HANDOUT Clinical, histologic, and immunohistochemical features in melanomas of childhood and adolescence.

Clinical, histologic, and immunohistochemical features in melanomas of childhood and
SPP and ASDP Companion Meeting. Baltimore. USCAP 2013
Victor G. Prieto, MD, PhD.
Departments of Pathology and Dermatology. The University of Texas M.D. Anderson Cancer
Center, Houston, TX (USA).
This companion meeting will be focused on neoplastic lesions in children and adolescents.
On the field of melanocytic lesions, this first lecture will focus on the clinical, histological,
and immunohistochemical findings of melanocytic lesions, and it will be followed by two
lectures describing the most recent developments in the fields of molecular pathology (CGH,
FISH, and proteomics).
Cutaneous melanoma in childhood is rare, particularly before puberty. The incidence under
15 years of age is approximately 1 per million (7% of cancers among patients aged 15-19
years vs. 1.2% in patients younger than 15 years of age).
As with adult population, childhood melanoma mainly affects Caucasian individuals; some
series have indicated that white race is associated with an impaired prognosis, although that
is not our experience. Some studies have suggested that there is a slight female predominance
and that male sex is associated with worse prognosis. However our experience, of the data
from cancer registry show equal sex distribution and prognosis. The extremities and the trunk
are the most common anatomic locations for cutaneous melanomas in children. On the other
hand it has been suggested that childhood melanomas arising from the head and neck may
have worse prognosis.
Like in the adult population, the most likely reason to suspect melanoma is the detection of
changes in the appearance of a “mole”. Changes described in decreasing order of frequency
are: rapid size increase, bleeding, color change, itch, lymph node enlargement, subcutaneous
mass, pain and distant metastases. Another common presentation is that of an amelanotic
lesion. Approximately 10% of melanomas develop in association with a pre-existing nevus
are asymptomatic. Since a diagnosis of melanoma is only rarely included in the differential
diagnosis of cutaneous lesions in children, some studies have reported that up to 50-60% of
cases have had a significant delay in the diagnosis.
Only a minority of childhood melanoma in children may also occur in mucosal surfaces.
Primary melanomas of mucosal sites (oral and nasal cavity, genitourinary, gastrointestinal
and conjunctiva) account for only 3-4% of all melanomas.
Many of known risk factors associated with melanoma in adults (e.g., large congenital nevi,
dysplastic nevus syndrome, numerous nevi) may be also identified during childhood. Such
conditions were demonstrated in about 50% of cases. As in adults, a large number of
melanocytic nevi in children are associated with increased risk of melanoma and a number of
childhood melanomas arise in either congenital or acquired nevi. There is some controversy
about the risk for progression of congenital nevi to melanoma. Historical estimates of
melanoma risk were likely exaggerated due to diagnosis of proliferative nodules as
developing in congenital nevi as melanoma. Regarding malignant transformation of
congenital nevi, approximately 10% of childhood melanoma arises from congenital nevi,
including 3.5% from giant nevi. In our experience, approximately half of all childhood
melanomas arise in children with sporadic dysplastic nevi and around 10% in children with
dysplastic nevus syndrome. In general, familial cases account for approximately 5-10% of
melanomas. This antecedent greatly increases risk of melanoma, both for patients with
dysplastic and benign nevi. Familial melanoma is associated with a lower age at diagnosis
than sporadic melanoma, presence of dysplastic nevi and the development of multiple
primary melanomas. The pattern of heredity is consistent with an autosomal dominant
inheritance with incomplete penetrance. Inactivating mutations of the CDKN2A gene (for
p16 and p14ARF) are present in 20-40% of such families and in 15% of individuals with
multiple primary melanomas but only rarely (less than 5%) of childhood melanomas. IN our
series 15 (22.4%) patients had 3 or more first-degree relatives with melanoma, but none of
them had multiple melanomas during the follow-up.
Another phenotype associated with melanoma is the presence of large number of melanocytic
nevi. In a study of 201 adolescents with melanoma, they reported a 34-fold increased risk in
patients with more than 100 nevi and 15-fold in those with 10 or more large nevi. In our
series we found such association only in 21% of patients.
A precursor nevus has been reported in up to 47% of childhood melanomas although it
appears that the percentage is much lower (15%). Small congenital nevi occur in 1% of newborns, and rarely develop melanoma before the second decade of life. In our large series, we
only found small nevi in 2/137 patients (3%). On the other hand, it is interesting that we have
shown that children in whom the melanoma is associated with a benign melanocytic nevus
have a significantly lower risk of metastases.
