Updates in the Management of Merkel Cell Carcinoma

Historically described by Toker in 1972 as a trabecular carcinoma of the skin, Merkel cell carcinomas (MCC) are rare and highly aggressive neuroendocrine cutaneous malignant neoplasms.1 2. MCC has an estimated annual incidence of 0.4 cases per 100.000 individuals in the American population, reaching 4.85 cases / 100.000 in the population over 80 years of age, with a tendency to increase as demonstrated in the last decades.3 However, MCC epidemiology exhibits a significant geographic heterogeneity: in Australia, incidence rates four times higher than in the American population have been reported, reaching almost 20.4 cases/100.000, also with a growing number of cases diagnosed over the past decades.4

Risk factors traditionally associated with the development of MCC include fair skin, chronic sun exposure, immunosuppression (e.g. solid organ transplantation, presence of HIV infection or concomitant malignant neoplasm) and advanced age, with a marked increase in the incidence rates in the 8^th and 9^th decades of life.2 5 In 2008, the presence of MCC-associated polyomavirus (MCPyV) has been linked to the development of a variable portion of these neoplasms. MCPyV is a human viral pathogen member of the Polyomavirus family, whose infection is common in childhood and typically self-limited. Approximately 60-80% of adults have positive serology for previous infection.6

In MCC, MCPyV circular double-strand DNA is capable of integrating into the host genome, resulting in the coding of viral oncoproteins known as Large T (LT) and Small T (ST) antigens.7 8 Such an event occurs early in the tumorigenesis, as the viral genome can be characterized in different cellular subclones. These antigens capability of directly binding to and negatively interfering with the function of important tumor suppressors such as pRb, p53 and FBXW7, encompasses the basis of carcinogenesis in MCPyV positive tumors.7 9 10 It is estimated that about 80% of MCC cases in the northern hemisphere are associated with the presence of MCPyV; in Australia, only about 30% of the cases are considered MCPV positive, being the vast majority attributed to the damage from chronic exposure to ultraviolet (UV) radiation.3 4 7 11 12 13 MCPyV positive MCCs are amongst the neoplasms with the lowest mutational burden (TMB), presenting on average 0.4-0.75 mutation/Mb and 12.5 somatic single nucleotide variants (SSNV) per exome, without highly recurrent alterations or UV damage signatures. This low TMB observed in MCPyV positive tumors suggests that the disorders caused by the activity of ST and LT oncoproteins are the most important factor in tumorigenesis. Nonetheless, the detection of serum antibodies against T antigens in this subgroup points towards to the presence of both recognition and response by the immune system, despite the low TMB, suggesting an intrinsic immunogenic role for MCPyV.9 14 MCPyV negative MCCs, on the other hand, exhibit both TMB and median number of neoantigens even higher than cutaneous melanomas, with about 40 mutations/ Mb, 1.121 SSNV per exome and 173 neoantigens per sample. Among the genes most frequently affected by pathogenic mutations in this subgroup are TP53 (70%), RB1 (45%), NOTCH1, FAT1, ATM, MSH2, BRCA1, MAP3K1, and TRAF7. Such profile reinforces the importance of p53 and pRb activity interference in the development of MCC, either by direct protein inactivation or somatic mutations. [9,12,13] It is also important to note that the frequency of C>T transitions, considered signatures of UV- induced damage, is significantly higher in MCPyV negative tumors, being observed in 87% of SSNV, reiterating the importance of chronic sun exposure in the carcinogenesis of this subgroup of neoplasms.12 14

The cellular origin of MCCs remains controversial. Due to the ultrastructural e immunohistochemical similarities with Merkel cells, it was initially proposed that these tumors would arise from these structures.15 Nonetheless, characteristics such as (1) very limited proliferative potential (with maintenance of its homeostasis by epidermal progenitor cells), (2) topography divergence between Merkel cells (located in the basal layer of the epidermis) and MCCs arising (dermis or hypodermis), as well as (3) differences in the arrangement of the cytoskeleton have been fostering discussions about its normal counterpart.9 10 Making this scenario even more complex, in a recent publication, Sunshine et al proposed that MCCs MCPyV positive and MCPyV negative would respectively derive from dermal and epidermal stem-cells, which, in turn, would converge into the process of neoplastic transformation to the neuroendocrine phenotype in the presence of deleterious interference with pRb function. According to the authors, this process of transdifferentiation would be, if corroborated by future studies, the first report of convergent evolution between two distinct embryonic lines (ectodermal and mesodermal) in malignant neoplasms.9 Finally, a possible origin from pre / pro-B cells has also been suggested, since an expressive percentage of these neoplasms exhibit B-cell line markers (e.g. PAX-5, TdT, bcl-2) and clonal rearrangement of immunoglobulins.10 16

Irrespectively of the underlying mechanisms, MCCs present an aggressive clinical course, with lethality rates even higher than paired-stage cutaneous melanomas. Both the presence of MCPyV, initial staging, and the absence of primary lesion at the time of diagnosis appear to be associated with better outcomes; the latter, in turn, highlights the possibility of an existing relationship between the primary lesion regression, higher TMB and a more efficient anti-tumor immune response, which would result in better 5-year overall survival rates favoring those with node-positive disease without an identifiable primary (42% vs. 27%).2 13 17 18 19 20 According to Harms et al, it is estimated that only about 8% of the patients present with distant metastasis at the time of diagnosis, being the vast majority diagnosed with local (65%) or locally-advanced (26%) disease.

