Herbert HM, Sun MT, Selva D, Fernando B, Saleh GM, Beaconsfield M, Collin R, Uddin J, Meligonis G, Leatherbarrow B, Ataullah S, Irion L, Mclean CJ, Huilgol SC, Davis G, Sullivan TJ. Merkel Cell Carcinoma of the EyelidManagement and Prognosis. JAMA Ophthalmol. 2014;132(2):197-204. doi:10.1001/jamaophthalmol.2013.6077
The literature on Merkel cell carcinoma (MCC) of the eyelid remains scarce, and there has yet to be a study using the most up-to-date TNM staging system for this rare but aggressive tumor.
To analyze the TNM stage, management, and outcomes of patients with MCC of the eyelid.
Design, Setting, and Participants
Retrospective case series of 21 patients from 5 tertiary referral centers in the United Kingdom and Australia with primary MCC of the eyelid presenting at a median age of 77 years, with median follow-up of 54 months. Tumors were staged according to the American Joint Committee on Cancer, 7th edition, TNM criteria for eyelid carcinoma and MCC.
Main Outcomes and Measures
TNM stage, treatment modalities, and clinical outcome.
The eyelid carcinoma TNM stages were T2aN0M0 for 5 patients, T2bN0M0 for 7 patients, T3aN0M0 for 4 patients, T3bN0M0 for 3 patients, T2bN1M0 for 1 patient, and T3aN1M0 for 1 patient. The MCC TNM stages were T1N0M0 for 12 patients, T2N0M0 for 7 patients, T1N1M0 for 1 patient, and T2N1M0 for 1 patient. One patient had a sentinel lymph node biopsy, and 8 patients underwent head/neck imaging. Eighteen patients underwent a wide local excision, 12 with a paraffin section and 6 with a frozen section. Two patients underwent Mohs surgery, 1 of whom required an orbital exenteration. Twelve patients (57%) received adjuvant radiotherapy, and 2 patients received chemotherapy. The local recurrence rate was 10%, the regional nodal recurrence rate was 10%, and the distant metastatic recurrence rate was 19%. The lowest T category tumor metastasizing to both regional nodes and distant locations was a T2a (eyelid TNM)/T1 (Merkel TNM) tumor measuring 8 mm. Two patients with T3a (eyelid TNM)/T2 (Merkel TNM) tumors died of metastatic MCC.
Conclusions and Relevance
The majority of patients with MCC of the eyelid present with localized eyelid disease of T category T2 (eyelid TNM)/T1 (Merkel TNM). A wide local excision with margin control remains the mainstay of treatment, whereas the use of radiotherapy is institution specific. Tumors with a low T category are associated with regional nodal and distant metastatic disease. It may therefore be reasonable to consider a sentinel lymph node biopsy or strict regional lymph node surveillance for all MCCs of the eyelid, regardless of T category or size.
Merkel cell carcinoma (MCC) is an aggressive skin tumor that was first described in 1972 and that has since been increasing in incidence.1,2 Established risk factors for MCC include immunosuppression and extensive sun exposure, and Merkel cell polyomavirus has recently been identified in 80% of MCCs.3- 5 The estimated mortality rate for all patients with MCC is 25% to 35%.6 Involvement of the eyelid occurs in 10% of cases, with an annual incidence of approximately 0.23 cases per 100 000 persons.7,8 There are no randomized clinical treatment trials determining the efficacy of treatment, and management is stage dependent.
Historically, a number of staging systems were used for patients with MCC.9,10 These have since been consolidated into the 2010 TNM staging system supported by both the American Joint Committee on Cancer (AJCC) and the International Union for Cancer Control (Table 1).11,12 In addition, in recognition of the fairly unique structure of the eyelid, the AJCC TNM staging for eyelid carcinoma, 7th edition, was established to provide a standardized approach to staging disease in this region (Table 2). However, to the best of our knowledge, there are no studies of MCC of the eyelid that have used either staging system. Given the few published series available to help guide us in the management of MCC of the eyelid, we present our experience with 21 cases of primary MCC of the eyelid using both the eyelid and MCC AJCC TNM staging systems.
