life with neurologic manifestations of increased intracranial pressure, mass lesions, or malignant

transformation to melanoma. Magnetic resonance imaging (MRI) findings of NCM may be present, even

in asymptomatic children. NCM is associated with an increased risk of development of central nervous

system melanoma. Symptomatic NCM is associated with a poor prognosis, even in the absence of

malignancy. Chemotherapy is not effective, but shunt placement to reduce intracranial pressure may be

palliative.

Clinical Diagnosis and Classification

3 Melanoma may have a characteristic appearance (Fig. 107-1). Early detection is key. The ABCD rule,

which outlines the warning signs of melanoma, was developed decades ago.11 A stands for asymmetry –

one-half of the lesion does not match the other half when drawing an imaginary line through the center

of the lesion. B is border irregularity – the edges of the lesion are ragged, notched, or fuzzy. C is color –

the pigmentation is not uniform throughout the lesion. Varying shades of tan, dark brown to black, or

shades of red, white, and/or blue may be present with a mottled appearance. While historically D has

been used for diameter (a width greater than 6 mm, or the size of a pencil eraser), today many use

difference – any change in the lesion. Alternatively, some maintain D as difference and add E for

evolution. The seven-point checklist used in Europe incorporates several of the ABCD features, including

change. Another useful diagnostic aid is the “ugly duckling” sign. A pigmented lesion that is different

from the others should be carefully approached with a high index of suspicion.

The goal of any diagnostic aid is early detection and the ability to differentiate between benign and

malignant, with 100% sensitivity combined with high specificity. The ABCD (or ABCDE) rule, sevenpoint checklist, and ugly duckling sign all provide relatively highly sensitivity but none fulfills all of

these criteria, especially with respect to specificity, to distinguish melanoma from benign lesions.

The ABCD criteria may fail to detect an important subset of nodular melanoma (NM), which often

does not exhibit the ABCDs and is often less than 6 mm in early development. The ABCD rule is also

relatively static without the important critical characteristic of change. Studies have documented early

signs of a change in size, shape, or color and the early symptom of persistent itching. One of the best

tools for diagnosis of early lesions that may fail to exhibit many of the standard melanoma features is

the importance of change, hence the reason the authors and others have used the D in the ABCD rule to

signify “difference.” A difference or change in lesions, especially with respect to size, shape, color, or

persistent itching, warrants evaluation. Furthermore, the nevi on the body should globally share a

common look or family resemblance. If one of the nevi seems different from the rest, it should be

evaluated with a high index of suspicion. In addition, if all of the nevi are atypical or irregular but are

morphologically similar to one another and are not changing, there may be no reason for individual

lesion concern (atypical or dysplastic nevi). However, the presence of atypical nevi represents a risk

factor for melanoma development anywhere on the skin surface with development in clinically normal

skin, normal nevi, and dysplastic nevi with relatively equal frequency. Thus, excision of all atypical nevi

does little for prevention. Education of the patient for early signs and symptoms, particularly change or

difference, is key in this clinical scenario.

It must be emphasized that critical diagnostic decisions concerning skin lesions should still always be

based on clinical assessments and the history, with attention to melanoma risk factors. The optimum

diagnostic aid that combines 100% sensitivity and specificity currently does not exist. The use of all

available means, including the ABCD and seven-point checklists, ugly duckling sign, difference,

photography, dermoscopy, and evolving digital and computer-assisted imaging technologies, enhances

the ability to diagnose melanoma at the earliest stage.11 The result of using these multiple diagnostic

aids is earlier detection of melanoma with higher overall survival rates.

