J Cutan Pathol 2010: 37: 28–34 Copyright © 2009 John Wiley & Sons A/Sdoi: 10.1111/j.1600-0560.2009.01352.xJohn Wiley & Sons. Printed in Singapore Journal of

Cutaneous Pathology

Prevalence of MCPyV in Merkel cellcarcinoma and non-MCC tumorsBackground: Merkel cell polyomavirus (MCPyV) is the likelycausative agent of Merkel cell carcinoma (MCC). However, theprevalence of MCPyV in non-MCC population and its possible rolein the pathogenesis of other skin cancers are not known yet.Methods: A molecular pathology study was performed in 33 MCCsamples and 33 age- and sex-matched samples of sun exposed non-MCC tumors [12 seborrheic keratoses (SK), 11 basal cell carcinomas(BCC) and 10 lentigo maligna melanomas (LMM)]. All tumorswere analyzed for presence of MCPyV-DNA by polymerase chainreaction (PCR) and Southern-Blot hybridization of PCR products.Results: MCPyV sequences were detected in 21 MCC samples(64%) and in 2 non-MCC tumors of sun exposed skin (6%; bothSK-patients). Neither the tissue samples from BCC nor LMM provedpositive for MCPyV sequences.Conclusion: We were able to confirm prior data on prevalence ofMCPyV-DNA in MCC. Furthermore, a female predominance ofMCPyV-positive MCC-patients was detected. There was no relevantassociation of MCPyV with SK, BCC and LMM. Speculative,prevalence of MCPyV in an age- and sex-matched non-MCCpopulation could average up to 6%.

Andres C, Belloni B, Puchta U, Sander CA, Flaig MJ. Is Merkel cellpolyomavirus also prevalent in non-Merkel cell carcinoma (MCC)tumors of sun exposed skin? A study of 66 patients.J Cutan Pathol 2010; 37: 28–34. © 2009 John Wiley & Sons A/S.

Christian Andres1, BenedettaBelloni2, Ursula Puchta1,Christian A. Sander3 and MichaelJ. Flaig1

1Department of Dermatology and Allergy,Ludwig-Maximilians-Universitat Munchen,Munich, Germany2Department of Dermatology and AllergyBiederstein, Technische Universitat Munchen,Munich, Germany, and3Department of Dermatology and Allergy,ASKLEPIOS Eduard-Arning hospital Hamburg,Germany

Dr. med. Michael J. Flaig, Klinik und Poliklinik furDermatologie und Allergologie derLudwig-Maximilians-Universitat Munchen, Frauenlobstr.9-11, D-80337 Munchen, Munich, GermanyTel: +49 (0)89 5160 6327/6181Fax: +49 (0)89 5160 6182e-mail: Michael.Flaig@med.uni-muenchen.de

Accepted for publication March 9, 2009

In humans persistent virus infections with one ormore viruses like herpesviruses, papillomavirusesor polyomaviruses are highly frequent and widelyspread.1 Normally these latent infections are notcausative to malignant neoplasms, but some of them,such as a subset of human papillomaviruses (HPV)or human T-lymphotropic virus type 1 (HTLV-1),may have serious consequences in terms of malignanttransformation of their host cells.1– 3

Polyomaviruses are double-stranded DNA-based,small viruses of commonly non-oncogenic characterfor their host, but maybe oncogenic for some speciesunder certain conditions.2 The first polyomavirus(murine polyomavirus, MuPyV) was discovered in1953 by Ludwik Gross,4 followed by Simian Virus40 (SV40) in 1960 by Bernice Eddy, Ben Sweetand Maurice Hilleman.5,6 A reasonable number of

polyomaviruses has been identified in the recent past,but only five human pathogenic members of thisfamily have been characterized until now.7 In 1971two human pathogenic polyomaviruses–BK and JCvirus–were identified, followed by KI and WU duringthe last 3 years.8 In January 2008, the latest species,Merkel cell polyomavirus (MCPyV), was decipheredby Feng et al. and described as the likely causativeagent of Merkel cell carcinoma (MCC).9

MCC is a very aggressive, primary skin cancer ofneuroendocrine differentiation which occurs mainlyin older, often immunosuppressed male Caucasiansin sun-exposed areas. The 2-year survival rate is50–70%. Incidence is growing steadily, with numbershaving tripled in the last 2 decades (0.44 per100,000 in 2001 in the USA).10 MCC is namedafter the putative cell of origin. The Merkel cell