We conducted a multicenter, noncomparative case review of all patients with MCC presenting to 5 oculoplastic units within the United Kingdom and Australia. Consecutive patients from Manchester Royal Eye Hospital, Royal Adelaide Hospital, Moorfields Eye Hospital in London, Royal Surrey County Hospital, and Royal Brisbane Hospital who presented between 1991 and 2012 with a histological diagnosis of MCC were included in our study. Of the 21 cases in this series, only 1 was reported previously.13
Clinical case notes were reviewed to obtain data, including patient demographics, duration of signs and symptoms, clinical presentation and staging, site of the tumor, imaging findings, histopathological appearances, treatment modalities, complications, and outcome. Each case was staged according to initial clinical presentation using AJCC TNM staging for both eyelid carcinoma and MCC, 7th edition. Institutional review board/ethics committee approval was obtained for our study from all centers. All patient data were de-identified.
There were 21 cases of MCC identified over the 21-year period. Of these cases, 7 were from the Manchester Royal Eye Hospital, 5 were from the Royal Adelaide Hospital, 5 were from Moorfields Eye Hospital, 3 were from the Royal Surrey County Hospital, and 1 was from the Royal Brisbane Hospital. The median age at presentation was 77 years (range, 58-91 years), and 52% (11 of 21 patients) were women. All patients were of European descent, with the exception of one who was of East African and Asian descent. The median duration of symptoms prior to presentation was 12 weeks (range, 4-39 weeks). The median follow-up period was 54 months (range, 4-252 months).
There were 2 patients who had associated hematological disease. One had a history of chronic lymphocytic leukemia, and the other received a diagnosis of multiple myeloma 10 months following the diagnosis of MCC.
The upper lid was involved in 17 cases (81%). Imaging was used for 10 patients (48%) to evaluate the extent of disease. Four patients underwent magnetic resonance imaging of the neck, 4 patients underwent computed tomography of the neck, and 2 patients underwent chest radiography. A sentinel lymph node (SLN) biopsy was performed in 1 patient, whereas 2 patients presenting with clinical lymphadenopathy underwent a nodal biopsy. The TNM stages for our patients at presentation are summarized in Table 3. Regional nodal metastases were present in 2 patients at presentation who had clinical lymphadenopathy, whereas the remaining 18 patients (86%) had disease localized to the eyelid.
Table 4 summarizes the management of MCC for our patients. Eighteen patients (86%) underwent a wide local excision (5-mm margins) with margin control. Of these, 12 specimens were processed with “fast paraffin” techniques, and 6 with frozen section. Two patients underwent a Mohs excision. One patient had a primary exenteration after invasion of the medial rectus detected by magnetic resonance imaging. Two patients initially treated with a wide local excision required exenteration owing to an incomplete primary excision with orbital involvement.
Twelve patients (57%) received adjuvant radiotherapy for the eyelids, regional nodes, and intervening tissue following a wide local excision. Dosages ranged from 55 to 59.4 Gy. Both patients with regional nodal disease were treated with adjunctive radiotherapy regardless of the size of their MCCs. One patient who was treated initially with an orbital exenteration also underwent radiotherapy. For the remaining 9 patients with disease localized to the eyelid, the use of adjuvant radiotherapy was center specific (Table 4). One patient underwent an evisceration for a painful blind eye following radiotherapy. Of the remaining 9 patients who did not receive adjuvant treatment, 1 developed systemic metastases 6 months following initial treatment (Table 5).
Two patients were recommended to undergo a functional neck dissection. One patient with a 15-mm T2bN0M0/T1N0M0 tumor underwent this procedure with a superficial parotidectomy after being found to have regional nodal metastases to the preauricular and parotid nodes 1 week after surgery. However, the second patient who had presented with a 15-mm T2bN1M0/T1N1M0 tumor involving the submandibular nodes refused this treatment. This patient was 1 of 2 who were treated with adjuvant chemotherapy. The second patient receiving chemotherapy presented initially with an 8-mm T2aN0M0/T1N0M0 tumor and developed thoracic metastasis at 6 months follow-up. There were no reported adverse events associated with chemotherapy.14
Nine patients (43%) had tumors greater than 2 cm in size, classifying as stage II disease for both eyelid carcinoma and MCC TNM staging. Of these 9 patients, 6 were treated with adjunctive radiotherapy, whereas 3 required an exenteration. One patient was recommended to undergo orbital exenteration but refused this treatment. There were 3 patients with tumors greater than 2 cm in size who developed locoregional or distant metastatic recurrences (Table 4), whereas the remaining 6 patients were free from recurrences at a median follow-up of 75 months. Of the remaining 12 patients (57%) with a tumor size of less than 2 cm, 1 with a 15-mm tumor presented with regional nodal involvement, whereas another with a 15-mm tumor was found to have regional nodal disease 1 week after surgery. Six patients were treated with adjunctive local radiotherapy. One patient with an 8-mm primary tumor who did not receive radiotherapy developed both regional nodal and distant metastatic recurrent disease 6 months after surgery. None of the remaining patients with a primary tumor size of less than 2 cm experienced recurrence of their MCC at last follow-up (median, 36 months).
