Chapter 120. Merkel Cell Carcinoma

Merkel cell carcinoma (MCC) is an increasingly common neuroendocrine skin cancer that is associated with ultraviolet (UV)-light exposure, advanced age, immune suppression and a recently discovered polyomavirus. The reported incidence of MCC has more than tripled in the past 20 years1 to approximately 1,500 US cases/year2 and is expected to grow with the aging population. Although it is 40 times less common than malignant melanoma, it carries a markedly poorer prognosis, with disease-associated mortality at 5 years of 46%3 as compared with 9% for invasive melanoma.4 Management of MCC is challenging and optimal therapy is controversial, at least in part due to a lack of prospective or randomized data on which to base treatment decisions.

MCC is a relatively recently described entity, although the Merkel cell was identified more than 100 years ago. In 1875, human Merkel cells were first described by Friedrich S. Merkel (1845–1919). He named these cells Tastzellen (touch cells) assuming that they had a sensory touch function within the skin due to their association with nerves. In 2009, a series of studies using elegant mouse models resolved several long-standing debates by conclusively establishing that (1) Merkel cells are essential for light touch responses,5 (2) Merkel cells have an epidermal origin,5,6 and (3) Merkel cells do not divide, but are renewed by a reservoir of epidermal progenitor cells.6

In terms of what we now call Merkel cell carcinoma, in 1972, Toker described five cases of “trabecular cell carcinoma of the skin.” In 1978, Tang and Toker found that cells from these tumors contained dense core granules on electron microscopy that were typical of Merkel cells and other neuroendocrine cells. In 1980, the name Merkel cell carcinoma (MCC) was first applied to this tumor because of the characteristic ultrastructural features it shares with normal Merkel cells. In 1992, it was found that antibodies to cytokeratin-20 stain normal skin Merkel cells as well as the vast majority of MCC tumors. This critical finding allows specific and relatively easy diagnosis of MCC to be made through immunohistochemistry. Since that time, electron microscopy is no longer used to make this diagnosis.

Over the 20-year period between 1986 and 2006, the reported incidence of MCC quadrupled from 0.15 per 100,0001 to 0.6 per 100,000.7 There are likely two factors that contribute to this increase in reported incidence that is more rapid than that of any other type of skin cancer. One factor is an increase in the accurate diagnosis of this malignancy through the routine use of cytokeratin-20 immunohistochemistry and the improved recognition of this malignancy by dermatopathologists. In the past, MCCs were often mischaracterized as lymphoma, melanoma, or undifferentiated carcinoma in the era before immunohistochemistry. A second likely reason is an increase in the number of people over age 65 years with extensive sun exposure history and people living with prolonged immune suppression. Each of these factors is a known risk factor for MCC and several are discussed in Section “Etiology and Pathogenesis.”

As of 2008, there are approximately 1,500 MCC cases per year in the US.2 This cancer is far more common in whites than in blacks, consistent with a known role for UV radiation in MCC pathogenesis. Specifically, the rates in whites have been reported as 0.23 per 100,000 as compared to 0.01 per 100,000 for blacks.8 Although not specifically reported, rates in Hispanics and Asians are likely intermediate between those in blacks and whites. MCC tends to be more common in men than in women (2:1 in ratio).9 Furthermore, men also tend to have a worse prognosis than women, with reported 10-year disease-associated survival rates of 51% for men and 65% for women.7

There are several known risk factors for MCC that will likely continue to lead to an increase in the incidence of this disease.

Age Greater Than 65 Years

The median age for diagnosis of MCC is 70 years, and there is a five- to tenfold increase in incidence after age 70 as compared with age less than 60 years, as shown in Fig. 120-1. Indeed, only about 10% of MCC cases present in patients under age 50 and it is extremely rare in childhood.10

Sun Exposure

Sunlight, prolonged UV exposure, and photochemotherapy are all associated with an increased risk of MCC. As seen in Fig. 120-2, the vast majority (81%) of MCC tumors present on sun-exposed skin.10 However, it is clear that sun is not required for MCC to develop. MCC cases can occur on sun-protected skin, including buttocks and vulva as well as portions of the scalp that are covered by hair.

