Merkel cell carcinoma: a clinicopathologic study with prognostic implications

Background: Merkel cell carcinoma (MCC) is a frequently aggressive neuroendocrine malignancy of the skin that presents in sun-exposed areas on elderly patients. Although originally described over 30 years ago, many aspects of MCC remain to be defined. Of particular importance is the need to identify prognostic factors capable of predicting the biological behavior of these tumors. Knowledge of these factors may help in determining which patients require more aggressive treatment regimens. In this study, we examined 25 cases of MCC with an attempt to identify clinical, histopathological, or immunohistochemical features capable of predicting disease outcome. Methods: Features that we evaluated in each case included age, gender, race, tumor location, tumor size, depth of invasion, growth pattern, lymphocytic infiltration, mitotic activity, ulceration, necrosis, vascular invasion, and perineural invasion. In addition, we examined neural cell adhesion molecule and cytokeratin-20 expression using immunohistochemical methods. Resualts: We found that most patients were males (84%) with an average age of 74 years. The tumors were located on the head and neck (68%) and upper extremities (32%). Overall, 64% of the patients developed metastatic disease to regional lymph nodes or distant sites (average follow-up time of 21 months). Local recurrence was also common, occurring in 29% of the patients. The overall 1- and 2-year survival rates were 80 and 53%, respectively. Histopathological examination revealed tumors with an average size of 7.2 mm. Common features included invasion into the subcutaneous adipose tissue, solid growth pattern, tumor necrosis, and vascular and perineural invasion. Findings that had a statistically significant correlation with poor outcome included tumor size greater than or equal to5 mm (p = 0.047), invasion into the subcutaneous adipose tissue (p = 0.005), diffuse growth pattern (p = 0.040), and heavy lymphocytic infiltration (p = 0.017). The remaining findings, including the immunohistochemical results, did not correlate with disease outcome. Using logistic regression models, we show that depth of invasion and degree of lymphocytic infiltration are strong predictors of disease outcome. Conclusions: The current controversies regarding the treatment of early-stage MCC (i.e., localized disease) underscore the importance of identifying clinicopathological features capable of predicting tumor behavior. In this study, we have identified several prognostic features in MCC. Perhaps, these features may prove useful in identifying patients who require more aggressive treatment regimens.