British Journal ofPlastic Surgery (1990).

43,608-611
0 1990 The Trustees of British Association of Plastic Surgeons

Malignant melanoma of the external ear

D.A. HUDSON, J. E. J. KRIGE, R. M. STROVER and H. S. KING
Departments of Plastic Surgery, General Surgery and Radiotherapy, University of Cape Town, and the
Melanoma Clinic, Groote Schuur Hospital, Cape Town, South Africa

Summary-Twelve patients with primary melanoma of the external ear were treated during a lo-year
period. Ten patients were males. Median melanoma depth was 2.1 mm and seven patients had
lesions deeper than 2 mm. Eight of 12 lesions were either Clark level IV or V. Six patients had wedge
excisions of the primary lesion and all developed recurrent disease (local: three). Six patients
underwent wide excision and none developed local recurrence. All patients with primary disease less
than 2 mm are alive. Only two of seven patients with lesions deeper than 2 mm have survived. The
definitive excision for melanoma of the ear should encompass wide margins despite the cosmetic
consequences. Lesser procedures inevitably lead to local recurrence. Melanoma deeper than 2 mm
penetration on the ear in our series had a poor prognosis.

Malignant melanoma of the external ear accounts Material and methods

for less than 1% of all cutaneous melanoma and
The records of all patients registered at the
perhaps because of infrequent occurrence, manage-
Melanoma Clinic at Groote Schuur Hospital, Cape
ment decisions are often dictated by limited
Town, between 1977 and 1986 were reviewed.
personal experience and anecdotal impressions.
During this period 1076 patients with primary
Earlier reports have suggested that melanoma of
cutaneous melanoma were treated, of whom 12
the ear has a poor prognosis (Batsakis, 1980).
(1 .l%) had melanoma of the external ear. Clinical
Detailed analyses of the natural history of mela-
data, pathological material and complete follow-
noma have provided a scientific rationale for the
up, performed by members of the Melanoma Clinic,
tailoring of treatment based on microstaging and
were available on all patients. Clinical data included
histological characteristics of the primary tumour
patient age, sex, symptoms, duration of tumour,
(Briele et al., 1985). The major histopathological
clinical stage, surgical treatment (width of primary
criteria shown to be of prognostic value are the
excision and details of lymph node dissection) and
histological subtype, tumour thickness, level of
subsequent clinical course, including sites and
invasion and the presence of ulceration (Wanebo et
timing of recurrence, management of recurrence
al., 1988). In addition, the specific anatomical
and survival. The histology of all primary lesions
location of the primary lesion has been proposed as
was reviewed and the recorded details included
an independent predictor of patient survival. The
histological subtype of melanoma (lentigo maligna
ear and scalp are considered high risk areas within
(LMM), superficial spreading (SSM) and nodular
the head and neck region (Wanebo et al., 1988).
(NM) level and thickness.
In addition to the thin skin and variable
lymphatic drainage, anatomical and cosmetic con-
straints may restrict traditional wide surgical
Results
excision margins on the ear. Recent data supporting
conservative local excision with narrower resection The average age at presentation of the 12 patients
margins (Day et al., 1982) may have important in the study was 61.6 years with a range of 40-78
sequelae in the management and prognosis of years. Ten patients were male and two were female.
lesions of the external ear. This study was under- The three major histogenetic varieties of melanoma
taken to evaluate the prognosis in 12 patients with were equally distributed and four patients each had
melanoma of the ear treated at a major referral either lentigo maligna melanoma, superficial
centre during a lo-year period, and to analyse the spreading or nodular melanoma. Eight patients
consequences of a limited resection margin. presented with either Clark level IV or V lesions