Immune Suppression

As compared to most cancers, MCC is strongly linked to immune suppression. Indeed, 7.8% of MCC patients are profoundly immune suppressed, a 16-fold overrepresentation compared with expected.10 Multiple forms of immune suppression are associated with an increase in MCC risk. These include HIV/AIDS,11 chronic lymphocytic leukemia,10 and the immune suppressive regimens associated with solid organ transplant.9 Additionally, there are 19 reported cases of complete spontaneous regression in the MCC literature, a far greater number than expected for its rarity, perhaps suggesting a sudden immune recognition and clearance of MCC.8,12

Merkel Cell Polyomavirus

In 2008, a new polyomavirus, the Merkel cell polyomavirus (abbreviated MCV or MCPyV), was discovered by extensive sequencing of MCC tumor RNA. MCPyV was reported to be present in 80% of MCC tumors compared to only 7% of skin controls.13 Numerous groups across several continents (over 25 literature citations) have verified the association between MCPyV and MCC, and all groups observed that the virus is strongly associated with MCC. MCPyV is a small, circular, double-stranded DNA virus related to other known polyomaviruses (SV40, BK, JC, KI, WU). MCPyV is unique among the human polyomaviruses as it is the only one proven to integrate into a human cancer. Monoclonal integration of MCPyV into MCC tumors suggests that the virus is present prior to or very early in tumorigenesis.

A large fraction of the population has been infected with MCPyV. By the age of 5, prevalence of antibodies is about 35%, but increases to 50% by the age of 15.14 Interestingly, among MCC patients, the prevalence is significantly higher (88%) than age and sex-matched controls (53%).15 Importantly, despite how common exposure is to this virus in the general population, the incidence of MCC is very low. Therefore, viral infection alone is clearly insufficient for the development of this cancer.

A systematic analysis of 195 MCC patients characterized clinical features that may serve as clues in the diagnosis of MCC.10 The most significant features can be summarized in an acronym: “AEIOU” (Table 120-1). This study found that 89% of 62 MCC cases exhibited at least three of the five features listed below. If a lesion exhibits at least three of these features, suspicion of MCC should increase and biopsy be considered. In particular, a lesion that is red or purple, rapidly growing, but nontender should be of concern. Strikingly, the clinician listed as clinical impression a benign diagnosis in over one half of cases. In particular, a cyst or acneiform lesion was the single most common of these (Box 120-1 and Figs. 120-3A and 120-3B) and nonmelanoma skin cancer was also relatively frequent as a presumptive diagnosis (Fig. 120-3C).

As shown in Fig. 120-4A, the classic histologic features of MCC include sheets of small basophilic cells with scant cytoplasm, fine chromatin, and no nucleoli. There are numerous mitotic figures and occasional individual necrotic cells. Lymphovascular invasion is a very common feature and often can be found when it is specifically searched for even in a “negative” margin. This helps to explain the high local recurrence rate in MCC for narrow or even relatively wide margin excision when adjuvant radiation therapy is not given.

Hematoxylin and Eosin Stain

Three histologic patterns have been described up to now and none is clearly associated with a better or worse prognosis. The most common type is the “intermediate type.” This has uniform small cells with minimal cytoplasm, pale nuclei, and a dispersed chromatin appearance. On hematoxylin and eosin staining, the differential diagnosis for this presentation is that of the small, blue-cell tumors including melanoma and lymphoma. The second most common pattern is the “small cell type.” This takes its name from small cell lung carcinoma, which is the principal differential diagnosis for this pattern. It shows irregular, hyperchromatic cells with scant cytoplasm and malignant cells that are arranged in linear patterns infiltrating stromal structures. The least common but perhaps most histologically distinctive type is the “trabecular” type. This is the pattern originally described by Toker in 1972. It has a lattice-like, or network appearance, and the differential diagnosis includes metastatic carcinoid tumor.

Immunohistochemical Stains

The use of antibody-based stains has greatly facilitated the ease and specificity of MCC diagnosis (Table 120-2). The single most useful of these stains is cytokeratin-20.

Cytokeratin-20

Intermediate filament protein is expressed in MCC as well as in adenocarcinomas of the colon, stomach, and pancreas. However, within the skin the expression of cytokeratin-20 is limited to Merkel cells. A “perinuclear dot” pattern of cytokeratin is essentially pathognomonic for MCC (see Fig. 120-4B).

Antibodies to CAM5.2

CAM5.2, a cocktail of antibodies that detects multiple human cytokeratin epitopes, typically reacts with both MCC and small cell lung carcinoma. Although it is useful as an initial screening tool to detect tumors of squamous origin, its lack of selectivity relative to cytokeratin-20 means that it cannot be used to definitively diagnose MCC.

