Introduction

Merkel cell carcinoma (MCC) is a rare and aggressive primary neuroendocrine skin cancer. In the USA, about 2000 new cases are diagnosed annually. The 5-year survival rate for local and metastatic disease are 70–80 and 20–30%, respectively [1]. Risk factors for MCC include ultraviolet (UV) exposure, older age (> 50 years), and immunosuppression (e.g., chronic lymphocytic leukemia) [2].

The discovery of the Merkel cell polyomavirus (MCPyV) in 2008 prompted further investigation of the biology and pathogenesis of MCC. Furthermore, as we gain experience with treating MCC patients, protocols for diagnosis and management are being refined. In this article, we provide updates on the pathogenesis, diagnosis, and management of MCC.

Pathogenesis

Merkel Cell Polyomavirus-Positive MCC Pathogenesis

Approximately 80% of MCC tumors in the USA are virus positive (VP-MCC) and have clonal integration of MCPyV into the host genome [3]. Like other polyomaviruses, MCPyV is a small (5.4 kb), double-stranded, non-enveloped DNA virus in the Orthopolyomavirus genus of the Polyomaviridae family [3]. The early region of the MCPyV virome encodes putative oncogenes including large T (LT) antigen, small T (sT) antigen, 57kT antigen, and a protein called alternative LT ORF (ALTO) [3,4,5,6,7]. The late region encodes the capsid proteins, VP1 and VP2 [3,4,5,6]. The LT protein contains multiple domains including conserved region 1 (CR1), DnaJ domain, RB binding site, origin binding domain (OBD), and a Helicase domain [4,5,6]. The sT protein contains CR1, DnaJ, the LT-stabilization domain (LSD), and the PP2A binding domain [8]. VP-MCC carries a low somatic mutation burden, suggesting that tumorigenesis is driven by viral oncogenes [9•]. The integration of MCPyV into the host genome occurs early in tumorigenesis (Fig. 1a). Interestingly, VP-MCC tumor development requires a mutation or deletion that disrupts LT helicase activity [5, 10]. Despite the demonstrated importance of MCPyV in MCC oncogenesis, little is known about the normal host cell of the virus, what drives viral integration, and the cell of origin for VP-MCC tumors.

Fig. 1
figure 1

Pathogenesis and management of Merkel cell carcinoma (MCC). a Schematic model of virus-positive (VP) and virus-negative (VN) MCC pathogenesis illustrating targets of viral small T antigen (sT) and large T antigen (LT) in VP-MCC tumors, as well as mutations impacting Rb and p53 function commonly found in VN-MCC. Antitumor T cells are disinhibited by targeting PD-1 or PD-L1 with monoclonal antibody immune checkpoint inhibitors in the treatment of MCC. TCR T cell receptor, MHC major histocompatibility complex, MCPyV Merkel cell polyomavirus. b Management flowchart for MCC. 1Excision and/or radiation to primary tumor should not occur prior to SLNB [2].Adjuvant radiation to the primary tumor bed is generally recommended. Consider observation without adjuvant radiation if the primary tumor is small (<1 cm), widely excised, and without high-risk features [3].Palliative option; recurrence is not uncommon [4].PET-CT is the preferred modality. MRI and CT are alternatives [5].Anti-PD-1/PD-L1 immune checkpoint inhibitors or referral to a clinical trial is preferred over chemotherapy for metastatic disease. Avelumab is FDA-approved for treating metastatic MCC [6].Consider radiation to nodal basin in high-risk patients (e.g., immunosuppression) or if false-negative SLNB is suspected [7].Consider routine imaging in high-risk patients [8].Serology testing is only useful in patents with detectable baseline titers assessed soon after treatment. SLNB sentinel lymph node biopsy, FNA fine needle aspiration, CLND complete lymph node dissection