608
MALIGNANT MELANOMA OF THE EXTERNAL EAR 609

and 7 patients had primary lesions measuring one-tailed Fisher’s exact test). Two of the 12
deeper than 2 mm. The median depth of the 12 patients developed positive regional lymph nodes
primary lesions was 2.1 mm (range 0.76-5 mm). An during follow-up and both had therapeutic node
equal number of patients by chance had either dissections, one of whom is alive at 89 months. Two
wedge excision (clearance margin < 10 mm) or of the three patients who developed local recurrence
wide local excision (margin > 10 mm). Six patients on the ear died. The mean follow-up period in
underwent wedge resection alone of the primary survivors is 60.3 months (range 21-133 months)
lesion (mean thickness 2.7 mm, range 0.75-5 mm) (Table 1). Five-year actuarial survival is 49’“.
and were histologically clear of tumour at the
resection line. All six patients with wedge excisions
Discussion
(mean clear margin 6 mm, range 5-8 mm) devel-
oped recurrent disease (local three, nodal two, Malignant melanoma of the external ear represents
systemic one) at a median of 15 months (range 2- 3312% of all melanomas originating in the head
24 months) and underwent subsequent therapeutic and neck region (Urist ez al., 1984; Cox er al., 1987)
regional node dissection (two patients) and total and is primarily a disease of men, usually in their
excision of the ear (three patients). Two of these six seventh decade, and often presents with advanced
patients are alive at 93 and 133 months. Six patients disease. Eighty-three per cent of patients with
had either wide local excision (five) or total excision primary melanoma of the ear in this study were
of the ear (one) as the initial definitive treatment of males compared to 69”< in the Scottish Melanoma
the primary lesion (mean thickness 2.00 mm, range Group series (Cox et al., 1987). This difference in
0.76-3.6 mm). The mean clear margin for lesions incidence by sex may reflect differing exposures of
excised by wide local excision was 12 mm (range the head and neck region in men and women to
IO- 16 mm). No patients have had local recurrence; sunlight, attributable in part to differences in hair
one patient developed systemic disease with lung coverage and outdoor occupation (Briele rl (II.,
metastases at 35 months and died at 45 months. All 1985). Fifty-eight per cent of lesions in this study
patients with primary lesions less than 2 mm were deeper than 2 mm and 68% were level IV or
penetration are alive without recurrent disease. V, with a median depth of penetration of 2.1 mm.
Only two of seven patients with lesions deeper than In a series of head and neck melanomas, patients
2 mm have survived. The difference in survival with ear lesions had a deeper median penetration
between these two groups is significant (p -0.02; (2.7 mm) and a greater proportion ofthicker lesions
Table 1 Patient data: melanoma of the ear

I 7XM LMM I 0.76 WLE Well 23 months

2 71F LMM IV 3.0 WLF Well 21 months
3 59M LMM IV I .o WLE .-. Well 57 months
4 SIM LMM IV 0.75 Wedge Nodal (15 months) Well 93 months
Neck dissection
5 43F SSM III I 1 WLE Well 15 months
h 63M SSM IV 18 Wedge Local (14 months) Well 133months
amputation
7 7XM SSM IV 3.6 WLE Lung mets (35 months) Died 41 months
X 5lM SSM III 1.3 Wedge Local (23 months) Died 57 months
amputation
Lung mets (34 months)
9 67M NM IV 7.5 Amputation Well 70 months
I0 66M NM IV 4.0 Wedge Lung mets (1 months) Died 4 month5
II 40M NM III 71 Wedge Nodal (2 months) amputation Died 37 month3
Neck d&section
Brain mets (34 months)
I7 72M NM V 50 Wedge Local (18 months)
nmputatlon
Brain mets (19 months)

WLE : wide local excision:

Wedge: wedge excision.
610 BRITISH JOURNAL OF PLASTIC SURGERY

(73% deeper than 1.5 mm) than melanomas on the their 102 patients had the depth of tumour invasion
face, neck or scalp (Wanebo et al., 1988). measured. The authors recommend wedge excision
Wanebo et al. (1988) and Byers et al. (1980) have after excisional biopsy for small lesions on the helix
previously noted the high risk nature of melanoma and either wedge excision or partial amputation for
of the ear compared to melanoma occurring on the larger lesions, but do not define the extent of wedge
face and neck. Multivariate analysis of clinical excision or relate its application to depth of
stage 1 head and neck melanoma showed that invasion. Balch and Milton (1985) describe the
location, thickness, level and ulceration were technique of wedge excision for a melanoma on the
important prognostic indicators (Wanebo et al., outer helix of the ear and recommend extending
1988). Thickness of the melanoma measured in the incision to the external auditory canal, and
millimetres is the single most accurate prognostic include the inner helix and a wedge of cartilage to
parameter (Briele et al., 1985) and analysis of our allow primary closure of the defect
data showed 2 mm to be the watershed level on the Currently no clear consensus exists regarding the
ear. All patients in our study with tumours less than management of the regional lymph nodes for
2 mm are alive, while five of seven patients with patients with primary melanoma of the ear.
lesions deeper than 2 mm have died. A number of Although a number of authors have reported a
reports have suggested that melanoma of the ear is survival benefit after elective regional node dissec-
associated with a poor prognosis, and overall 5-year tion for head and neck melanoma (Ames et al.,
survival rates have ranged from 22% (Fitzpatrick 1976; Urist et al., 1984) others found no survival
et al., 1972) to 31% (Wanebo et al., 1988). Byers et benefit (Ballantyne, 1970). Only two prospectively
al. (1980) report better results and their data suggest randomised trials for elective regional lymph node
that thickness of the lesion is a more sensitive dissection have been reported and neither of these
indicator of disease-free survival than the level of involved patients with head and neck melanoma
invasion. Of their 18 patients with lesions less than (Veronesi et al., 1982; Sim et al., 1986). Historically
1 mm thick, 65% survived 5 years while 45% of 13 those who oppose prophylactic lymph node dissec-
patients with lesions greater than 3 mm thick tion in patients with stage 1 melanoma cite the fact
survived for 5 years. Urist et al. (1984) found an that the great majority of patients will undergo
SO%, 5-year survival for SSM and NM of the ear needless surgical treatment. Of 33 patients with
and 100% in six patients with LMM. primary melanoma of the ear undergoing prophy-
Although the basic principle of wide excision of lactic block dissection, only four (12%) had micro-
the primary site including surrounding normal skin scopic positive nodes (Byers et al., 1980). In
is generally accepted, the traditional surgical addition, the removal of subclinical node metastases
margins may be more difficult to achieve on the ear is of questionable benefit. In patients with primary
in an attempt to preserve function and appearance. melanoma of the ear, the prognosis was no better
All six patients in this series who had a wedge when elective lymph node dissection was performed
excision alone developed recurrent disease, of and positive clinical metastases found, than in
which half were local recurrence. In the six patients patients who underwent subsequent neck dissection
who had either wide or total excision of the ear, no after nodes became clinically palpable (Byers et al.,
local recurrence developed. The mean thickness of 1980). Only two of 12 patients in our series with ear
the wedge excision group, however, was thicker melanoma developed positive regional nodes dur-
(2.7 mm) than the wide excision group (2 mm). One ing prolonged follow-up, and one is alive after
hundred and two cases of external ear melanoma therapeutic node dissection. The majority of pa-
treated over a 34-year period at the M. D. Anderson tients with lesions deeper than 2 mm developed
Hospital were reviewed by Byers et al. (1980). systemic metastases without evidence of regional
Seventy-six per cent of patients had the lesion lymph node involvement, offsetting any advantage
removed either as a biopsy or as definitive treatment prophylactic neck dissection might offer.
before referral, 13% had only an incisional biopsy Our findings support the premise that melanoma
while 11% had had prior re-excision. Thirty-six arising on the external ear, especiaIly those deeper
patients subsequently had a wedge excision and 30 than 2 mm penetration, have an aggressive biolog-
a partial or total amputation of the ear. Local ical behaviour. The optimal treatment for primary
recurrence occurred in 8%. Thick lesions (> 3 mm) melanoma of the ear should be based on adequate
and level IV invasion were associated with a higher removal of microscopic disease. Wedge excision
incidence of positive nodes. However, only 51 of per se is insufficient therapy and should only be
MALIGNANT MELANOMA OF THE EXTERNAL EAR 611