Based on histologic patterns and clinical characteristics, melanomas can be classified into multiple

categories, with the four major subtypes being lentigo maligna melanoma (LMM), superficial spreading

melanoma (SSM), NM, and acral lentiginous melanoma (ALM).12 In general, these subtypes derive their

prognostic biologic behavior and risk of metastasis based on a secondary correlation with Breslow depth

and are thus not independent prognostic factors. Invasive LMM constitutes 10% to 15% of cutaneous

melanomas and typically occurs on the chronically sun-exposed areas of the head and neck, most often

in older individuals. Clinically, the LMM pattern is associated with a significantly higher rate of often

extensive and asymmetrical subclinical growth several centimeters beyond the clinical lesion. Failure to

completely excise the entire lesion with meticulous margin control results in higher recurrence rates

with lethal potential. In addition, amelanotic and desmoplastic melanoma with neurotropism is more

frequent in the lentigo maligna (LM) lesion subtype. LM is not a benign disease as once thought. LM

3148

represents in situ melanoma and, with time, progresses to invasive LMM. LMM often begins as a flat

brown lesion on chronically sun-exposed areas with progression to black and other classic melanoma

features over time.

3149

Figure 107-1. Examples of skin lesions. A: Giant congenital nevus. B: Lentigo maligna (Hutchinson freckle). C: Melanoma arising

in a lentigo maligna (lentigo maligna melanoma). D: Superficial spreading melanoma. E: Nodular melanoma. F: Acral lentiginous

melanoma (ulcerated nodular plantar melanoma with satellite lesions). G: Subungual melanoma. H: Pigmented basal cell

carcinoma.

SSM accounts for approximately 70% of cutaneous melanomas. This pattern is the most common in

melanomas arising in a pre-existing nevus. SSM is typically characterized by variation in color, irregular

borders, and irregular surface and often exhibits the classic melanoma clinical features.

NM occurs in approximately 15% to 30% of patients with cutaneous melanoma and is often the most

aggressive of the four types of melanoma because of later detection and rapid growth. In general, NM

may be more uniform in coloration than the other types, have regular borders with size less than 6 mm

in early lesions, and lack the classic melanoma features initially. A change in a lesion is important for

early detection.

ALM is a distinct clinicopathologic variant of melanoma that most commonly occurs on the hands,

feet, fingers, and toes as well as in subungual locations (Fig. 107-1). ALM represents only 2% to 8% of

melanoma in whites but 35% to 60% in people of color. Later detection frequently occurs because of

location that is not easily or routinely examined and a low index of suspicion. When corrected for

Breslow depth, the overall prognosis for ALM appears similar to the other subtype categories. In

subungual locations, ALM can appear as an irregular, tan-brown streak in the nail that originates from

the base of the nail bed. More than three fourths of subungual melanomas involve the great toe or

thumb, and they can be confused with subungual hematoma.

Other rare melanoma subtypes deserve brief mention. Mucosal, anal, and vulvovaginal melanomas

are associated with a poorer prognosis, possibly related to advanced disease at presentation with late

detection and high vascularity at the lesion site. Desmoplastic melanoma is associated with a higher rate

of neurotropism and lower rate of lymph node metastases.13–15 Small cell or nevoid melanoma often

lacks any of the classic features, is difficult to accurately diagnose without expert dermatopathology

interpretation, and thus may be initially more frequently misdiagnosed. And, amelanotic melanoma,

occurring in 3% to 4% of cases, is often associated with late detection due to lack of pigment and failure

to diagnose early.

3150

Staging and Prognostic Factors

4 A great deal of information is available regarding various factors that correlate with the clinical

outcome of patients with melanoma. Some of these prognostic factors, such as microstaging and nodal

status, are of sufficient independent significance to be incorporated into staging systems with known

survival rates. Other prognostic factors such as tumor ulceration and microscopic versus macroscopic

nodal disease have also been found to be significant variables that influence survival and have been

incorporated into the staging system.