Other risk factors include inherited and acquired immunodeficiency. In pediatric population,
melanoma account for up to 15% of all post-transplantation skin cancers. Regarding sun
exposure, unless the patient suffers from inherited conditions that increase sensitivity to
ultraviolet light, such as xeroderma pigmentosum, it is unlikely that sun-exposure will be
associated with childhood melanoma.
Survivors of childhood cancer have increased risk for a second neoplasm, including
melanoma. Such lesions tend to be deeper and have poor prognosis, possibly associated with
tumor-related immunodeficiency mechanisms mediated by the neoplasm itself.
The most important finding regarding prognosis is age. In most studies, older age, likely after
puberty, is associated with worse survival. In several series, the large majority patients who
died from melanoma were older than 10 years at time of diagnosis. For example, Sander et al
reviewed 126 cases of melanoma in young patients, and found that all 13 deceased patients
were older than 12 years of age. It is also our experience, since in our series all our patients
who died from melanoma were older than 10 years at time of diagnosis. Thus it seems that
postpuberal age is one of the main factors determining death risk.
In general, it seems that the same histopathologic criteria developed for adult melanoma
should be used for diagnosis of childhood melanoma: size >10 mm, asymmetry, lateral
borders poorly demarcated, intraepidermal pagetoid spread of single cells prominent in the
center or else present at the periphery of the lesion, irregular intraepidermal nest formation,
“consumption”/atrophy or ulceration of the epidermis, absence of maturation, deep extension
in dermis/subcutaneous tissue, expansive deep border, expansive / diffuse sheets of cohesive
cells in the dermis, cellular atypia, and deeply located or atypical mitotic figures. However,
these features should be evaluated together since none of them is sufficient for a diagnosis of
Regarding histologic subtypes, several series have reported a predominance of nodular
lesions, but more recent series indicate that most childhood melanomas are of superficial
spreading type. It is possible that some of the cases in prior series were actually superficial
spreading melanomas with large, dermal, nodular component. It is interesting that some
studies have indicated that true nodular morphology is associated with impaired prognosis.
Interestingly, in contrast with melanomas in adults, in whom histologic subtype is not related
to prognosis, nodular melanomas appear to show poor survival in children. There appear to
be higher rate of metastases in melanomas with fusiform or spitzoid cytology. Also, some
studies have indicated impaired prognosis of lesions with melanomas of small cell type,
similar to small round cell tumors.
A high percentage of reported childhood and adolescent melanomas are relatively thick,
possibly due to delay in suspicion of malignancy. When comparing adult and pediatric
melanomas, the latter show significantly thicker tumors and such thicker lesions appear to
have poor prognosis. Nodal metastases have been detected in up to 2/3 of children with Clark
level IV-V or Breslow >1.5 mm melanomas. Conversely, metastases or recurrence are
unusual in patients with melanoma <1.5 mm thick. As in adults, ulceration, VGP, vascular
invasion and high mitotic activity are also associated with higher recurrence rate, although in
our experience it did not correlate with decreased survival. Brisk lymphocytic infiltrate and
regression do not appear to have significant correlation with survival.
The most challenging diagnosis is the subtype spitzoid melanoma. There appears to be a
spectrum of lesions from obvious benign (i.e., Spitz nevus) to obvious malignant (spitzoid
melanoma), with a group of lesions falling in between the two poles of the spectrum. It has
been reported that up to 40% of melanomas originally diagnosed as nevi actually correspond
to these spitzoid melanomas.
Classic Spitz nevus is considered as a benign lesion with some nuclear and cytological
pleomorphism but absence of deep/atypical mitotic figures, expansile pattern of growth, etc.
(see above). In contrast, spitzoid melanoma is a malignant melanocytic lesion with some
cytological attributes of Spitz nevus (spindle and/or epithelioid melanocytic cells) but with a
preponderance of the atypical features summarized above. Patients with lesions sharing some
classic features of Spitz nevus and some of these abnormal pathology features, in which a
definite diagnosis of nevus or melanoma is not firmly established, may be described as
atypical spitzoid lesions or atypical Spitz tumors (of unknown biological potential).
The term “nevoid melanoma” applies to lesions in which there is little proliferation of
melanocytes in the epidermis and a dermal proliferation that mimics some features of a
compound or intradermal nevus. Due to the lack of prominent intraepidermal component
some of these lesions are diagnosed as nevi. The main histological features to look for in
these lesions include: dermal mitotic figures, sheet-like growth pattern, nuclear
pleomorphism and lack of conventional maturation (monomorphous appearance of the
melanocytes throughout the lesion.