Full size image

sT Antigen

The primary MCPyV oncogenes are thought to be the sT and LT antigens. The sT antigen is expressed from two of four alternative spliced mRNA of the MCPyV virome [11• ]. While the MCPyV LT and sT share exon 1 of the T antigen locus, the DnaJ, CR1, and Hsc70 domains found in exon 1 appear to be dispensable for sT-driven tumorigenesis [8]. In most polyomaviruses, sT antigen binds protein phosphatase 2A (PP2A) to promote AKT activation and thereby increases cell survival [12]. Like other polyomaviruses, the sT antigen of MCPyV binds PP2A, but unlike other polyomaviruses, this binding has no observed effect on in vitro or in vivo tumorigenesis [12]. Interestingly, MCPyV sT binds protein phosphatase 4C (PP4C) and protein phosphatase 4 regulatory subunit 1 (PP4R1) to form a complex targeting the NF-κB regulator NEMO, leading to reduced NF-κB translocation and transcriptional activity [13• ]. Regulation of PP4C and PP4R1 by the sT could be a mechanism by which MCPyV modulates host anti-viral response or autoimmunity [13•].

The transformative activity of MCPyV sT is due, in part, to the large T stabilization domain (LSD) located between amino acids 91–95 [14]. The LSD binds and inhibits E3 ubiquitin ligase (SCFFbw7) [14]. Because MCPyV LT is a target of Fbw7-mediated ubiquitination to induce proteasomal degradation, sT increases the half-life of LT protein [14]. Furthermore, inhibition of Fbw7 by MCPyV sT also increases the levels of other Fbw7 target proteins such as c-Myc and cyclin E which contribute to increased cell proliferation [14].

The MCPyV sT LSD also increases hyperphosphorylated 4E-binding protein 1 (4E-BP1), a regulator of cap-dependent translation. Active 4E-BP1 inhibits eukaryotic translation initiation factor 4E (eFI4E) function, thereby inhibiting translation [12]. 4E-BP1 is inactivated through phosphorylation by Mammalian target of complex 1 (mTOR1) [12]. Shuda et al. showed that, independent of mTOR1, MCPyV sT promotes hyperphosphorylation of 4E-BP1, thereby increasing protein translation [12]. Thus, MCPyV sT increases protein translation, reduces proteasomal degradation, and thereby promotes tumor cell survival via increased levels of c-Myc, cyclin E, and MCPyV LT antigen.

LT Antigen

MCPyV LT also plays specific fundamental roles in promoting VP-MCC tumorigenesis. The C-terminal domain of LT contains a helicase region important for initiating viral DNA replication [6, 8]. Importantly, loss of viral replication is necessary for tumorigenesis because initiating replication within integrated virus promotes DNA damage responses and host cell death [5, 6, 8, 10, 14, 15]. Therefore, C-terminal mutations disrupting LT helicase function are essential for VP-MCC tumor development [15]. The preservation of N-terminal LT in tumors suggests that it is necessary for host cell transformation.

MCPyV LT antigen can inhibit tumor suppressors and activate oncoproteins in MCC. First, the Rb binding domain (amino acids 211–217) of LT is never mutated in MCC [16]. This domain binds and inhibits Rb, thereby preventing Rb-mediated suppression of E2F activity [16, 17]. Release of E2F, by LT sequestering Rb, allows transcription of cell cycle genes involved in G1 to S-phase transitions, thus promoting tumor growth [16, 17]. The LT Rb-binding domain is also required for upregulation of the anti-apoptotic oncoprotein, survivin, which can be targeted pharmacologically to delay MCC xenograft tumor growth in mice [18]. Another target of the MCPyV LT is the tumor suppressor p53 [19]. Although MCPyV LT is not known to bind p53 directly, Borchert et al. showed that the LT expression leads to reduced p53 transactivation activity [19]. Taken together, integrated MCPyV employs multiple sT and LT-mediated mechanisms to promote tumor development and growth.

Merkel Cell Polyomavirus-Negative MCC Pathogenesis

The 20% of MCC that are polyomavirus negative (VN-MCC) [3] contain an exceptionally high somatic mutational burden, enriched for ultraviolet signature C > T transitions throughout the genome (Fig. 1a) [9•, 20•, 21•, 22•]. Interestingly, the two most commonly mutated tumor suppressor proteins in VN-MCC, Rb and p53, are both targets of MCPyV LT antigen [18, 19, 21•]. In VN-MCC, Rb is generally lost through genome deletion or epigenetic hypermethylation [20•, 21•, 23, 24•]. Loss of Rb in VN-MCC leads to increased E2F activity and thus increased tumor growth [17]. The majority of p53 mutations occurring in MCC are in VN-MCC tumors, which often show high p53 immunostaining [20•, 21•, 24•]. The inactivating mutations in p53 result in downregulation of p53 targets, thereby preventing tumor cell senescence, cell cycle arrest, DNA damage repair, and apoptosis [18, 20•, 21•].