28

Prevalence of MCPyV in Merkel cell carcinoma (MCC) and non-MCC tumors

was first described by Sigmund Merkel in 1875 astastzellen in the skin with putative mechanoreceptorfunction, which are found in the basal layer ofthe epidermis.11 It is not yet clarified, whether themalignant tumor originates from the Merkel cell itselfor from a pluripotent dermal stem cell.12 The tumorsare composed of small basophilic, uniformly sized,highly mitotic malignant cells having round to ovalnuclei and scanty cytoplasm. Rarity of epidermalinvolvement of the usual solid, dermal located tumornodules supports the dermal stem cell-assumption,while the presence of dense-core granules and positivestaining for anti-cytokeratin 20 and neurofilamentsmakes the Merkel cell itself supposable as the cell oftumor origin.10,12

Feng et al. detected integration of viral DNAwithin the tumor genome in a clonal pattern inMCPyV-positive MCC, suggesting that MCPyV-infection and integration precedes clonal expansionof the tumor cells.9 Meanwhile several studiesconfirmed these data.8,13,14 Geographical differencesof MCPyV-prevalence in MCC have been published.Becker et al. showed presence of MCPyV in 45of 53 (84.9%) European-MCC-patients,13 whileGarneski and colleagues observed presence ofMCPyV in nearly 70% of North American MCC(11 of 16 specimen) and only 24% in AustralianMCC (5 of 21 specimen).15 They speculate thatowing to increased sun exposure in Australiaviral pathogenesis could be a less importantcofactor.15 Recently Kassem and colleagues observeda very interesting correlation between MCPyV-prevalence and immunosuppresion. They detecteda significantly higher frequency of MCPyV innon-melanoma-skin-cancer (Morbus Bowen andBCC) of immunosuppressed patients compared toimmunocompetent patients.16

However, the method of infection with MCPyVand the prevalence of MCPyV in non-MCCpopulation and its possible role in the pathogenesisof other skin cancers are not profoundly knownyet. In the present study, we assigned every MCC(n = 33) to a different neoplasm of a sex and age-matched patient being either a seborrheic keratosis(SK, benign skin tumor, n = 12), basal cell carcinoma(BCC, ‘‘semimalignant’’ skin tumor, n = 11) orlentigo maligna melanoma (LMM, malignant skintumor, n = 10). Knowing well about the risk factorsof MCC, all matched tumors were located in sun-exposed skin. As MCC is seen primarily in the elderlyCaucasian population, we matched every MCC-patient to a same-aged and same-gender Caucasianpatient. Subsequently we investigated all tumors forpresence of specific MCPyV-DNA sequences in order

to gain insight into the prevalence and the possiblereservoir of this polyomavirus.

Materials and methodsPatients’ characteristics and tissue specimen handlingThe study was carried out using formalin-fixedand paraffin-embedded (FFPE) tumor tissue from33 patients with MCC, 12 patients with SK, 11patients with BCC and 10 patients with LMM.Samples have been selected from the archivesof the Department of Dermatology and Allergy,ASKLEPIOS Eduard-Arning hospital Hamburg,Germany, from the Department of Dermatologyand Allergy Biederstein, Technische UniversitatMunchen, Munich, Germany and in the mainfrom the Department of Dermatology and Allergyof the Ludwig-Maximilians-Universitat Munchen,Munich, Germany. Details of clinicopathologiccharacteristics are summarized in Tables 1 and2. Serial sections of all specimens were used forHematoxylin & Eosin (H&E) staining and DNApreparation.

DNA preparationH&E stains of the selected specimens were inde-pendently reviewed by two experienced pathologists(M.F. and C.A.) to select paraffin material contain-ing >95% tumor tissue. Three consecutive 10-μmparaffin sections from each specimen were subjectedto DNA extraction. After deparaffinization, all tissueswere lysed by proteinase K overnight (55◦C) untilcomplete tissue lysis.