Microstaging

One of the most important prognostic features of cutaneous melanoma is the stage of development of

the primary tumor. The microstaging method that is used routinely was originally described by

Breslow.16 This method classifies the primary tumor according to its thickness in millimeters, as

measured with an ocular micrometer, from the top of the granular layer to the base of the tumor. Many

investigators have documented an inverse correlation between tumor thickness and survival. The largest

series of 17,600 patients reported by Balch et al.17 is summarized in Table 107-1. The ulceration status

and mitotic rate are also independent factors that are used to determine microstaging of the primary

lesion.17,18 Prior to the use of the Breslow microstaging method, melanomas were staged according to

the level of invasion into the histologic layer of skin. This was known as the Clark level of invasion and

comprised five levels (I, in situ lesions; V, subcutaneous involvement). Several studies have confirmed

that Breslow thickness conveys more accurate prognostic information than does the determination of

Clark level.17,19

The presence of regional lymph node metastases is associated with a worsening prognosis. The tumor

burden (microscopic vs. macroscopic disease) of involved lymph nodes has an inverse correlation with

long-term survival.17,19 The 5-year survival rate for patients with involved lymph nodes ranges from

70% to 25%, based primarily on tumor burden and ulceration status of the primary lesion (Table 107-1).

The use of sentinel lymph node biopsy (SLNB) has identified a subgroup of patients with

micrometastatic nodal disease who have a favorable prognosis compared with patients with macroscopic

nodal involvement. This information has been incorporated in the staging system as well.

RESULTS

Table 107-1 Survival Rates for Melanoma TNM and Staging Categories

Clinical and Pathologic Staging

The American Joint Committee on Cancer (AJCC) developed a five-stage system that divides

melanomas according to tumor thickness (T), nodal status (N), and metastatic disease (M).17,19 The

3151

current system was updated in 2009 and is summarized in Tables 107-2 and 107-3. There are five stages

based on prognosis: stage 0 (in situ melanoma), stage I (local disease), stage II (local disease), stage III

(regional nodal, in-transit, or satellite metastases), and stage IV (distant metastases).

Other Prognostic Factors

The major prognostic factors that predict survival in melanoma patients have been accounted for in the

AJCC staging system – namely, tumor microstaging, ulceration, nodal status, and distant

metastases.17,19 The presence of ulceration in a melanoma appears to be associated with a poorer

prognosis. Men have a higher proportion of ulcerated lesions than women (27% vs. 19%, respectively).

Although ulceration appears to correlate with thickness of the melanoma, the presence of ulceration is

an independent prognostic factor and has been included in the staging system.17,19 In addition, a higher

mitotic rate of 1.0 mm2 or greater is an independent prognostic factor of both nodal involvement and

survival.18,20–22 The presence of angiolymphatic invasion is also a poor prognostic sign.

Treatment of Primary Melanoma

Biopsy

For localized melanoma, the tumor thickness (Breslow depth of invasion) is the single variable that

most accurately determines therapy and prognosis. The ulceration status and mitotic rate also represent

independent prognostic variables. A full-thickness biopsy with 1- to 2-mm margins only, to the adipose

tissue, is preferred for any lesion highly suspect for melanoma.23 If the melanoma is transected with a

partial-thickness shave biopsy, the ability to obtain an accurate measurement of tumor thickness is lost.

Therefore, a superficial shave biopsy is never recommended for a suspect pigmented lesion.

Saucerization, which uses a curved blade to perform a deeper shave biopsy down to the subcutaneous

fat, is acceptable.

Excisional biopsy with 1- to 2-mm margins is the preferred method for suspect lesions to provide the

pathologist a total specimen for histologic interpretation and accurate microstaging. Performing a wide

excision as the first step, especially on the trunk or head and neck, is not recommended, as several

benign and malignant lesions can mimic melanoma and because doing so may result in the inability to

accurately perform SLNB. Formalin-fixed, paraffin-embedded, permanent sections should be used for

biopsy diagnosis of primary cutaneous melanoma to accurately determine tumor thickness and other

histopathologic prognostic variables. Frozen sections have no role in the diagnosis or microstaging of

primary melanoma. If the lesion is a melanoma, the excisional biopsy represents the first stage of a twostage procedure. The second stage is reexcision with margins generally ranging from 0.5 to 2.0 cm, with

or without SLNB, depending primarily on the tumor thickness. When biopsying a suspected melanoma

on the extremities, it is critical that the biopsy be oriented on the long axis of the extremity, parallel to

the lymphatics. Performing a biopsy oriented transversely on the extremity may not only compromise

the accuracy of lymphatic mapping but may also obligate the patient to a skin graft to close the

subsequent wide excision defect.