Combined cutaneous squamous and Merkel cell carcinoma is an established variant of MCC that are consistently MCPyV-negative [25]. Pulitzer et al. found that, like pure VN-MCC, combined cutaneous squamous and Merkel cell carcinoma tumors have a high mutational burden including frequent mutations in both p53 and RB1, leading to increased expression of p53 and decreased expression of Rb [22•]. Cutaneous squamous and Merkel cell carcinoma also show high expression of p63 in the squamous component of the tumor [22•]. Interestingly, the ΔNp63α isoform of p63 is known to be overexpressed in squamous cell carcinoma and is thought to be a hallmark of both squamous cell and basal cell carcinoma tumorigenesis [26,27,28,29,30,31,32,33, 34•].

Taken together, loss of Rb and p53 tumor suppressors function seems to be a consistent feature of both VP-MCC and VN-MCC tumors. Whereas a number of additional oncogenic processes are consistently activated by viral oncogenes in VP-MCC, there is little commonality among the numerous additional UV-induced driver mutations found in VN-MCC tumors. A few oncogenic pathways such as MYCL and phosphoinositide 3-kinase (PI3K) activation have been implicated in VN-MCC tumors; nonetheless, only subsets of VN-MCC show activation or upregulation of these oncogenic pathways [20•, 21•]. Thus, most MCCs are MCPyV-driven tumors with low mutational burdens, and the remainder are a heterogeneous set of VN-MCC tumors driven by UV-signature mutations.

Diagnosis and Staging

MCC typically presents as a rapidly growing, asymptomatic, erythematous nodule (Fig. 2a). Polymorphous vessels and milky-red structureless areas are often seen on dermoscopy (Fig. 2b). Because MCC is rare and clinically nondescript, it is seldom suspected on clinical exam. Diagnosis is best done by an experienced dermatopathologist. Because of its aggressive nature, management of MCC requires multidisciplinary care. Referral to a cancer center with a cutaneous oncology tumor board should be considered. The general staging and management of MCC are summarized in Fig. 1b.

Fig. 2
figure 2

Clinical and histopathological appearance of Merkel cell carcinoma. The clinical (a) and dermatoscopic (b) presentation of a MCC tumor on the leg of a patient. The central crust is a healing punch biopsy site. c Hematoxylin and eosin (H&E) staining of a tissue section from a MCC tumor biopsy from another patient showing sheets of small blue cells with neuroendocrine features. d Immunostaining for cytokeratin 20 (CK20) shows staining in the cytoplasm of the cells with frequent paranuclear dots. Original magnification ×200

Full size image

Diagnosis

Primary Tumor

MCC diagnosis requires biopsy and tissue examination with immunohistochemistry (IHC). Histologically, MCC is comprised of small round blue cells with neuroendocrine features arranged in sheets or narrow cords (Fig. 2c). IHC markers are needed to distinguish MCC from other tumor types. MCC expresses nonspecific neuroendocrine (e.g., synaptophysin, chromogranin, neuron-specific enolase) and epithelial (e.g., pancytokeratin, epithelial membrane antigen) markers. CK20 is currently the most sensitive and specific marker for MCC and is often expressed in a paranuclear dot-like pattern (Fig. 2d). Excluding cutaneous metastases from other neuroendocrine tumors is important. Negative TTF-1 staining generally excludes metastatic small cell lung cancer [35]. Pathology reports for MCC should include comments on tumor size and depth, margin status, lymphovascular invasion, and extracutaneous extension. Other factors to report include mitotic index, tumor-infiltrating lymphocytes, tumor growth pattern, and presence of a second malignancy [36].

Lymph Nodes

As regional involvement is common, sentinel lymph node biopsy (SLNB) should be considered for all MCC tumors. Immunostaining with CK20 or pancytokeratin should be used to evaluate lymph nodes. IHC with CK20 increases the sensitivity for detecting micrometastatic disease [37, 38]. The pattern of lymph node involvement may have prognostic implications, with worse overall survival when sheet-like patterns of tumor cells infiltrate the lymph node [39].

