1

Introduction

Cutaneous Merkel cell carcinoma (MCC) is a rare skin cancer of neuroendocrine origin that was first described by Toker in 1972 . The age-adjusted incidence is approximately 0.24 to 0.44 per 100 000 person years . Although the incidence is reportedly rising, it is unclear whether this represents a true increased incidence or an increased recognition of MCC. Risk factors for MCC are sun exposure and immune suppression, including chronic lymphocytic leukemia, solid organ transplant, and HIV . Human polyoma virus appears to be etiologic in a significant proportion of patients with MCC; the presence of human polyoma virus DNA in the MCC cells may be associated with an improved prognosis . Heath et al reported on 195 patients with MCC treated at 5 institutions (Boston, 3; Seattle, 2) between 1980 and 2007; 81% were located in ultraviolet-exposed sites. MCC exhibits a slight male preponderance . Mojica et al reported on 1665 patients from the SEER database from 1973 to 2000; men accounted for 59% and women for 41% of patients. The vast majority (more than 90%–95%) was white and approximately 90% were older than 50 years . Mojica et al reported a median age of 76 years (range, 22–101 years). The most common sites include the head and neck and extremities. Andea et al reported on 156 patients treated at the Memorial Sloan Kettering Cancer Center between 1980 and 2005. They observed the following site distribution: extremity, 42%; head and neck, 37%; buttocks, 16%; and trunk, 5% .

Most MCCs appear relatively innocuous at diagnosis. Heath et al reported a median maximum tumor diameter of 1.8 cm; 88% of patients were asymptomatic despite a history of rapid growth in 63% during the 3 months before diagnosis. The most common color was red/pink in 56%, followed by blue/violaceous in 26% . The lesion was thought to be benign before biopsy in 56% .

The diagnostic evaluation of the patient with MCC includes taking a thorough history, physical examination, chest radiograph, and computed tomography of the primary site and regional lymphatics. Fluorodeoxyglucose positron emission tomography–computed tomography will likely contribute to altered staging and a change in the treatment plan and should be obtained in most patients . The value of sentinel lymph node biopsy (SLNB) in the routine management of patients with MCC is debatable and depends on treatment philosophy . On one hand, if patients with a pathologically negative SLNB were followed up and adjuvant nodal radiotherapy (RT) was withheld, then SLNB would be valuable to define this subset of patients. In addition, one could argue that those with pathologically positive SLNBs could be considered for adjuvant chemotherapy because of the increased risk of distant metastases. Alternatively, adjuvant chemotherapy has not been shown to improve outcome in high-risk patients and, because of the high likelihood of subclinical disease in clinically negative regional nodes, it is the authors’ bias to electively irradiate these regions regardless of SLNB status . Thus, in the latter instance, SLNB does not meaningfully contribute to management decisions.

Several staging systems have been described for MCC . The staging system described by Yiengpruksawan et al is straightforward and has been widely used: stage I, local disease; stage II, regional disease; and stage III, distant metastasis. The staging system described by the American Joint Committee in Cancer is more complex and is ill-suited to an entity that is relatively rare and where the number of patients included in most single-institution outcome studies is relatively small . Mojica et al reported the following stage distribution in 1665 patients from the SEER database: stage I, 55%; stage II, 31%; stage III, 6%; and no data, 8%.

Surgery and RT are the mainstays of treatment for patients with stage I and II MCC .

Although a subset of patients with stage I disease may be managed with surgery alone, the high likelihood of subclinical disease in the clinically negative regional lymphatics and the modest risk of in-transit metastases suggest that most patients benefit from the addition of RT . Patients with stage II disease have approximately a 75% local-regional control rate after RT alone or combined with surgery . Although debatable, the addition of surgery to RT probably results in improved local-regional control .

The purposes of this article are to report the outcomes of a series of patients with MCC treated at our institution and to discuss the optimal management of patients with this disease.

2

Materials and methods

Between December 1984 and November 2009, 40 previously untreated patients with cutaneous MCC were treated with curative intent at the University of Florida.

During the same period, 10 patients were treated with curative intent for a local-regional recurrence after surgery alone. Their outcomes are reported separately because the prognosis is likely worse for patients with recurrent disease.

There were 32 men (80%) and 8 women (20%). The median age was 69 years (range, 42–87 years). All patients were white. Primary sites were the following: head and neck, 32 patients (80%); extremities, 6 patients (15%); and trunk, 2 patients (5%).

Patients were staged according to the Yiengpruksawan staging system: stage I, 24 patients (60%); and stage II, 16 patients (40%) .

Surgery was part of the initial treatment in 37 patients and included resection of the primary lesion in all 37 patients and also included regional lymph node dissection in 10 patients (27%). Of those with stage I disease, 1 of 23 patients underwent a regional node dissection, compared with 10 of 14 patients with stage II disease. Resection margins were the following: negative (R0), 27 patients (73%); microscopically positive (R1), 6 patients (16%); and grossly positive (R2), 4 patients (11%).