For suspicious lesions that are too large for complete excision and those that are located where the

amount of skin is critical in terms of functional or cosmetic results, an incisional biopsy may be

performed with an elliptical incision, deep saucerization shave to adipose or deep dermis in thick skin

areas, or a punch biopsy. Incisional biopsies for melanoma do not increase the risk of local recurrence

and distant metastasis or affect patient survival. They should generally be performed on the most raised

or most pigmented area of the lesion to maximize the obtainable diagnostic and prognostic information.

The most raised area usually corresponds to the maximal thickness of the lesion, but not always. Several

incisional biopsies can be obtained from different areas for large lesions with multiple morphologic

features. In the scenario of incisional biopsy with significant remaining lesion, complete excision for

accurate microstaging should be considered prior to definitive treatment, unless the lesion is already 1

mm or greater in Breslow depth.

Metastatic Workup

In an attempt to standardize staging workup for melanoma, the National Comprehensive Cancer

Network (NCCN) has published guidelines.24 There are three basic reasons to perform a metastatic

workup following the diagnosis of primary cutaneous melanoma: (a) for staging and prognosis, (b) to

detect an early metastasis with potential survival benefit, and (c) to avoid morbidity of an extensive

surgical procedure by detection of a distant metastasis.24 The best test for the staging workup still starts

3152

with a history (a focused review concentrating on constitutional, respiratory, neurologic, hepatic,

musculoskeletal, gastrointestinal, skin, and lymphatic systems) and physical examination (total body

skin examination, palpation of lymph nodes). Routine imaging (chest x-ray, computed tomography [CT]

scans, positron emission tomography [PET] scans) and blood studies (complete blood cell count,

comprehensive metabolic profile, liver function tests, serum lactate dehydrogenase [LDH] levels) in

asymptomatic, clinically node-negative patients are low in both sensitivity and specificity and are not

necessary.25–28 False-positive staging tests are common and lead to more tests and patient distress. SLNB

represents the best baseline staging test, with both relatively high sensitivity and specificity in patients

at significant risk for metastasis. The ability to detect stage IV disease with routine studies is small if the

SLNB is negative. Both ultrasound and PET scanning have been examined as methods of identifying

nodal metastases prior to surgery. While they detect disease in a minority of patients, the sizes of most

metastases to the regional nodes are below the threshold of ultrasound or PET to detect, so this is not a

cost-effective strategy.29–31 If a thorough history and physical and detailed review of systems reveal no

signs or symptoms suspicious for regional or distant disease, then no further staging is necessary.

However, a high index of suspicion should be maintained and a thorough symptom-directed workup

should be initiated for any worrisome finding on the history and physical.

STAGING

Table 107-2 American Joint Commission on Cancer Melanoma Staging System,

TNM Definitions

3153

Approximately 5% to 10% of newly diagnosed melanoma patients present with clinically involved

lymph nodes. Any clinically suspicious node should be confirmed by fine needle aspiration (FNA)

biopsy, as enlarged lymph nodes may be reactive (particularly after a biopsy). Ultrasound-guided FNA

biopsy can be useful in cases where the suspicious node is difficult to biopsy by hand. Patients

documented to have macroscopic regional disease should undergo complete staging prior to definitive

surgery, as a significant percentage may have more advanced regional or distant disease that may alter

surgical therapy.32,33 Staging should include a serum LDH level; MRI of the head and either CT scan of

the chest, abdomen, and pelvis; PET scan; or combined CT/PET scan. For patients with involved groin

nodes, CT scan is recommended to visualize the pelvic nodes, as this may help determine the extent of

inguinal node dissection.