used as a diagnostic excisional biopsy. If a mela- Fitzpatrick? P. J., Brown, T. C. and Reid, J. (1972). Malignant
melanoma of the head and neck: a clinicopathological :;tudy.
noma is confirmed. the definitive procedure should
Canadian JournalofSurgery, 15,90.
be a wide local excision encompassing margins of Sim, F. H., Taylor, W. F., Pritchard, D. J. and Soule, E. H.
at least IO mm. Lesser resection margins in the ear, (1986). Lymphadenectomy in management of stage 1 malig-
in our series, inevitably led to either local or regional nant melanoma: a prospective randomized study. Ma~~o C/ink
Proceedings, 61,697.
recurrence. Usually less than total amputation of
Uris& M. M., Balch, C. M., Soong, S-J. Milton, G. W., Shaw. H.
the ear is sufficient and preferred. especially if a M., McGovern, V. J., Murad, T. M., McCarthy, W. H. and
portion of the ear is required to support spectacles. Maddox, W. A. (1984). Head and neck melanoma in S34
clinical stage 1 patients. AnnaL~of Surgery. 200,769.
Veronesi, U., Adamus, J., Bandiera, D. C. et al. (1982). Delayed
References regional lymph node dissection in stage I melanoma of the
skin of the lower extremities. Cancer, 49,242O.
Ames, F. C., Sugarbaker. E. V. and Ballantyne, A. J. (1976). Wanebo, H. J., Cooper, P. H., Young, D. V., Harpole, D. H. and
Analysis of survival and disease control in stage I melanoma Kaiser, D. L. (1988). Prognostic factors in head and neck
of the head and neck. American JournalojSurgery, 132,484. melanoma. Effect of lesion location. Cwcrr. 62,831
Balch, C. M. and Milton, G. W. (1985). Cutaneous Melunoma.
Philadelphia: J. B. Lippincott
Ballantyne, A. J. (I 970). Malignant melanoma of the skin of the
head-and neck : an analysisof 405 cases. American Journal of The Authors
Surperr. 120.425.
Batsa& J. G.. (1980). Tumour.s qf the Head and Neck. 2nd D. A. Hudson, FRCS, Registrar, Plastic Surgery
Edition. Baltimore: Williams and Wilkins. J. E. J. Krige, FRCS, FCS(SA), Consultant, General Surgery
Briele, H. A., Walker, M. J. and Das Gupta, T. K. (1985). R. M. Strover, FRCS, Consultant, Plastic Surgery
Melanoma of the head and neck. Clinics in P/a.sic Surgery, 12, Helen S. King, M.Med(Rad), Consultant, Radiotherapy
495.
Byers, R. M.. Smith, J. L., Russell, N. and Rosenberg, V. (1980).
Departments of Plastic Surgery, General Surgery and Radio-
Malignant melanoma of the external ear. American Journal of
therapy. University of Cape Town, and the Melanoma Clinic,
Surgery. 140,518.
Groote Schuur Hospital, Cape Town.
Cox, N. H., Jones, S. K. and MacKie, R. M. (1987). Malignant
melanoma of the head and neck in Scotland. An eight-year
analysis of trends in prevalence, distribution and prognosis. Requests for reprints to: Mr D. Hudson. FRCS. Department of
Quarwriy Journal oj Medicine. 64.66 I Plastic Surgery. Groote Schuur Hospital. Observatory 7925,
Day, C. L. Jr, Mihm, M. C. Jr, Sober, A. J., Fitzpatrick, T. B. Cape Town, South Africa.
and Malt, R. A. (1981). Narrower margins for clinical stage I
malignant melanoma New England Journal qf Medicine. 306, Paper received 15December 1989.
479. Accepted 17 April 1990.