MCPyV detection by polymerase chain reaction (PCR)PCR was performed with genomic DNA using 45cycles for each primer set with the Hot-start BlueTaq(Euroclone) polymerase in a final volume of 30 μl.DNA quality was confirmed by ß-globin PCR usingthe GH20 (5′-GAAGAGCCAAGGACAGGTAC-3′)and PCO4 (5′-CAACTTCATCCACGTTCACC-3)primer set. For MCPyV detection we used two setsof primers resulting in a 138 base-pair product(MCV138 forward: 5′-GGTTAGAGATGCTGGA-AATGACC-3′; reverse: 5′-CAAATAAGCAGCAG-TACCAGGC-3′) and a 191 base-pair product(MCV191 forward: 5′-CCACTTTATTATCTT-AGCCCAT-3′; reverse: 5′-TCCTTTTGGCTAG-AACAGTGTC-3′) respectively, targeting the largeand small T-antigen region (isolate MCC-Mpt-LS2of MCPyV). The exact location for the MCV138-primer is forward 2185-2207 and reverse 2301-2322and for the MCV191-primer forward 1994-2015 andreverse 2163-2184 regarding reference sequences of

29

Andres et al.

Table 1. Characteristics of Merkel cell carcinoma (MCC) patients, results of Merkel cell polyomavirus (MCPyV) PCR and confirmation bySouthern-Blot analysis. MCPyV-positive patients marked bold and cursive

MCV 138 MCV 191

MCC- Patient Sex Age PCR Southern-Blot PCR Southern-Blot

1 W 59 + + + +2 M 65 − − + +3 W 86 + + + +4 W 72 + + + +5 M 44 − − − −6 M 64 − − − −7 W 64 + + + +8 W 66 − − − −9 W 67 + + − −10 W 75 + + + +11 M 88 − − − −12 M 89 − − + +13 M 78 + + + +14 W 79 − − −15 W 67 + + − −16 W 90 − − + +17 M 74 + + − −18 W 73 + + + +19 W 83 + + + +20 M 68 − + − +21 M 82 + + + +22 M 85 − − − −23 W 81 + + + +24 M 87 + + + +25 W 88 − − − −26 W 81 + + + +27 M 97 + − − −28 M 68 − − − −29 W 83 + + + +30 M 57 − − − −31 W 80 + + + +32 W 82 − − − −33 W 91 − − − −Pos/tot 18/33 18/33 17/33 18/33% 54.5% 54.5% 51.5% 54.5%M (n = 14), F (n = 19) Total pos: 21/33 (63.6%)

M, male; F, female; PCR, polymerase chain reaction; pos, MCPyV-positive; tot, total.

the National Center for Biotechnology Information(NCBI) Entrez Nucleotide database gb|EU375803.1MCPyV isolate MCC350, using the NCBI Blastprogram. For PCR-negative controls we used waterinstead of DNA, containing all PCR componentsexcept DNA.

Southern-BlottingTo increase sensitivity and specificity we per-formed a Southern-Blot analysis using the PCRproducts. These were transferred to positivelycharged nylon membranes after electrophoresis ina 1.5% agarose gel and then hybridized withthe specific probes MCV138 and MCV191, whichcarry a 5′digoxigenin label (DIG-MCV138 forward:

5′-GTAAGAAGTTTAAGAAGCACCTAG-3′ andDIG-MCV191 forward: 5′-GATCTCGCCTCAAA-CCTCACAAG-3′) omitting the primer sequences,using the Roti-Hybri-Quick solution (Carl RothGmbH & CoKG), Anti-Digoxigenin-AP, Fab frag-ments (Roche Diagnostics, Mannheim, Germany)and CDP-Star (Roche Diagnostics, Mannheim,Germany), accordingly to the manufacturer’sinstructions.

ResultsMCPyV-DNA was detected in 21 of 33 (64%) tissuespecimens in the MCC cohort [18/33 (55%) forMCV138 and 18/33 (55%) for MCV191; positivefor both primer sets 15/33 (46%)]. Of the 33

30

Prevalence of MCPyV in Merkel cell carcinoma (MCC) and non-MCC tumors

Table 2. Characteristics of 33 sex and age-matched patients with different neoplasms in sun-exposed anatomic region, such as seborrheickeratoses (SK, benign skin tumor), basal cell carcinoma (BCC, ‘‘semimalignant’’ skin tumor) and lentigo maligna melanoma (LMM, malignantskin tumor). Results of MCPyV-PCR and Southern-Blot analysis. MCPyV-positive patients marked bold and cursive