Thyroid Transcription Factor-1

Thyroid transcription factor-1 is negative in MCC and positive in small cell lung cancer and thus useful for the differential diagnosis between these two tumors that can look identical by routine histology.

Cytokeratin-7

Cytokeratin-7 has the same staining pattern as thyroid transcription factor-1, that is, typically negative in MCC and positive in small cell lung carcinoma; it is also expressed in epithelial cells of the lung, ovary, and breast.

In 2010, the American Joint Committee on Cancer (AJCC) adopted a new staging system for MCC that replaced five existing—and conflicting—staging systems.16 According to the new AJCC guidelines, there are four stages for MCC based on clinical and pathological features at the time of presentation. This new staging system is summarized in Fig. 120-5 and Table 120-3. Survival after a diagnosis of MCC is highly dependent on the stage at presentation. Patients with local disease have estimated 5-year relative survival between 60% and 79%. Survival decreases markedly with nodal involvement and metastatic disease (Fig. 120-5). Unlike malignant melanoma, if MCC recurs it tends to do so rapidly with ∼80% of recurrences occurring within 2 years of diagnosis.17

Until 2009, MCC did not have its own specific International Classification of Diseases (ICD) code. Instead, MCC was often coded as “173.x: other malignant neoplasm of the skin.” Specific ICD codes are important because they allow for more streamlined therapy, faster insurance approvals and improved tracking of patients and costs for clinical research purposes. As of 2010, MCC now has its own set of ICD codes (209.3x) to better facilitate treatment and tracking of patients with MCC.18

Sentinel Lymph Node Biopsy and Staging of Merkel Cell Carcinoma

Over the past 20 years, sentinel lymph node biopsy (SLNB) has become quite common in staging malignant melanoma presenting with a depth of greater than 1 millimeter. More recently, several studies have indicated that SLNB is also a sensitive test for detecting MCC spread to the lymph nodes.19–21 Interestingly, MCC is far more likely to have occult lymph node involvement (approximately 30% for the average 1.7 cm MCC) than melanoma (∼1% for melanomas with the average Breslow thickness of 0.63 mm).21,22 SLNB clearly has less morbidity than an elective lymph node dissection. Ideally, an SLNB should be performed at the time of the wide resection as opposed to after the wide excision when local lymphatics have been disturbed. The importance of performing an SLNB is that one-third of patients with local-only disease as determined by clinical palpation of lymph nodes, in fact, have microscopic disease in the lymph nodes that very much affects their subsequent survival (Fig. 120-5, Table 120-3).21 Specifically, one-third of patients would be understaged by clinical exam only, and would shift to stage IIIA disease with microscopic examination. Identification of microscopic nodal disease also importantly alters therapy for the node bed, as discussed below.

Radiologic Imaging Studies

Although many patients who present for management of MCC have not had an SLNB, the majority have had computed tomographic (CT) scans. We studied our own series of patients to determine the sensitivity and specificity of CT scans in patients presenting with MCC.21 We found that scans were not sensitive (missing 90% of positive cases) in terms of detecting nodal involvement of MCC as compared with SLNB. Among patients who had low-risk disease (negative sentinel lymph nodes or very small primary tumors) all of the “positive” scan results were, in fact, false positives.21 Therefore, we typically reserve CT or PET-CT scans for patients presenting with more advanced disease, such as positive nodal involvement or clinical evidence of metastatic disease.

Prognostic Findings on Histology

A recently published retrospective study evaluated several histologic features of MCC and performed survival analyses on these features. This large study had over 4 years of average follow-up time and analyzed the primary tumor lesions from 156 patients with MCC. In addition to tumor stage, two new histologic features were of particular interest: (1) lymphovascular invasion (LVI) and (2) tumor growth pattern. LVI was defined as tumor emboli within vascular spaces. Persons with MCC tumors with detectable LVI had a worse overall survival as compared to those with tumors without LVI (hazard ratio, 3.84; P = 0.007).23 Tumor growth pattern was described as nodular (well-circumscribed interface between tumor and surrounding tissue) or infiltrative (rows, trabeculae, or single cells that penetrate the dermis). Tumors that exhibited both features were considered infiltrative. An infiltrative tumor growth pattern was associated with poor outcomes as compared to MCC tumors with a nodular growth pattern (hazard ratio, 6.85; P = 0.001).23

Optimal therapy for MCC is controversial and there is no broad consensus on how to manage this disease. The best summary of consensus treatment for MCC is available through the National Comprehensive Cancer Network (http://www.nccn.org) and is updated annually (these guidelines and other useful information can also be found at http://www.merkelcell.org/usefulInfo/index.php). The best outcome for MCC is clearly obtained when multidisciplinary management is carried out by an experienced team. Each major treatment modality is summarized below.