STAGING

Table 107-3 Stage Groupings for Cutaneous Melanoma

Surgical Excision of Primary Melanoma

5 The primary purpose of a melanoma excision is to prevent local recurrence due to persistent disease.

For melanoma in situ, excision margins of 0.5 to 1.0 cm are indicated. For invasive melanoma, wide

excision of the primary tumor, with margins generally ranging from 1 to 2 cm, is indicated for local

control. The optimal margin width remains somewhat controversial. The historical approach of excising

all primary melanomas with a 3- to 5-cm margin is extinct. At least five randomized controlled trials

failed to demonstrate a difference in overall survival or local recurrence with narrow (1- to 2-cm)

versus wide (3- to 5-cm) margins (Table 107-4). Of note, one trial demonstrated an increased risk of

locoregional recurrence with 1-cm versus 3-cm margins for melanoma more than 2 mm thick.34 Based

on these trials, current consensus guidelines recommend margins of 1 cm for melanomas 1 mm or less

thick, 2 cm for melanomas greater than 2 mm thick, and 1 to 2 cm for melanomas 1 to 2 mm thick. For

this last group, the size of the margins should be based on the anatomic location of the melanoma and

the morbidity of excising 2 cm versus 1 cm, including the need for an advancement flap or skin graft

3154

(Table 107-5). For several locations, such as melanomas on the head and neck, the maximum margins

attainable are dictated by anatomic limitations. Most subungual invasive melanomas require amputation

of the digit. Perianal melanomas should be treated by wide excision, as these achieve similar long-term

results compared with more radical surgery.35,36 Abdominoperineal resection (APR) may be necessary

for locally advanced disease or melanoma of the anal mucosa. A complete staging workup should be

performed before proceeding with APR, as these patients have a very high rate of regional and distant

disease.37

RESULTS

Table 107-4 Randomized Controlled Trials: Narrow Versus Wide Excision Margins

TREATMENT

Table 107-5 Recommended Surgical Margins for Melanoma Excision

For the clinically ill-defined LMM, histologic confirmation of negative margins is important, and it

should be emphasized that these margin recommendations may not apply to in situ LM and invasive

LMM on the head and neck.38,39 LM and LMM are often associated with extensive subclinical

involvement to as much as several centimeters beyond the clinical component. Careful and complete

margin control, including excision of the lesional trailing edge of atypical melanocytic hyperplasia,

often beyond the standard margins, may be necessary for complete surgical resection of LM/LMM on

the head and neck.

Treatment of Regional Metastatic Melanoma

Lymphadenectomy Results and Indications

6 Surgical excision of metastases to regional lymph nodes is potentially curative therapy. The 5-year

survival rate for patients who undergo lymphadenectomy for clinically positive involved nodes (AJCC

stage III) ranges from 25% to 70%. In addition, for those patients not cured by lymphadenectomy,

resection can avoid potential pain associated with tumor enlargement, skin breakdown, and tumor

necrosis (Fig. 107-2). Only 5% to 10% of patients who first present with the diagnosis of melanoma

have clinical evidence of nodal metastases, approximately 85% to 90% have localized disease, and the

remaining 5% have distant metastases. In less than 3% of patients, a diagnosis of melanoma is made in

the absence of a definable primary lesion.40 When patients present with isolated nodal disease from an

unknown primary site (an occult primary melanoma), the results of lymphadenectomy are similar to or

better than those for patients with known primary tumors.41 For patients with melanoma 1 mm thick or

greater who present with clinically negative nodes, SLNB should be considered to determine whether

therapeutic lymphadenectomy is indicated.42

3155

Figure 107-2. A patient with large melanoma axillary involvement with skin ulceration.

Sentinel Lymph Node Biopsy

Morton et al.43 were the first to develop the concept of “sentinel lymph nodes” in the nodal basins

draining cutaneous melanomas. They hypothesized that melanoma involvement of a nodal basin

develops in an orderly fashion with metastasis to the sentinel lymph node as the first step in that

process. With the intradermal injection of a blue dye, these investigators were able to identify sentinel

lymph nodes 90% of the time. If the sentinel node was negative for melanoma, the remaining lymph

nodes were also free of involvement in at least 96% of cases. These results have been confirmed by

multiple other groups.44,45

Two tracers are typically used for identifying the sentinel lymph nodes: a radiolabeled colloid

solution, and a blue dye – either methylene blue or 1% isosulfan blue (Lymphazurin). Lymphazurin