MCV 138 MCV 191

Tumor Sex Age PCR Southern-Blot PCR Southern-Blot

LMM-1 W 59 − − − −LMM-2 W 72 − − − −LMM-3 W 64 − − − −LMM-4 W 81 − − − −LMM-5 W 67 + − + −LMM-6 M 78 − − − −LMM-7 M 44 − − − −LMM-8 M 68 − − − −LMM-9 M 85 − − − −LMM-10 W 80 − − − −Pos/tot 1/10 0/10 1/10 0/10% 10% 0% 10% 0%BCC-1 W 86 − − − −BCC-2 M 64 − − − −BCC-3 M 74 − − − −BCC-4 W 88 − − − −BCC-5 W 81 − − − −BCC-6 M 97 + − − −BCC-7 W 82 − − − −BCC-8 W 66 − − − −BCC-9 W 79 − − − −BCC-10 M 82 + − − −BCC-11 W 91 + − − −Pos/tot 3/11 0/11 0/11 0/11% 27.3% 0% 0% 0%SK-1 M 65 − − − −SK-2 M 57 − − − −SK-3 W 75 − − − −SK-4 M 88 − + − −SK-5 W 83 − − − −SK-6 W 90 − − − −SK-7 W 73 − − − −SK-8 W 83 − − − −SK-9 M 68 − − − −SK-10 W 67 − − − −SK-11 M 89 + − − −SK-12 M 87 + + − −Pos/tot 2/12 2/12 0/12 0/12% 16.7% 16.7% 0% 0%

M, male; F, female; PCR, polymerase chain reaction; pos, MCPyV-positive; tot, total.

sex and age-matched non-MCC tumors of sunexposed skin, 2 were positive for MCPyV-DNAusing the MCV138 primer set. Interestingly bothsamples belonged to the SK-group. The PCR-negative controls were constantly negative in allexperiments. Prior to MCPyV analysis all 66specimens revealed amplifyable DNA with a ß-globinPCR. Clinicopathologic data are summarized inTables 1 and 2; 1.5% agarose gels and correspondingSouthern-Blot analysis are shown for MCV191 in

10 patients with MCC (Fig. 1) and for MCV138 in10 patients with LMM (Fig. 2).

58% (19/33) of the MCC-patients were female.In 14 of these 19 MCC tumors MCPyV-DNA wasdetected (74%), while only 7 of 14 male patients(50%) showed prevalence for MCPyV-DNA intumor tissue. The mean age of all MCC-patientswas 76 ± 11 years. The mean age of the femaleMCPyV-positive patients was 76 ± 9 years, while themean age of the male MCPyV-positive patients was78 ± 9 years.

31

Andres et al.

(A)

(B)

Fig. 1. A) 1.5% agarose gel for MCV191 in ten patients with MCCincluding positive (+), negative control (−) and size ladder HAEIII (φX174 RF DNA/Hae III Fragments) (L). B) CorrespondingSouthern-Blot.Note: Patients 7, 10, 26, 29 and 31 display specific PCR products,confirmed by Southern-Blot analysis.

DiscussionAlthough human polyomaviruses are known forhalf a century, until recently a pivotal role inhuman malignancies has not been shown yet.1

Feng et al. indentified a new human polyomavirusand coined it Merkel cell Polyomavirus (MCPyV),because of its strong association to MCC. MCCis a rare but highly aggressive skin cancer whichoccurs mainly in the elderly white population andimmunosuppressed individuals. These are facts, thatpoint toward a possible infectous etiology, alikeKaposi’s sarcoma and human herpesvirus 8 (HHV-8).17 Furthermore, Feng et al. showed monoclonalintegration of MCPyV-genome in 8 of 10 MCC.9

We found presence of MCPyV in 64% of MCCspecimens (21/33) and 6% in sun-exposed non-MCCtumors (2/33). PCR amplification with our primersets resulted in a 138 base-pair product and a 191base-pair product. Adapting to the situation of usingFFPE specimens primer sets were designed to producerelatively small PCR-products (138-bp and 191-bp)in order to optimize the detection rate MCPyV.18

Previous studies suggest a slightly higher prevalenceof MCPyV in MCC.8,9,13,14 Kassem A et al. detectedprevalence for MCPyV in 30 of 39 MCC. Theyinvestigated also FFPE samples using four different

(A)

(B)

Fig. 2. A) 1.5% agarose gel for MCV138 in ten patients with LMMincluding positive (+), negative control (−) and size ladder HAEIII (φX174 RF DNA/Hae III Fragments) (L). B) CorrespondingSouthern-Blot.Note: Lane 5 displays an unspecific PCR product, not confirmedby Southern-Blot analysis. All patients are negative for Merkel cellpolyomavirus.