Staging

MCC can be clinically and pathologically staged using the American Joint Cancer Committee (AJCC) Cancer Staging Manual. Updates to the upcoming eighth edition of AJCC guidelines will reflect the observation that nodal disease (stage III) of unknown primary has a better prognosis than nodal disease with known primary [40]. In addition to AJCC staging, the patient’s immune status significantly impacts prognosis, with worse survival in those with impaired immunity [41]. A relatively new adjunct test uses serology for antibodies against the MCPyV sT antigen at diagnosis as an independent predictor of decreased MCC recurrence [42].

Management

Surgery

Wide local excision is the treatment approach for primary MCC. No studies have correlated margin size and recurrence risk, but 1–2-cm margins and excision down to fascia are recommended [36]. Moh’s micrographic surgery (MMS) can be considered if tissue sparing is of priority [36]. A small retrospective study of 22 patients treated with MMS showed a 5% recurrence rate after a median follow-up of 37.5 months [43]. If MMS is performed, it should be done so as to not interfere with SLNB. Extensive reconstructive procedures that involve undermining tissue should be delayed until histologic margins are confirmed negative by IHC on permanent sections and SLNB is performed [36].

Lymph Nodes

Clinical Node Negative Disease

All patients with clinically node negative, primary MCC should be offered a SLNB. A study of 240 cases suggested that tumor size does not reliably predict nodal involvement [44]. In addition, clinically occult lymph node disease is not uncommon; a study of 403 cases showed that about one third have nodal involvement on SLNB [45••]. Patients with lymph node disease have increased risk of recurrence, but the impact on overall survival is unclear [36, 46]. Nonetheless, lymph node status determines adjuvant treatment options for regional control.

Adjuvant therapy is recommended in patients with positive SLNB. A study of 29 patients with positive SLNB showed that those who received adjuvant treatment (radiation therapy [RT], chemotherapy, or complete lymph node dissection [CLND]) had a 3-year relapse-free survival of 51 versus 0% in those who did not [46]. Patients with positive SLNB should undergo CLND if they are appropriate surgical candidates. If CLND cannot be performed, RT to the nodal basin is an alternative option [36]. In a study of 171 patients with nodal disease, RT to the nodal bed was associated with higher 3-year disease-specific survival (76.2 versus 48.1%) [47••]. Patients with nodal disease should also receive baseline imaging studies.

Those with negative SLNB can be managed with regional observation alone. Recurrence in the same nodal basin is low (4–14%) after negative SLNB [48, 49]. Multiple studies have suggested that compared to observation alone, RT does not decrease regional recurrence if the SLNB is negative [48, 49]. In addition, RT to the nodal bed in SLNB negative patients does not impact 3-year disease-specific survival [47••]. In high-risk patients (e.g., immunosuppression) or if a false-negative SLNB is suspected, nodal RT can be considered.

Clinical Node Positive Disease

Clinically evident lymph node disease should be biopsied for confirmation of metastatic disease. Fine needle aspiration (FNA) or core biopsy with immunostaining is recommended and, if positive, should be followed by CLND and imaging studies. Adjuvant RT should also be considered if there is extracapsular extension or if multiple nodes are involved [36]. If the FNA or core needle biopsy is negative, an open biopsy should be considered.

Radiation

Adjuvant radiation is an option for all stages of MCC [36]. Radiation to the primary tumor bed should be initiated within weeks of surgical excision, with higher doses if resection margins are positive [36, 50]. Several studies suggest that RT decreases local and regional recurrence and may have an overall survival benefit [47••, 51]. A recent retrospective study of 171 MCC patients with known lymph node status and without distant metastases showed that RT to the primary tumor bed and/or lymph node basin improved 3-year local and regional control, disease-free survival, and overall survival, but had no effect on disease-specific survival [47••]. Another retrospective study of 1254 patients showed that local RT post-resection decreases local and regional recurrence compared to surgery alone [51]. Two large studies based on the Surveillance, Epidemiology, and End Results database suggest that adjuvant radiation improves overall survival, but not disease-specific survival [52, 53]. The only prospective randomized controlled trial investigating RT in MCC involved 83 patients with stage I disease who all underwent WLE and RT to the tumor bed. This study evaluated the effects of additional RT to the regional lymph nodes compared to observation. Prophylactic radiation to the regional nodes decreased regional recurrence compared to observation, but there was no survival benefit. However, this study dropped in recruitment and terminated prematurely after SLNB became a standard of care [54].