Table 5 describes features of recurrent MCC in our patients. Locoregional recurrences occurred in 4 patients and included 2 local recurrences (10%) and 2 regional nodal recurrences (10%). All initial recurrences were within an 18-month period. Neither patient with regional nodal recurrences was investigated at presentation with a SLN biopsy or neck imaging. There were 4 patients (19%) who developed systemic metastatic recurrences.
Two patients (10%) with stage T3a (eyelid TNM)/T2 (Merkel TNM) tumors died of metastatic MCC, and 6 patients (29%) died during follow-up of other causes. The remaining 13 patients (62%) are alive and without evidence of disease at last median follow-up of 54 months.
Our multicenter study represents, to the best of our knowledge, the largest case series of primary MCCs of the eyelid to date, with a median follow-up period of 54 months. Our series is also the first to use the AJCC TNM, 7th edition, staging system for eyelid carcinoma and MCC. Although the primary treatment for all our cases was a wide local excision with margin control, the use of prophylactic radiotherapy to prevent recurrent disease was institution specific. The smallest tumor associated with both regional nodal and distant metastatic disease was of T category T2a (eyelid TNM)/T1 (Merkel TNM). We demonstrated lower rates of locoregional and distant metastatic recurrence compared with MCCs in other body locations. In addition, we found a disease-specific mortality rate of 10%, which is also significantly lower than the reported rates for MCCs in general.
Data specific to MCC of the eyelid is scarce, and, to date, a total of 97 cases of primary MCC of the eyelid have been reported in the literature (Table 6). Among these 97 cases, there was a local recurrence rate of 14%, a regional nodal recurrence rate of 20%, and a distant metastatic recurrence rate of 5%. Only 2 patients (2%) with primary MCC of the eyelid had regional nodal or distant metastatic involvement at presentation, and the disease-specific mortality rate was 6%. In our study, we report lower rates of local and regional nodal recurrence, 10% for both, but found a higher rate of metastatic recurrence at 19%. In addition, we had 2 patients (10%) presenting initially with clinically evident regional nodal disease and a higher disease-specific mortality rate of 10%. However, these figures remain lower than that reported for MCCs in general, suggesting that MCCs of the eyelid carry a better prognosis. This may be related to earlier detection of eyelid lesions, with 52% of our patients presenting with primary lesions less than 2 cm in size (stage I).
Primary treatment of MCC in our series consisted of a wide excision with margin control or Mohs micrographic surgery with or without adjunctive radiotherapy. A wide local excision with 2.5- to 3-cm margins is generally recommended for cutaneous MCC.23,40,41 However, Peters et al27 demonstrated effective local control with 5-mm margins for MCC of the eyelid, in keeping with management guidelines for other aggressive malignant eyelid lesions. Although a Mohs excision has been used to treat MCCs affecting other cutaneous sites, there are only 3 reports27,31,38 of Mohs surgery for MCC of the eyelid, and one of these involved a recurrent MCC in which the initial surgical management was excision. The size of the surgical margin has not been shown to correlate with the locoregional recurrence rate because recurrences have been reported even after a wide excision with negative margins.42,43 In our series, 3 of 4 patients who developed recurrences underwent a wide local excision with margin control as their initial surgical management.
The role of radiotherapy in the management of MCC remains controversial. Although a number of studies have shown improvement in locoregional control,23,43,44 the evidence for improvement in disease-specific mortality is inconclusive.43 Over half of our patients (57%) received adjuvant radiotherapy following surgery. However, of the 4 patients who developed locoregional or distant metastatic recurrences, 3 had received radiotherapy. In addition, both patients who died of metastatic MCC had received adjuvant radiotherapy after surgery. Although there may be a role for the use of adjuvant radiotherapy to prevent recurrence in localized disease, further studies are required to determine patient selection for this treatment modality.