primer sets, showing great heterogeneity in theirperformance.8 Their detection rates varied from11/39 by LT1 (440-bp), 14/39 by VP1 (351-bp) and15/39 by M1/2 (178-bp) up to 28/39 by LT3 (308-bp).8 Possible explanations for low detection ratesare differences in DNA integrity because of differentfixation and processing protocols, targeting only oneregion of the MCPyV (T-antigen-region) and useof two instead of four primer sets. Nevertheless,the primer sets used in our study (MCV138 andMCV191) proved to be equally effective for MCPyVdetection in MCC (both each 18/33). Kassem et al.speculated that differences in the detection frequencymight be caused by changes due to the viralintegration process–to us a plausible assumption. Wecould detect MCPyV-DNA in 15 of 21 samples withboth primer sets, while either MCV138 or MCV191was positive in a total of 6 cases (Patient 9, 15,17 for MCV138 and 2, 12, 16 for MCV191). In allpositive cases specificity was proven by Southern-Blotanalysis of the PCR products. Individual cases wereconsidered positive only, if positive PCR-result wasconfirmed by Southern-Blot (Fig. 1 and 2).

The prevalence of MCPyV showed no significantassociation to non-MCC skin tumors. Interestingly,

32

Prevalence of MCPyV in Merkel cell carcinoma (MCC) and non-MCC tumors

we detected MCPyV in 2 of 12 (17%) SK, while allLMM and BCC tissue specimens were negative forMCPyV. SK are considered benign skin tumors incontrast to the highly malignant MCC.19 LMM is asubtype of malignant melanoma that typically devel-ops on sun-damaged skin in elderly patients20 –twocommon risk factors with MCC. Nevertheless, all10 investigated LMM of our cohort were MCPyVnegative (Fig. 2). While no sufficient data exist in theliterature for MCPyV-prevalence in SK and LMMso far, BCC was investigated in this regard. Beckeret al. showed presence of MCPyV in 3 of 24 BCC13

and recently Kassem et al. showed a surprisingly highprevalence of MCPyV in 47 of 147 sporadic BCC(32%).16 Nevertheless we could not detect MCPyV-DNA in 12 BCC by PCR and Southern-Blot analysis.We speculate that the prevalence of MCPyV in non-MCC tumors (2/33; 6%) might display the dimensionof latent infection with MCPyV in the elderly whitepopulation irrespective of the underlying neoplasminvestigated. It is very likely that other risk factorssuch as immunosuppression, chronic sun damage orgenetic disposition are further factors for oncogenictransformation.

Several studies have shown a slight male predom-inance for MCC.21– 23 Nevertheless, 58% (19/33)of our MCC-population were female. Interestingly,we could show in 14 of these 19 female patientsprevalence of MCPyV (73.7%), while only 7 of 14male patients (50%) showed prevalence for MCPyV.Kassem et al. included 20 male and 19 female patientsin their investigations. They reported similar to ourresults a female predominance concerning prevalenceof MCPyV in MCC. In 84% (16/19) female MCC incontrast to 70% (14/20) male MCC MCPyV-DNAwas detectable.8 Nevertheless, more data are neededto verify this observation. Incidentally, both MCPyV-positive SK-patients were male. In other recentlypublished studies no information was given in termsof sex-distribution of non-MCC samples analysed.

In summary, our investigations underline a strongassociation between MCPyV and MCC. The analysisof a matched cohort of LMM, BCC and SK points toa possible MCPyV-prevalence in non-MCC patientsof 6%. Our non-MCC-cohort is so far too small topredict MCPyV-prevalence, but still our data suggest,that MCPyV infects humans in a low but relevantpercentage. We assume that MCPyV infection is oneamongst other risk factors, leading to a malignanttransformation of Merkel cells in human. This isthe first reported observation presenting data ofprevalence of MCPyV in a matched non-MCC-population. Profound studies are needed to elicitthe natural reservoir of MCPyV to get insight intothe transfection mode and the targeted cells in

the human organism. In consequence, methods ofprevention and possibly targeted therapy could beelucidated.

AcknowledgementsWe would like to thank Dres. M. Gummer and W. Maciejewski(Munchen) and Dres. D. Dienemann and J. Holzl (Munchen)for providing additional tissue samples. This study was in partsupported by the Dr. H. Legerlotz, the R. Bartling and the M. LackasFoundation.