Observation without RT to the tumor bed may be considered if the primary tumor is small (< 1 cm), widely excised, and without other high-risk features [55]. When appropriate, patients with positive SLNB should receive RT to the regional lymph node basin, while those with negative SLNB can be observed [46, 48, 49].

In surgically unresectable cases, RT can be used as monotherapy, but recurrence is not uncommon [56]. For palliative purposes, hypofractionated or single-fraction RT can be used [36].

It is worth noting that most studies of RT in MCC are retrospective in design and are limited by variability in cohorts, with inconsistent specification of lymph node status, additional treatments (e.g., chemotherapy), and RT parameters (e.g., dose, timing, site). Many studies were performed prior to the routine use of the SLNB as a standard of care. Currently, there is a lack of prospective data investigating the role of RT in MCC management.

Imaging

Baseline imaging should be performed in patients with positive SLNB or if there is clinical suspicion of metastatic disease. FDG-PET/CT is the preferred modality [57,58,59]. PET enables detection of bone and bone marrow involvement and should be combined with CT to detect liver and lung metastases. If FDG-PET/CT is unavailable, MRI or CT are alternatives.

Somatostatin receptors (SSTR) are expressed in most MCC and are experimental targets for molecular imaging. Small studies have investigated the use of somatostatin analogues (e.g., [68Ga]DOTA-D-Phe1-Tyr3-octreotide (DOTATOC) or -octreotate (DOTATATE)) combined with PET (SSTR-PET) to identify loci of metastatic MCC. SSTR-PET has high sensitivity for the bone, soft tissue, and brain metastases, but CT is still required for reliable detection of liver and lung lesions [60].

Chemotherapy

Chemotherapy is reserved for distant metastatic disease. Guidelines are not well defined, and regimens are based on MCC’s similarity to small cell lung cancer. The most common regimen includes a platinum-based agent with or without etoposide. Others may use cyclophosphamide, doxorubicin, and vincristine (CAV), or topotecan. A recent review showed that the objective response rate (ORR) with chemotherapy is higher in the first-line compared to second-line setting (53–61 versus 23–45%). Although MCC is chemosensitive, there are high toxicity and responses that are seldom durable. The duration of response is less than 8 months with most recurrences occurring within 6 months [61•]. The effect of chemotherapy on overall survival is unclear. Chemotherapy’s shortcomings and studies of MCC biology prompted the investigation of targeted therapies as well as immunotherapies for treatment of metastatic disease.

Targeted Therapy

PI3K/Akt/mTOR Pathway Inhibition

Signaling through the PI3K/Akt/mTOR pathway is upregulated in a small subset (4–10%) of MCC due to activating mutations in PIK3CA and AKT1 [62,63,64]. There is a case report of a patient with metastatic MCC with a known PIK3CA mutation that had complete response after treatment with idelalisib, a PI3Kδ inhibitor [65]. A phase I/II clinical trial investigating the mTOR inhibitor MLN0128 for metastatic MCC is ongoing (NCT02514824).

Tyrosine Kinase Inhibition

MCC tumors frequently express tyrosine kinase receptors for signaling molecules including VEGF-A, VEGF-C, VEGFR-2, PDGF-A, and c-kit. As such, tyrosine kinase inhibitors have been studied as potential therapeutics [66, 67]. Pazopanib, a multitargeted tyrosine kinase inhibitor, was used in one patient who had progressed on chemotherapy, with a partial response at 6-month follow-up [66]. Imatinib was studied in c-kit-positive metastatic MCC in a phase II trial and showed a brief partial response in 1/23 (4%) of patients [67]. A phase II trial with cabozantinib, a small molecule inhibitor of c-MET and VEGFR-2, in patients with recurrent or metastatic MCC that progressed on chemotherapy is ongoing (NCT02036476).