Although the diagnosis of melanoma continues to rest on microscopic morphologic features,
immunohistochemistry may be helpful in a number of cases, especially in like spitzoid or
nevoid melanoma. HMB-45 usually shows a stratified pattern with diminished expression
toward the base of the lesion (maturation) in most nevi, including Spitz nevi, in contrast with
the patchy pattern of expression throughout melanomas. Blue nevi, deep penetrating nevi,
plexiform nevi, and a minority of Spitz nevi show diffuse labeling with HMB45. Spitz nevi
also have low proliferation rates (1-2%) when assessed with Ki-67, compared with much
higher rates in melanomas (15-30%). Moreover, melanoma has labelling throughout the
lesion while nevi have more staining at the top of the lesion than at the bottom (maturation).
S100 protein and MART-1 (melanoma antigen recognized by T cells-1) show diffuse
expression throughout both Spitz nevus and melanoma; however, desmoplastic melanomas
may lose MART1 expression while preserving S100; this finding may be diagnostically
helpful. Any spindle cell melanocytic lesion that shows either patchy or negative expression
of MART1 should be carefully evaluated to rule out melanoma.
Other markers have been studied in spitzoid and other childhood melanocytic lesions. p16
and its related protein CDK4 show a mostly inverse pattern of expression; many Spitz nevi
express p16 but not CDK4 and the opposite is true for melanoma. However, more recent
studies question the usefulness of immunohistochemical analysis of p16 to distinguish Spitz
nevus from spitzoid melanoma.
Expression of CD133, a stem-cell marker, has been reported to be associated with aggressive
behavior (lymph node and visceral metastasis) in childhood melanoma.
During the period 1968-2004, there were 643 deaths attributed to melanoma among children
aged 0-19 years in US (2.25/year/million at-risk individuals). We have shown that presence
of metastases at the time of diagnosis is one of the two main factors that influence overall
survival in children with melanoma. Along these lines, it has been reported that pediatric
patients have a survival at 10 years of with 94.5% in stage I-II and 60.1% in stage III.
Regarding sentinel lymph node (SLN), recent studies indicate that this procedure has a less
significant predictive value in childhood melanomas than in adult lesions. While
approximately 20% of melanomas in adults are associated with positive SLN, pediatric
lesions show values between 25 and 60. This seems to indicate that even with positive SLN,
those children appear to have a better prognosis than adults. Although it is unclear why, it
have been suggested possible differences in lymphatic flow density and immune system,
which would be able to clear micrometastases in children. Another possibility is that some of
these positive SLN actually contain benign melanocytes (nodal nevi). These nodal nevi are
usually small clusters of benign-appearing melanocytes within the capsule (and rarely in the
parenchyma) of lymph nodes, in contrast with the common subcapsular location of metastatic
melanoma. Overall, up to 20% of lymphadenectomies from the axilla or groin contain such
benign collections of melanocytes. These benign nodal melanocytes usually lack gp100
expression (with HMB45), and show very low Ki67 expression, thus consistent with benign
melanocytes. In order to facilitate the identification of proliferating melanocytes, it is
possible to use a cocktail that includes anti-MART1 and MIB1 (against Ki67). Since these
two markers are expressed in different cellular components (Ki67 in the nucleus and MART1
in the cytoplasm) it is relatively easy to determine how many of the melanocytes (i.e., cells
expressing MART1) are proliferating (i.e., expressing Ki67). In view of the apparent
differences between prognostic information derived from SLN examination in children and
adults, only long-term follow-up studies will be able to determine the actual value of SLN in
melanocytic lesions from children.
In conclusion, careful analysis of histological features as well as the additional
immunohistochemistry should allow establishing the right diagnosis in most cases of
childhood melanoma. Although it seems that pediatric melanoma patients still a number of
children will develop metastasis and die of their disease, particularly when melanoma is
diagnosed after puberty.
 Al Dhaybi R, Sartelet H, Powell J, Kokta V. Expression of CD133+ cancer stem cells in childhood
malignant melanoma and its correlation with metastasis. Mod Pathol. Mar 2010;23(3):376-380.
 Aldrink JH, Selim MA, Diesen DL, et al. Pediatric melanoma: a single-institution experience of 150
patients. J Pediatr Surg. Aug 2009;44(8):1514-1521.
 Conti EM, Cercato MC, Gatta G, Ramazzotti V, Roscioni S; EUROCARE Working Group.
Childhood melanoma in Europe since 1978: a population-based survival study. Eur J Cancer
 de Vries E, Steliarova-Foucher E, Spatz A, et al. Skin cancer incidence and survival in European
children and adolescents (1978–1997): report from the Automated Childhood Cancer Information
System project. Eur J Cancer. 2006; 42: 2170–2182
 Ducharme EE, Silverberg NB. Pediatric malignant melanoma: an update on epidemiology,
detection, and prevention. Cutis. Oct 2009;84(4):192-198.
