Clinical characteristics of Merkel cell carcinoma at diagnosis in 195 patients: the “AEIOU” features

Introduction

Merkel cell carcinoma (MCC) is a highly aggressive skin cancer with a mortality of approximately 33% at 3 years¹, higher than that of melanoma (approximately 15%). Data from Surveillance, Epidemiology, and End Results (SEER) show a three-fold increase in MCC from 0.15 to 0.44 per 100,000 annually from the years 1986 to 2001². This trend is continuing, and approximately 1000–1500 new cases will be diagnosed in the United States in 2007^{3, 4}. Several factors likely contribute to this including an aging population, increased aggregate sun exposure and a higher number of immune suppressed individuals. Furthermore, the advent of the immunohistochemical marker cytokeratin-20 (CK-20) improved recognition of this disease. In the era before widespread CK-20 immunohistochemistry, laborious electron microscopy was required to make an accurate MCC diagnosis. Indeed, 66% of MCC cases in one series would have been misdiagnosed (as metastatic small cell lung cancer, basal cell carcinoma, lymphoma or other metastatic carcinoma) if electron microscopy had not been performed demonstrating the characteristic “neurosecretory granules within cytoplasmic extensions”⁵.

Management of MCC is controversial. To date there have been no controlled therapeutic trials in this disease. In most cases, surgical excision with sentinel lymph node biopsy^{1, 6} followed by radiation^{7, 8} is indicated. Conventional adjuvant chemotherapy lacks evidence of survival benefit and may in fact be associated with poorer outcomes^{1, 9}. A consensus treatment algorithm has been developed by the National Comprehensive Cancer Network¹⁰.

MCC prognosis is highly associated with the extent of disease at presentation. Disease-specific survival rates for local disease are greater than 90%, falling to 52% with nodal involvement³. If distant metastatic disease is present, expected survival is typically less than 10% at three years¹. As delay in diagnosis could allow disease progression, early detection and clinician recognition of this disease may improve survival rates. At present, a detailed description of the clinical characteristics of MCC at the time of diagnosis has not been published. Specifically, a PubMed search of “Merkel cell carcinoma clinical features” (performed on 10/24/07) yielded 87 studies, none of which described the clinician’s presumptive diagnosis, the color or symptomatic nature of the lesion, or the time to biopsy after lesion appearance.

The purpose of this study is to identify key clinical features that may assist the clinician in recognizing this aggressive skin cancer at an earlier and potentially more curable point.

Patients and Methods

Institutional review board approval was obtained from each institution. Tumor registry data and prospective patient identification (beginning in 2003) were used to identify 195 patients from 3 medical centers in Boston (Dana Farber Cancer Institute, Brigham and Women’s Hospital, Massachusetts General Hospital) and 2 medical centers in Seattle (Seattle Cancer Care Alliance, and University of Washington Medical Center). The study included patients with a pathologic diagnosis of MCC between 1980 and 2007.

Patient characteristics, clinical features of the lesion (i.e. site, tenderness, color, growing time, diameter), stage at presentation, interval from appearance to biopsy and the clinician’s impression at the time of biopsy were reviewed. The initial clinical impression was that recorded by the physician in either the clinical notes or on the pathology requisition. Presumptive diagnoses were listed in 106 patients. Of these, 78 had a single diagnosis while 28 had multiple clinical impressions reported; each of the 141 impressions was considered independently. The clinical impressions were stratified into benign, malignant and indeterminate lesions.

Estimation of age-adjusted chronic lymphocytic leukemia (CLL) prevalence in the US was determined for the 50–69 and ≥70 year old age groups using the SEER¹¹ database (2004 data; 30,465, and 51,579 cases, respectively) and dividing this by the US Census Bureau¹² 2004 population estimate for those age groups (60,489,662, and 26,653,288 persons, respectively). For solid organ transplant patient prevalence, a specific query was submitted to UNOS¹³ (United Network for Organ Sharing) which provided data regarding all living solid organ transplant recipients engrafted between 10/1/1986 and 6/30/2006 who have not had a reported graft failure (223,307 cases). This total number was divided by the 2006 US Census Bureau’s¹² estimated total US population (299,398,484 persons).

Results

Discussion

This study of 195 patients was conducted to identify key features of MCC that may aid clinicians in recognizing when a biopsy may be warranted. The basic demographic profile of our cohort is similar to that described in other studies: mostly elderly, Caucasian and with a slight male predominance.