All patients received RT: RT alone, 3 patients; and adjuvant RT, 37 patients. The 3 patients who received RT alone included 1 patient with a stage I MCC of the tip of the nose, 1 patient with a stage II scalp MCC with a satellite lesion, and 1 patient with MCC of the arm and axillary node metastases. This last patient was scheduled for post-RT resection and declined. Of those who received adjuvant RT, 33 patients were treated postoperatively, and the remaining 4 patients were treated preoperatively. For those treated with both surgery and RT, the median interval between the modalities was 38 days (range, 13–90 days). The median RT dose was 60.0 Gy (range, 33.0–74.0 Gy). Thirty-three patients (83%) were treated with once-daily fractionation and the remainder with either hyperfractionation or the concomitant boost technique. Radiotherapy was administered to the regional nodes in 22 of 24 patients with stage I disease and 14 of 16 patients with stage II disease. Four patients with positive regional nodes received RT to the nodes without a node dissection. One patient with a positive SLNB has no gross residual disease; the remaining 3 patients had grossly positive nodes treated with RT alone.

Adjuvant chemotherapy was administered to 11 patients (28%). The agents used included etoposide (VP-16) (9), cisplatin (4), carboplatin (3), cisplatinum (1), and isotretinoin (Accutane) (1). The number of cycles ranged from 1 to 4.

The median follow-up for all patients was 3.0 years (range, 0.6–15.6 years). The median follow-up for surviving patients was 4.2 years (range, 2.2–14.2 years). No patients were lost to follow-up.

SAS and JMP software (SAS Institute, Cary, NC) were used for all statistical computations. The rates of local control, regional control, local-regional control, distant metastasis-free survival, cause-specific survival, and overall survival were calculated according to the Kaplan-Meier product-limit method . Complications were coded as severe if they resulted in hospitalization, necessitated surgical intervention, and/or resulted in death .

2

Materials and methods

Between December 1984 and November 2009, 40 previously untreated patients with cutaneous MCC were treated with curative intent at the University of Florida.

During the same period, 10 patients were treated with curative intent for a local-regional recurrence after surgery alone. Their outcomes are reported separately because the prognosis is likely worse for patients with recurrent disease.

There were 32 men (80%) and 8 women (20%). The median age was 69 years (range, 42–87 years). All patients were white. Primary sites were the following: head and neck, 32 patients (80%); extremities, 6 patients (15%); and trunk, 2 patients (5%).

Patients were staged according to the Yiengpruksawan staging system: stage I, 24 patients (60%); and stage II, 16 patients (40%) .

Surgery was part of the initial treatment in 37 patients and included resection of the primary lesion in all 37 patients and also included regional lymph node dissection in 10 patients (27%). Of those with stage I disease, 1 of 23 patients underwent a regional node dissection, compared with 10 of 14 patients with stage II disease. Resection margins were the following: negative (R0), 27 patients (73%); microscopically positive (R1), 6 patients (16%); and grossly positive (R2), 4 patients (11%).

All patients received RT: RT alone, 3 patients; and adjuvant RT, 37 patients. The 3 patients who received RT alone included 1 patient with a stage I MCC of the tip of the nose, 1 patient with a stage II scalp MCC with a satellite lesion, and 1 patient with MCC of the arm and axillary node metastases. This last patient was scheduled for post-RT resection and declined. Of those who received adjuvant RT, 33 patients were treated postoperatively, and the remaining 4 patients were treated preoperatively. For those treated with both surgery and RT, the median interval between the modalities was 38 days (range, 13–90 days). The median RT dose was 60.0 Gy (range, 33.0–74.0 Gy). Thirty-three patients (83%) were treated with once-daily fractionation and the remainder with either hyperfractionation or the concomitant boost technique. Radiotherapy was administered to the regional nodes in 22 of 24 patients with stage I disease and 14 of 16 patients with stage II disease. Four patients with positive regional nodes received RT to the nodes without a node dissection. One patient with a positive SLNB has no gross residual disease; the remaining 3 patients had grossly positive nodes treated with RT alone.

Adjuvant chemotherapy was administered to 11 patients (28%). The agents used included etoposide (VP-16) (9), cisplatin (4), carboplatin (3), cisplatinum (1), and isotretinoin (Accutane) (1). The number of cycles ranged from 1 to 4.

The median follow-up for all patients was 3.0 years (range, 0.6–15.6 years). The median follow-up for surviving patients was 4.2 years (range, 2.2–14.2 years). No patients were lost to follow-up.

SAS and JMP software (SAS Institute, Cary, NC) were used for all statistical computations. The rates of local control, regional control, local-regional control, distant metastasis-free survival, cause-specific survival, and overall survival were calculated according to the Kaplan-Meier product-limit method . Complications were coded as severe if they resulted in hospitalization, necessitated surgical intervention, and/or resulted in death .