References

1. Allander T, Andreasson K, Gupta S, et al. Identification of athird human polyomavirus. J Virol 2007; 81: 4130.

2. zur Hausen H. A specific signature of Merkel cell polyomaviruspersistence in human cancer cells. Proc Natl Acad Sci U S A2008; 105: 16063.

3. Flaig MJ, Rommel F, Campe H, Emmerich B, Sander CA. AdultT-cell lymphoma/leukemia: a very rare disease in Europe. DtschMed Wochenschr 2001; 126: 389.

4. Gross L. A filterable agent, recovered from Ak leukemic extracts,causing salivary gland carcinomas in C3H mice. Proc Soc ExpBiol Med 1953; 83: 414.

5. Eddy BE, ROWE WP, Hartley JW, Stewart SE, Huebner RJ.Hemagglutination with the SE polyoma virus. Virology 1958; 6:290.

6. Sweet BH, Hilleman MR. The vacuolating virus, S.V. 40. ProcSoc Exp Biol Med 1960; 105: 420.

7. zur Hausen H. Novel human polyomaviruses – re-emergence ofa well known virus family as possible human carcinogens. Int JCancer 2008; 123: 247.

8. Kassem A, Schopflin A, DIAZ C, et al. Frequent detection ofMerkel cell polyomavirus in human Merkel cell carcinomas andidentification of a unique deletion in the VP1 gene. Cancer Res2008; 68: 5009.

9. Feng H, Shuda M, Chang Y, Moore PS. Clonal integration ofa polyomavirus in human Merkel cell carcinoma. Science 2008;319: 1096.

10. Bichakjian CK, Lowe L, Lao CD, et al. Merkel cell carcinoma:critical review with guidelines for multidisciplinary management.Cancer 2007; 110: 1.

11. Merkel S. Tastzellen und Tastkorperchen bei den Haustierenund beim Menschen. Arch Mikrosk Anat 1875; 11: 636.

12. Ferringer T, Rogers HC, Metcalf JS. Merkel cell carcinomain situ. J Cutan Pathol 2005; 32: 162.

13. Becker JC, Houben R, Ugurel S, Trefzer U, Pfohler C, SchramaD. MC Polyomavirus is frequently present in Merkel cellcarcinoma of European patients. J Invest Dermatol 2009; 129:248.

14. Foulongne V, Kluger N, Dereure O, Brieu N, Guillot B, SegondyM. Merkel cell polyomavirus and Merkel cell carcinoma, France.Emerg Infect Dis 2008; 14: 1491.

15. Garneski KM, Warcola AH, Feng Q, Kiviat NB, Leonard JH,Nghiem P. Merkel Cell polyomavirus is more frequently presentin North American than Australian Merkel Cell carcinomatumors. J Invest Dermatol 2009; 129: 246.

16. Kassem A, Technau K, Kurz AK, et al. Merkel cell poly-omavirus sequences are frequently detected in nonmelanomaskin cancer of immunosuppressed patients. Int J Cancer 2009;Jul 15; 125(2): 356.

33

Andres et al.

17. Andres C, Hofmann H, Ring J. Kaposi’s sarcoma of the tongue.JEADV (2009) In press.

18. Srinivasan M, Sedmak D, Jewell S. Effect of fixatives and tissueprocessing on the content and integrity of nucleic acids. Am JPathol 2002; 161: 1961.

19. Hafner C, Vogt T. Seborrheic keratosis. J Dtsch Dermatol Ges2008; 6: 664.

20. Mckenna JK, Florell SR, Goldman GD, Bowen GM. Lentigomaligna/lentigo maligna melanoma: current state of diagnosisand treatment. Dermatol Surg 2006; 32: 493.

21. Allen PJ, Bowne WB, Jaques DP, Brennan MF, Busam K,Coit DG. Merkel cell carcinoma: prognosis and treatmentof patients from a single institution. J Clin Oncol 2005; 23:2300.

22. Hodgson NC. Merkel cell carcinoma: changing incidence trends.J Surg Oncol 2005; 89: 1.

23. Medina-Franco H, Urist MM, Fiveash J, Heslin MJ, Bland KI,Beenken SW. Multimodality treatment of Merkel cell carcinoma:case series and literature review of 1024 cases. Ann Surg Oncol2001; 8: 204.

34