Bcl-2 Antisense Oligodexoyribonucleotide Therapy

Preclinical studies found that Bcl-2, an anti-apoptotic protein, is upregulated in MCC. A phase II study of olimersen, a Bcl-2 antisense oligodexoyribonucleotide, showed no objective response in 12 patients with metastatic or regionally recurrent MCC [68].

Somatostatin Analogues

Somatostatin (SST) has an anti-proliferative effect on neuroendocrine tumor cells. Because ~ 90% of MCC express somatostatin receptors (SSTR), SST analogues are being investigated [69]. There was an early report of a complete response after treatment with lanreotide in a patient with metastatic MCC [70]. Current clinical trials include a recently completed phase I trial with pasireotide for metastatic MCC (NCT01652547) and an ongoing phase II trial with lanreotide for unresectable and/or metastatic MCC (NCT02351128).

SST analogues can also be paired with radionucleotide therapy, in which a radiolabeled SST analogue is taken up by SSTR-expressing tumor cells, leading to emission of local radiation. A phase II study investigating this approach with 177Lu-DOTATATE for treatment of SSTR-expressing NETs, including MCC, recently completed with results pending (NCT01237457).

Immunotherapy

The immunogenicity of MCC has prompted multiple immunotherapy trials in MCC [71].

Intratumoral Vaccines

Intratumoral vaccines stimulate the host immune response while circumventing systemic toxicity. Talimogene laherparepvec (T-VEC) is an oncolytic virus engineered from attenuated HSV-1 that has been modified to secrete GM-CSF. T-VEC injections are FDA-approved for unresectable and injectable melanomas [72]. An ongoing phase II trial is investigating T-VEC with or without hypofractionated RT in melanoma, MCC, and other solid tumors with skin metastases (NCT02819843).

In mouse melanoma, intratumoral injection with IL-12 plasmid vaccine followed by electroporation increases IL-12 and IFN γ, and reduces tumor vascularity [73]. This same treatment regimen was subsequently studied in a phase I trial in patients with metastatic melanoma, and the majority had stable disease, partial response, or complete response [74]. A phase II trial of IL-12 plasmid vaccine followed by in vivo electroporation for MCC recently completed enrollment (NCT01440816).

Adoptive Cell Transfer

Immunotherapy with adoptive cell transfer involves harvesting, expanding, and infusing immune cells to mount an antitumor response. This process can be autologous, in which case T cells are harvested from the patient, modified or cultured in vitro, and then reinjected into the patient. Transferred cells can also be from allogenic sources.

One ongoing phase I/II trial for metastatic MCC combines transfer of autologous MCPyV T-antigen-specific CD8+ T cells with the immune checkpoint inhibitor avelumab, in addition to either localized RT or recombinant IFN-β (NCT02584829). Another phase II study involves infusions of innate natural killer (NK) NK-92 cells and the NK stimulating cytokine IL-15 (NCT02465957).

Immune Checkpoint Inhibition

Programmed Cell Death Receptor 1 (PD-1)/Programmed Cell Death Receptor Ligand 1 (PD-L1)

Immune checkpoints prevent chronic activation of immune responses. For example, signaling to the T cell receptor PD-1 by PD-L1 expressed on tumor or immune cells leads to T cell inactivation and local immune tolerance. Blocking PD-1 signaling with monoclonal antibodies targeting PD-1 or PD-L1 enhances T cell activation and antitumor activity (Fig. 1a). In MCC, PD-L1 is expressed on tumor cells and local immune cell infiltrates [75]. In addition, MCPyV-specific tumor infiltrating lymphocytes have high levels of PD-1 expression [76]. These findings rationalize using PD-1/PD-L1 blocking antibodies in MCC.

A phase II trial studied the use of pembrolizumab, a monoclonal antibody targeting PD-1, as first-line therapy for 26 patients with metastatic MCC [77••]. The ORR was 56%, with responses observed independent of tumor viral status or PD-L1 expression. Progression-free survival at 6 months was 67%. The most common side effects were fatigue and laboratory abnormalities. By Common Terminology Criteria for Adverse Events (CTCAE), grade 3 or 4 drug-related adverse events occurred in 15% of patients, including myocarditis and elevated liver enzymes.