The AJCC staging system for MCC was first introduced in 2010 to standardize staging and improve comparability with staging systems for other skin cancers.11 In addition, the AJCC updated their staging for eyelid carcinoma in the 7th edition to improve its practical application.45 Because the AJCC eyelid TNM stages allow for finer discrimination of tumor size and extent to reflect the unique structure of the eyelid, this may be preferred over the Merkel TNM staging system. However, additional studies using the 2 staging systems are required to further investigate their practical application. Stage II disease for the AJCC TNM eyelid carcinoma and MCC staging system refers to localized disease where primary tumor size exceeds 2 cm. Previously established poor prognostic factors include tumors greater than 2 cm in size, being older than 65 years of age, being a male patient, locoregional recurrence, and distant metastases.27,44,46 In keeping with previous reports, both MCC-related deaths within our series occurred in patients with a primary tumor greater than 2 cm in size and distant metastatic disease. Of these 2 patients, 1 was a man presenting with stage III disease, whereas the other was a woman presenting with stage II disease. We also had 1 patient with eyelid T category T2a and Merkel T category T1 measuring 8 mm who subsequently developed regional nodal and distant metastatic recurrences. This patient died of an unrelated disease, but this highlights the importance of careful surveillance for MCC regardless of stage and size.
A pathologic nodal evaluation has been shown to improve prognostic accuracy, and previous studies have suggested that SLN mapping and excision may be a useful adjunct in the treatment of MCC.12,47 However, there is conflicting evidence regarding the association of SLN status and disease-specific survival.42,48 Despite this, an SLN biopsy has been recommended for all MCCs of the eyelid and conjunctiva, although further prospective studies have been advocated.49 Only 1 patient in our series underwent an SLN biopsy, with head and neck imaging being the more common form of regional disease assessment. None of the 4 patients who developed locoregional nodal and/or distant metastatic recurrences were investigated initially with an SLN biopsy. Furthermore, the smallest tumor in our series associated with regional nodal and distant metastatic disease was an 8-mm T2a (eyelid TNM)/T1 (Merkel TNM) tumor. This is in keeping with 3 previous reports of MCC of the eyelid in the literature,27,31,38 in which small tumors (≤10 mm) localized to the eyelid have been associated with regional nodal recurrences.14,27 This suggests that a nodal evaluation via an SLN biopsy or imaging followed by strict nodal surveillance may be appropriate for all patients with MCC of the eyelid. There has also been increasing evidence to support the use of positron emission tomography in the staging of MCC, with one retrospective study50 demonstrating a change in staging for 33% of their patients, and a change in management for 43% of their patients following a positron emission tomographic scan. Although the most appropriate follow-up method and frequency for patients with MCC have not been studied prospectively, most studies recommended regular examinations of the local disease site and the regional node, clinically and with imaging if indicated.14,51 It should be noted, however, that there is no evidence to suggest that early detection of regional recurrence improves survival. For high-risk cases (higher T category or nodal involvement), consideration could be given to regular (2- to 6-month) nodal surveillance with ultrasonography for at least the first 3 years. Reports of nodal surveillance with ultrasonography for head and neck short course chemotherapy suggest that this modality, in the hands of an experienced operator, has good sensitivity and specificity compared with computed tomography/magnetic resonance imaging.52
Two patients in our series presented with regional nodal disease, and a functional neck dissection was recommended for both; however, 1 patient refused this treatment and was instead treated with chemoradiotherapy. This patient was 1 of the 2 patients who received chemotherapy, a treatment that is generally reserved for distant metastatic disease for palliation rather than cure.23,27 Voog et al53 found a 60% response rate to chemotherapy in patients with MCC, in which the median overall survival after starting chemotherapy was 9 months for patients with distant metastatic disease and 24 months for patients with locoregional disease. In our series, 1 patient with regional nodal and distant metastatic recurrences was successfully treated with chemotherapy and died of unrelated causes 75 months after treatment. The second patient is also without recurrent disease at 12 months, although continued follow-up and additional studies are required to better define the role of chemotherapy for MCC.