Surgery at Primary Site

The initial management in most cases of MCC is surgical excision of the tumor. When carried out without subsequent adjuvant radiation, surgery may have a relatively high recurrence rate depending on the margins chosen and the risk profile for the tumor. Surgery alone with 0.5 cm margins resulted in a 100% recurrence rate in 38 patients.24 When wide local excision was carried out with 2.5 cm margins, the recurrence rate fell to 49%.25 For Mohs surgical excision, the local recurrence rate was 16% for surgery alone and 0% in patients who also had adjuvant radiation therapy, although this difference was not statistically significant.26 There are no data to suggest that extremely wide excision margins improve overall survival. Depending on the location of the tumor, significant morbidity can result when 2–3 cm margins are taken. Numerous studies show that if surgery is the sole treatment, recurrence rates are significantly higher than if radiation therapy is added to the regimen. Excision with narrow but clear margins (carried out at the time of SLNB) followed by adjuvant radiation therapy is a reasonable approach to management in many cases. Overly aggressive surgery, including amputation, or very wide margins in cosmetically sensitive areas, decreases quality of life, increases morbidity, delays time to initiation of adjuvant radiation, and does not appear to improve survival or local control rates.

Surgery at the Draining Node Bed

Completion lymphadenectomy is typically carried out if there is gross involvement of the draining node bed detected clinically. However, the role of completion lymphadenectomy in MCC is controversial. For example, a 2010 study suggests that patients with nodal MCC disease who underwent complete lymphadenectomy had comparable outcomes with patients who only underwent radiation therapy to the nodal bed.27 These two options have similar and excellent control rates for node-positive disease. Although either can be chosen, depending on the clinical situation, it is clear that for microscopic nodal disease, only one of these two modalities should be performed. This is because regional control rates were 100% for each modality and the combination of radiation and surgery to the lymph node bed greatly increases risk of chronic lymphedema.

Radiation Therapy

MCC is an unusually radiosensitive tumor.25,28 Recently, there have been reports of successful treatment of MCC with radiation as monotherapy. One study of 43 patients with MCC for whom surgical excision was not possible demonstrated an excellent infield control rate (75%) after treatment with radiation therapy alone.29 In a separate cohort, there were no recurrences among nine patients treated with radiation monotherapy.28 One of our cases treated with radiation monotherapy is shown in Fig. 120-6. In most of these cases, the lesion was felt to be inoperable and radiation was given for palliation, but typically resulted in long-lasting local control.

The much more common use of radiation is as an adjuvant to surgery. Adjuvant radiation clearly is critical if surgical margins are positive or if microscopic margins are relatively narrow (<0.5 cm). A retrospective review of the literature indicates a statistically significant improved local and nodal rate of control in this cancer if radiation is added.30 A cancer registry-based study also indicates improved survival in patients with adjuvant radiation therapy.31 In this study, patients who received adjuvant radiation therapy had a median survival of 63 months compared to median survival of 45 months in patients who did not receive adjuvant radiation. The typical doses of radiation for MCC are 50–56 Gy for a primary site with negative excision margins. Detailed dosing regimens can be found in the National Comprehensive Cancer Network's MCC Guidelines. Radiation doses are typically given in 2-Gy fractions, five times/week over 4–6 weeks. Acute side effects from radiation therapy include erythema at the site and mild-to-moderate fatigue that peaks toward the end of radiation and usually resolves within 1–2 months of completing a 5-week course. Chronic radiation skin changes include temporary or permanent alopecia within the irradiated field, epidermal atrophy, loss of adnexal structures leading to skin or mucosal dryness, and risk of subsequent secondary skin cancers in the irradiated region in patients with a life expectancy of greater than 20 years after the radiation treatment. Perhaps the most significant potential side effect is lymphedema. This is more commonly an issue in lower extremities, when radiation therapy is given to the inguinal lymph nodes, especially after surgery has also been carried out in that region. Early referral to a physical therapist trained in lymphedema management is indicated to minimize the severity and incidence of this potential complication in higher risk cases.