A larger phase II trial tested avelumab, an anti-PD-L1 monoclonal antibody, as a second-line agent in 88 patients with chemotherapy-resistant metastatic MCC [78••]. The ORR was 31.8%, with several patients achieving complete response. There was no difference in response rates across subgroups (e.g., tumor PD-L1 expression, tumor MCPyV status, baseline disease burden). Response to treatment with avelumab was durable, with an estimated 6-month response durability of 92%. Treatment with avelumab was well tolerated, and the most common side effects were fatigue and infusion reactions. Five patients had grade 3 adverse events, including lymphopenia and elevated liver enzymes, creatinine phosphokinase, and blood cholesterol. There were no grade 4 adverse events. Avelumab is now the first FDA- and EMA-approved drug for metastatic MCC. A phase III trial testing adjuvant avelumab in resected stage III MCC is ongoing (NCT03271372).

Other immune checkpoint inhibitors for metastatic MCC are currently under investigation. A phase I/II trial of nivolumab (anti-PD-1) for metastatic virus-associated tumors including MCC is ongoing. Preliminary results with 22 patients with MCC revealed an ORR of 68% [79]. A phase II trial with atezolizumab (anti-PD-L1) in combination with bevacizumab (anti-VEGF) in rare solid tumors including MCC is recruiting participants (NCT03074513).

Cytotoxic T-Lymphocyte Associated Protein 4 (CTLA-4)

Signaling through CTLA-4, another immune checkpoint receptor on T cells, downregulates T cell activity and is permissive to tumor cells’ immune evasion. In a series of five cases of metastatic MCC treated with the anti-CTLA-4 antibody, ipilimumab (3 mg/kg every 3 weeks for 4 cycles), two patients had a complete response, two had stable disease, and one had disease progression [80]. There is an ongoing phase II trial that initially was comparing adjuvant ipilimumab versus observation after complete MCC resection. However, due to the higher toxicity of anti-CTLA-4, patients are now randomized to receive either nivolumab (anti PD-1) or observation (NCT02196961).

Combinations of Checkpoint Inhibitors

Ongoing trials combining multiple immune checkpoint inhibitors for metastatic MCC include a phase II trial combining nivolumab and ipilimumab, with and without stereotactic body RT (NCT03071406), and a phase I/II trial combining tremelimumab (anti-CTLA-4) and durvalumab (anti-PD-L1) with a tumor microenvironment immunostimulant, poly-ICLC (NCT02643303).

Disease Monitoring and Follow-up

All MCC patients should be monitored closely by a dermatologist. For the first 3 years following diagnosis, a full body skin and lymph node exam should be performed every 3–6 months, and every 6–12 months thereon afterwards. Close follow-up is critical, as median time to recurrence is 9 months, with 90% of recurrences occurring within 24 months [36]. Routine imaging should be considered in high-risk patients.

Circulating biomarkers for monitoring disease progression or recurrence have recently been investigated. Based on a retrospective study of 60 MCC patients, blood levels of neuron specific enolase and chromogranin A are not useful for predicting outcomes or detecting recurrence [81]. However, circulating tumor cells have potential utility in following MCC disease course [81, 82]. The only validated biomarker to monitor for MCC recurrence is the serial evaluation of circulating anti-MCPyV sT antibody titers [42]. Increasing titers have a positive predictive value of 66% for clinical recurrence, and titers may rise before recurrence is detectable on physical exam. However, there are limitations with this approach. Circulating sT antibodies are only found in some patients with MCPyV-positive tumors (52% of MCC patients), and patients must be tested soon after initial treatment because titers may fall below detection post-treatment.

Conclusions

MCC is a rare disease requiring multi-institutional collaborations to investigate its natural history and develop treatment protocols. For patients with advanced disease, first-line immunotherapy with anti-PD-1/PD-L1 checkpoint inhibitors is well tolerated and offers a durable response compared to traditional chemotherapy. Currently, the only approved drug for metastatic MCC is avelumab. However, not all patients respond to immune checkpoint inhibitors, and additional therapies are needed. As our understanding of VP-MCC and VN-MCC biology and pathogenesis develops, additional treatments can be investigated.