In a recent review of the literature, Missotten et al34 identified 3 previous reports of patients with chronic lymphatic leukemia (CLL) and MCC of the eyelid. Our study adds an additional patient to this group. This patient had established CLL and was receiving immunosuppressive therapy prior to developing MCC. He presented with tumor invasion into the medial rectus, which required an orbital exenteration, and subsequently developed metastatic MCC but died 5 months after diagnosis of complications of CLL. Recent studies have demonstrated that patients with CLL have a high risk of developing MCCs containing Merkel cell polyomavirus owing to immunosuppression related to CLL and viral infection.54 In addition, patients with CLL who subsequently develop MCC have been shown to have significantly worse overall and cause-specific survival compared with patients without CLL.55 Our patient with CLL developed systemic metastases within the shortest time from initial diagnosis, compared with all other patients with systemic metastatic recurrences, and survived the shortest time after development of metastases.
Our study has a number of limitations that warrant mentioning. Our multicenter, retrospective study design included centers with differing treatment protocols, particularly with respect to the use of prophylactic radiotherapy and nodal assessment, thus making the correlation between T category in the TNM staging system and outcome difficult. This, however, reflects the lack of standardized treatment approaches to MCC of the eyelid, owing to its rarity, and encourages future research to use standard reporting with TNM staging. This would enable us to better estimate the prognosis for individual patients and would also enhance our ability to determine which patients may benefit from lymph node surveillance or an SLN biopsy. In addition, previous reports have documented that 90% of recurrences tend to occur within the first 2 years after primary diagnosis,10,42 and although our median follow-up period was 54 months, further recurrences may occur beyond this period. Additional follow-up is required to assess the recurrence rate in our patients in order to assess the efficacy of treatment.
Merkel cell carcinoma of the eyelid is associated with significant rates of recurrence and metastases, as well as a significant mortality rate. Compared with MCCs occurring in other locations, MCCs of the eyelid appear to be associated with a better prognosis, which may be related to earlier detection. A wide local excision with 5-mm margins and histological margin control remains the mainstay of treatment, and adjuvant radiotherapy may play a role in prevention of recurrent disease for selected patients. However, recurrent disease and regional/distant metastases do occur in patients who have disease localized to the eyelid or a lower T category tumor and in those who receive prophylactic radiotherapy. Hence, an SLN biopsy or a radiological nodal assessment should be considered for all patients with MCC of the eyelid as part of the initial staging process, and strict nodal surveillance may be useful in detecting early recurrent disease. TNM staging should be used to plan the management of patients and also should be adopted in future studies to better enhance comparison between studies. Because the TNM staging for eyelid carcinoma enables a finer discrimination of tumor size and extent, we would favor its use over the Merkel staging system.
Submitted for Publication: April 13, 2013; final revision received June 18, 2013; accepted June 18, 2013.
Corresponding Author: Michelle T. Sun, South Australian Institute of Ophthalmology, Royal Adelaide Hospital, University of Adelaide, Level 8, East Wing, Adelaide, South Australia, Australia 5000 (firstname.lastname@example.org).
Published Online: November 28, 2013. doi:10.1001/jamaophthalmol.2013.6077.
Author Contributions: Miss Sun had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Herbert, Selva, Saleh, Sullivan.
Acquisition of data: Herbert, Sun, Fernando, Beaconsfield, Collin, Uddin, Meligonis, Leatherbarrow, Ataullah, Irion, Mclean, Huilgol, Davis, Sullivan.
Analysis and interpretation of data: Herbert, Sun, Selva, Beaconsfield, Irion, Sullivan.
Drafting of the manuscript: Herbert, Sun, Fernando, Saleh, Collin, Meligonis, Ataullah, Irion, Sullivan.
Critical revision of the manuscript for important intellectual content: Herbert, Sun, Selva, Beaconsfield, Uddin, Leatherbarrow, Mclean, Huilgol, Davis, Sullivan.
Statistical analysis: Herbert, Sun.
Administrative, technical, or material support: Fernando, Saleh, Irion.
Study supervision: Selva, Beaconsfield, Collin, Uddin, Meligonis, Leatherbarrow, Ataullah, Mclean, Huilgol, Davis, Sullivan.
Conflict of Interest Disclosures: None reported.