum Pathol 2000;31:11–22. 8. Jackler RK, Brackmann DE, editors. Neurotology. Saint Louis: Mosby; 1994. p. 1323– 4. 9. Omojola MF, Al Hawashim NS, Al Zuwayed M, et al. CT and MRI features of cavernous haemangioma of the internal auditory canal. Br J Radiol 1997;70:1184 –7. 10. Baker LL, Dillon WP, Hieshima GB, et al. Hemangiomas and vascular malformations of the head and neck: MR characterization. Am J Neuroradiol 1993;14:307–14. 11. Dubois J, Garel L. Imaging and therapeutic approach of hemangiomas and vascular malformations in the pediatric age group. Pediatr Radiol 1999;29:879 –93. 12. Tille JC, Pepper MS. Hereditary vascular anomalies: new insights into their pathogenesis. Arterioscl Thromb Vasc Biol 2004;24:1578 –90. 13. North PE, Mihm MC. Histopathological diagnosis of infantile hemangiomas and vascular malformations. Facial Plast Surg Clin North Am 2001;9:505–24. 14. Kahn HJ, Bailey D, Marks A. Monoclonal antibody D2-40, a new marker of lymphatic endothelium, reacts with Kaposi’s sarcoma and a subset of angiosarcomas. Mod Pathol 2002;15:434 – 40. 15. Eby TL, Fisch U, Makek MS. Facial nerve management in temporal bone hemangiomas. Am J Otol 1992;13:223–32. 16. Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000;407:249 –57. 17. Mangham CA, Carberry JN, Brackmann DE. Management of intratemporal vascular tumors. Laryngoscope 1981;91:867–76. 18. Yeo A, Majithia A, Kalan A. Haemangioma or vascular malformation of the tympanic membrane? A case report and review of the literature. Indian J Otolaryngol Head Neck Surg 2008;60:59 – 61. 19. Greene AK, Rogers GF, Mulliken JB. Intraosseus “hemangiomas” are malformations and not tumors. Plast Reconstr Surg 2007;119: 1949 –50.