Chemotherapy

The most commonly used chemotherapeutic regimen for MCC is the combination of etoposide and either cisplatin (perhaps more clinically effective) or carboplatin (less nephrotoxic). Chemotherapy is often useful in palliation for symptomatic disease that is otherwise inoperable. MCC patients receiving chemotherapy for the first time usually have a significant response with shrinkage of the tumor. Unfortunately, in almost all cases, the tumor grows back and is resistant to chemotherapy even if entirely different agents are used on a subsequent round of chemotherapy. After careful analysis of the literature, we currently do not recommend adjuvant chemotherapy for patients whose MCC has been treated with surgery, radiation therapy, or both. There are six reasons why we currently do not recommend adjuvant chemotherapy:

1. Mortality: There is a 4%–7% acute death rate due to adjuvant chemotherapy in MCC partly due to the fact that these patients are often elderly.32,33

2. Morbidity: Neutropenia has been reported to occur in 60% of patients with fever, and sepsis in 40%.34

3. Decreased quality of life: This can be quite severe in this older population, including fatigue, hair loss, nausea, and vomiting.

4. Resistance to chemotherapy: MCC that recurs after chemotherapy is less responsive to later palliative chemotherapy.

5. Immunity: Chemotherapy suppresses immune function, and this is known in general to be very important in preventing and controlling MCC.

6. Apparent poorer outcomes: Among patients with nodal disease, there was a 60% survival if chemotherapy was not given among 53 patients. In contrast, survival was only 40% among node positive MCC patients who did receive adjuvant chemotherapy.17 Although this is not a randomized trial and was not statistically significant, it certainly does not suggest a major survival benefit for administrating adjuvant chemotherapy.

Optimal Treatment for Merkel Cell Carcinoma

In general, optimal treatment for MCC should involve obtaining pathologically clear margins by surgery, typically with 1- to 2-cm margins as possible, depending on the site. More narrow margins or even positive margins can often be effectively treated by local radiation, typically extending 3–5 cm beyond the tumor bed. We also recommend treating the draining lymph node bed, likely with radiation therapy, for patients with high-risk disease including a positive SLNB, immune suppression, or a tumor greater than 2 cm in diameter. Although still controversial, we currently do not recommend adjuvant radiation therapy for MCC patients with all of the following five good prognostic features: (1) primary tumor diameter ≤1 cm; (2) microscopic margins that are confidently negative following surgery; (3) no lymphovascular invasion noted in the tumor; (4) no profound immune suppression (HIV, chronic lymphocytic leukemia, etc.); and (5) SLNB that was negative with proper immunohistochemistry studies.

In more than 90% of cases, MCC is an unanticipated diagnosis at the time the pathology results become known. Most commonly, the lesion in question was thought to be a cyst or acneiform lesion. Although most patients and physicians are not familiar with this disease or its specific management, MCC can be lethal and there is a need to initiate therapy rapidly. In one Australian study, a high fraction of patients (45%) developed progressive disease while waiting for adjuvant radiation therapy to begin (median wait time 41 days).35 Because of the rarity of this cancer, patients and physicians increasingly resort to the Internet for medical information. One Web site, http://www.merkelcell.org, is devoted to aiding patients and physicians by presenting current data and referral center availability.

Among patients who experience a recurrence of their MCC, ∼80% of these happen within 2 years of diagnosis. The most common site of recurrence is the draining nodal basin or adjacent skin.17 For those who have recurrences, the locations and frequencies are skin (28%), lymph nodes (27%), liver (13%), lung (10%), bone (10%), brain (6%), bone marrow (2%), pleura (2%), and other sites (4%).33 Once MCC has spread to viscera, it is typically incurable.

Fortunately, for more than 50% of patients, a good outcome can be anticipated. Those who have had no recurrences for 3–5 years, unlike for melanoma, enjoy a greatly reduced risk of recurrence. The complications for those who do not experience a metastasis depend very much on the therapy they received. We believe that treatment with surgical excision and radiation therapy as outlined in this chapter has relatively minimal complication rates and the best possible chance of cure based on current literature. However, for those who receive very aggressive surgical excision, amputation or chemotherapy for low-risk disease, complication rates tend to be higher without improved outcomes.