 Ellis ZM, Jassim AD, Wick MR. Anorectal melanoma in childhood and adolescence. Ann Diagn
Pathol. Apr 2010;14(2):69-73.
 Hilliard NJ, Krahl D, Sellheyer K. p16 expression differentiates between desmoplastic Spitz nevus
and desmoplastic melanoma. J Cutan Pathol. Jul 2009;36(7):753-759.
 Howman-Giles R, Shaw HM, Scolyer RA, et al. Sentinel lymph node biopsy in pediatric and
adolescent cutaneous melanoma patients. Ann Surg Oncol. Jan 2010;17(1):138-143.
 Jen M, Murphy M, Grant-Kels JM. Childhood melanoma. Clin Dermatol. Nov-Dec
 Juhl AL, Byers TE, Robinson WA, Morelli JG, Crane LA. The anatomic distribution of melanoma
and relationships with childhood nevus distribution in Colorado. Melanoma Res. Aug
 Kirkwood JM, Jukic DM, Averbook BJ, Sender LS. Melanoma in pediatric, adolescent, and young
adult patients. Semin Oncol. Oct 2009;36(5):419-431.
 Mason A, Wititsuwannakul J, Klump VR, Lott J, Lazova R. Expression of p16 alone does not
differentiate between Spitz nevi and Spitzoid melanoma. J Cutan Pathol. Dec 2012;39(12):10621074.
 McNutt NS, Urmacher C, Hakimian J, Hoss DM, Lugo J. Nevoid malignant melanoma:
morphologic patterns and immunohistochemical reactivity. J Cutan Pathol. 1995;22(6):502–517
 Mills O, Messina JL. Pediatric melanoma: a review. Cancer Control. Jul 2009;16(3):225-233.
 Moore-Olufemi S, Herzog C, Warneke C, et al. Outcomes in pediatric melanoma: comparing
prepubertal to adolescent pediatric patients. Ann Surg. Jun 2011;253(6):1211-1215.
 Murali R, Sharma RN, Thompson JF, et al. Sentinel lymph node biopsy in histologically ambiguous
melanocytic tumors with spitzoid features (so-called atypical spitzoid tumors). Ann Surg Oncol.
2008; 15: 302–309
 Pappo AS. Melanoma in children and adolescents. Eur J Cancer 2003;39(18):2651-61.
 Paradela S, Fonseca E, Pita S, Kantrow SM, Goncharuk VN, Diwan H, Prieto VG. Spitzoid
melanoma in children: clinicopathological study and application of immunohistochemistry as an
adjunct diagnostic tool. J Cutan Pathol 2009; 36: 740-52
 Paradela S, Fonseca E, Pita-Fernández S, Kantrow SM, Diwan AH, Herzog C, Prieto VG.
Prognostic factors for melanoma in children and adolescents: a clinicopathologic, single-center
study of 137 Patients. Cancer 2010; 116: 4334-44
 Paradela S, Fonseca E, Prieto VG. Melanoma in children. Arch Pathol Lab Med 2011; 135: 307-16
 Paradela S, Fonseca E, Pita-Fernández S, Prieto VG. Spitzoid and non-spitzoid melanoma in
children. A prognostic comparative study. J Eur Acad Dermatol Venereol. e-Pub 8/2012
 Prieto VG, Shea CR. Immunohistochemistry of melanocytic proliferations. Arch Pathol Lab Med
2011; 135: 853-9
 Somers GR, Chilton-MacNeill S, Zheng K, et al. Increased detection of metastatic melanoma in
pediatric sentinel lymph node biopsies using RT-PCR on paraffin-embedded tissue. Appl
Immunohistochem Mol Morphol. Jul 2010;18(4):365-370.
 Strouse JJ, Fears TR, Tucker MA, Wayne AS. Pediatric melanoma: risk factor and survival analysis
of the surveillance, epidemiology and end results database. J Clin Oncol 2005 20;23(21):4735-41.
 Tcheung WJ, Marcello JE, Puri PK, Abernethy AP, Nelson KC. Evaluation of 39 cases of pediatric
cutaneous head and neck melanoma. J Am Acad Dermatol. Aug 2011;65(2):e37-42.
 Toro J, Ranieri JM, Havlik RJ, et al. Sentinel lymph node biopsy in children and adolescents with
malignant melanoma. J Pediatr Surg. 2003; 38: 1063–1065
 Wang YL, Uhara H, Yamazaki Y, Nikaido T, Saida T. Immunohistochemical detection of CDK4
and p16INK4 proteins in cutaneous malignant melanoma. Br J Dermatol. Feb 1996;134(2):269275.