carries a risk of allergic reaction in 0.5% to 2.0% of patients, which can be life-threatening.45 Methylene

blue dye has an extremely low risk of allergic reaction, although this should be diluted as it can cause

skin necrosis if not completely resected. One to 4 hours before surgery, a radiolabeled colloid solution is

injected intradermally around the primary tumor site or excision site, taking care to avoid injecting

directly into any residual tumor. Lymphoscintigraphy imaging is then performed, which is critical in

that it can identify sentinel nodes outside the traditional nodal basin. The blue dye is injected

intradermally in the operating room a few minutes before the procedure. Using a handheld gamma

detector probe, the surgeon is able to identify the sentinel lymph node location through the skin,

thereby limiting the incision necessary to find the node. Blue-stained lymphatics and lymph nodes can

be directly visualized (Fig. 107-3). The combined use of the blue dye plus the radiolabeled colloid

enables the detection of the sentinel node in 97.3% to 98.6%.46 The technique of SLNB is most often

performed in conjunction with the wide excision of the primary tumor. It can routinely be performed as

an outpatient procedure.

3156

Figure 107-3. Blue dye tracking up a lymphatic vessel to a blue-stained lymph node.

Once the sentinel lymph node is removed, it is processed by the pathologist by step sectioning the

entire node into multiple sections for routine hematoxylin and eosin staining. This is done to examine

for micrometastases that could be missed by bivalving the node and performing an examination of only

one section. If serial sectioning and staining are negative for metastasis, then immunohistochemical

staining for melanoma markers such as S-100, Melan-A, and HMB-45 is performed.47 These stains can

identify microscopic clusters of tumor cells that are hard or impossible to identify by hematoxylin and

eosin staining. The thoroughness of assessing the sentinel lymph nodes for metastatic disease has

resulted in much more accurate staging of patients. Micrometastases to lymph nodes have been

incorporated in the current staging system as described earlier.

If the sentinel lymph node is positive for melanoma, then complete lymph node dissection (CLND)

(Fig. 107-4) is indicated on the basis of potential survival and clinical benefit. The final results of the

Multicenter Selective Lymphadenectomy Trial I (MSLT-I), which randomized patients to wide local

excision (WLE) alone or WLE in combination with SLNB and CLND for a positive SLN, demonstrated

that disease-free survival rates were significantly improved in the SLN group, compared with the

observation group, for both patients with intermediate-thickness melanoma (71.3 ± 1.8% vs. 64.7 ±

2.3%; HR = 0.76, P = 0.01) and thick melanoma (50.7 ± 4.0% vs. 40.5 ± 4.7%; HR = 0.70, P =

0.03).48 Although MSLT-I showed no benefit on overall survival, subset analysis showed that node

positive patients discovered on SLN biopsy had a 10-year melanoma-specific survival of 62.1 ± 4.8%

compared with 41.5 ± 5.6% for patients who recurred during observation. Even after factoring for

false-negative sentinel nodes, this remained significant (HR = 0.67; 95% CI, 0.46 to 0.97; P = 0.04).

Although this subset analysis has been questioned as being valid,49,50 latent-subgroup analysis showed a

clear, significant benefit of sentinel node biopsy.48

An unresolved question is whether every patient with a positive sentinel lymph node requires CLND.

Several attempts to retrospectively identify clinical or histologic factors that may predict the absence or

presence of disease within the nonsentinel nodes have not identified any absolute way to determine who

requires additional surgery, although a combination of the Breslow depth and burden of disease within

the sentinel lymph node seems promising.51–54 In addition, retrospective studies of SLN-positive patients

who did not undergo CLND appear to have similar outcomes to those who did.55–57 The Multicenter

Selective Lymphadenectomy Trial-II (MSLT-2), which randomized patients with a positive sentinel

lymph node to CLND or serial ultrasonography of the regional basin, has completed accrual and will

provide more information regarding the benefit of CLND for sentinel lymph node–positive patients.

Until that time, CLND for a positive sentinel lymph node remains the standard of care in the treatment

of melanoma.

3157