The anatomic distribution of the tumors seen in our study further supports sun-exposure as a risk factor for the development of MCC, consistent with prior studies. Agelli et al. used SEER Registry data to demonstrate a positive association between geographic UV-B radiation indices and age-adjusted MCC incidence among White patients in a variety of US cities³. In patients receiving PUVA (psoralen + UV-A) for psoriasis, Lunder and Stern reported the incidence of MCC to be approximately 100 times that expected in the general population¹⁴. In our series, 81% of cases presented on UV-exposed skin. While sun exposure is strongly associated with MCC, as in melanoma, MCC can arise in the absence of significant UV exposure. Importantly, 5% of patients had tumors on highly sun-protected sites (buttock or vulva), and 14% had tumors arise on partially protected sites (abdomen, thighs and hair-bearing scalp).

Profound immunosuppression also appears to be an important risk factor for MCC. Indeed, 7.8% of our patients had some form of immunosuppression, including HIV, CLL, or iatrogenic suppression secondary to solid organ transplantation. This frequency is a 16-fold over-representation of that expected in the general US population, in which the estimated prevalences are 0.4% for HIV¹⁵, 0.029% for CLL^{11, 12}, and 0.075%^{12, 13} for solid organ transplantation.

An association of MCC with HIV and solid organ transplantation is documented in the literature. Miller et al. reported a roughly ten-fold increase in MCC after solid organ transplantation¹⁶, and Engels found a 13-fold increase among HIV-positive patients¹⁷. Further highlighting the importance of immune function in MCC, Friedlaender et al. described regression of MCC metastases after discontinuation of cyclosporine in a kidney transplant patient¹⁸.

Among our immunosuppressed patients, CLL was particularly over-represented (4.1%). The age-adjusted incidence of CLL in our cohort was 2.4% for 50–69 year olds and and 6.5% for ≤ 70 year olds, representing a respective 48-fold and 34-fold increase above that expected in the US population for those age groups. MCC arising in the setting of CLL has been described^19–23, but to the best of our knowledge, this is the first quantitation of the degree to which MCC is over-represented in this disease. We acknowledge the potential for referral bias in our series because immune suppressed patients may be more likely to be seen in tertiary medical centers. However, in 3 of the 15 cases, MCC was diagnosed first, and the immune suppressed state (HIV or CLL) was discovered as part of the MCC workup. This finding also supports the need for practitioners to consider further workup for immunosuppression in patients presenting with MCC.

In agreement with a prior study of transplant patients, we observed more advanced disease at time of presentation in the immunosuppressed group; ten of the fifteen (67%) immunocompromised patients presented with either nodal or distant metastatic disease as compared to 42% in the immunocompetent group (difference not statistically significant). In contrast to reports of organ transplant recipients and HIV positive patients developing MCC at an earlier age^{17, 24, 25}, we did not find a difference in the mean ages of immunosuppressed and immunocompetent MCC patients. This likely reflects the inclusion of 8 patients with CLL, a disease with a mean age of 65.

Clinicians thought most lesions were benign prior to biopsy, which may have contributed to a delay in the diagnosis. Although not quantified, many of our patients reported they had been reassured by a physician about the benign nature of their lesion. Indeed, several of the characteristic MCC features were present at that earlier visit. It is hoped that publicizing the clinical characteristics of MCC may result in an earlier diagnosis in some cases.

We identified several tumor characteristics that may aid a clinician in suspecting MCC, summarized as AEIOU (Table V). Of particular note, we consider lack of tenderness an important feature in MCC as many other lesions that are rapidly growing and red or pink (such as an inflamed cyst, the most common presumed diagnosis) would likely be tender.

Here we describe the first systematic analysis of the clinical features of Merkel cell carcinoma. We have identified several characteristics that in combination are highly sensitive for the diagnosis of MCC. While we do not have data to address the specificity of these features, it is likely to be low given the rarity of MCC. Among lesions encountered in routine clinical practice that display multiple of these features, only a few may be MCC while others, such as squamous cell carcinoma or keratoacanthoma, would also have required a biopsy. Thus, the use of these criteria to aid in the decision to perform a biopsy may be appropriate. Given the correlation between survival and stage at presentation¹, identifying patients at an earlier and potentially more curable point is highly desirable for Merkel cell carcinoma.