Historically, the 5-year overall survival rates range from 14% to 51% for those with local and metastatic disease.20

MCC often presents as a fast-growing violaceous nodule, usually asymptomatic, that appears in sun-exposed skin. Nevertheless, the initial presentation can be non-specific. Multiple cohort studies describe head and neck as the most common anatomic site of the primary tumor.21 In order to facilitate its early detection, key clinical features were summarized in an acronym: AEIOU - asymptomatic, expanding rapidly (significant growth in = 3 months), immune suppression, older than 50 years age, and UV-exposed area in a fairskinned individual. The presence of at least three of these characteristics increases the suspicion of MCC.22

The differential diagnoses encompass benign and malignant cutaneous lesions, including basal cell carcinoma, squamous cell carcinoma, keratoacanthoma, pyogenic granuloma, lipoma and adnexal tumors. Because of their similar microscopic appearance, other possibilities are metastases of small cell carcinoma of the lung, small cell melanoma and Ewing's sarcoma.23 24

On histopathologic examination, the tumor presents as a poorly defined, sheet-like mass involving the dermis. Generally, the epidermis is spared and neoplastic infiltrate is confined to the papillary dermis and subcutaneous tissues. Intraepithelial layer impairment may occur. Increased mitotic activity and necrosis are often seen. In hematoxylin-eosin staining (H&E), tumor cells are monotonously small, round, and blue, with a finely dispersed nuclear chromatin pattern and scant cytoplasm, forming peculiar trabeculae.10 25 Immunohistochemistry demonstrates expression of neuroendocrine markers and particular low-molecular weight cytokeratin immunoreactivity, which is expressed in a characteristic paranuclear punctate or “dot-like” pattern. The main marker of MCC is cytokeratin 20 (CK20), with a sensitivity greater than 90%.26 Virtually all MCC are positive for neuron-specific enolase (NSE), whereas chromogranin B and A are found in 100% and 72% of the tumors, respectively. Some studies have shown c-kit positivity in up to 95% of cases. S-100 protein, HMB-45, CK7, TTF-1, desmin, and actin are typically negative.27 28

MCPyV LT antigen can be detected by immunohistochemistry using mouse monoclonal antibody CM2B4 and the viral DNA by real-time polymerase chain reaction in tumor tissue.29 Nevertheless, the investigation of MCPyV is not mandatory for the diagnosis of MCC. Antibodies to MCPyV-oncoproteins can also be detected in peripheral blood and are associated with the risk of recurrence. Paulson et al, in a prospective validation study, evaluated 219 patients with newly-diagnosed MCC and the presence of antibody to MCPyV-oncoproteins at diagnosis was an independent predictive factor of decreased recurrence risk (HR 0.58; p =0.04). Additionally, among patients with positive antibodies, an increasing titer after definitive therapy was associated with clinically recurrence (positive predictive value of 66%) and it could be useful for ongoing surveillance. Patients with negative antibodies upon initial presentation are at higher risk for recurrence and may benefit from closer follow-up with imaging, in contrast to patients with positive antibodies, for whom alternating serial MCPyV antibody titer assessment with imaging could be an option.6 30

Once the diagnosis of MCC is established, initial work-up should include imaging tests. Positron emission tomography with fluorodeoxyglucose (¹⁸ F-FDG-PET/CT), computed tomography (CT) and magnetic resonance imaging (MRI) are viable options. There is no consensus on which method is best for work-up and staging of MCC. Treglia et al. conducted a systematic review and meta-analysis that examined the performance of PET-CT. It showed elevated accuracy and effectiveness in detecting regional lymph node and distant metastatic disease with a sensitivity of 90% and specificity of 98 %.31 The extent of disease at presentation is the most important predictor of survival for MCC and determines management of these patients. According to the American Joint Commission on Cancer (AJCC) 8th Edition, MCC is classified by tumor size, involvement of regional lymph nodes and presence of distant metastases. (Table 1)32.

It is known that lymphatic dissemination is frequent and responsible for a considerable amount of regional recurrences. In about 20-30% of patients, nodal involvement is present at initial presentation and, in multiple series, there is a major agreement that lymphatic invasion is an unfavorable prognostic factor.36 46 47 48 In addition, 30-50% of the patients develop nodal recurrence in the course of the disease.49 Sentinel lymph node (SLN) mapping and biopsy is a well-established technique in the context of melanoma. It aims to detect subclinical nodal disease and, therefore, changes treatment decision. Due to the rarity of MCC, the experience of SLN biopsy is mostly extrapolated from the experience with melanoma.50 Gupta et al. published a single institution series of SLN biopsy in MCC with clinically negative nodes and demonstrated a positive SLN in 32% of the cases; there was a difference in relapse free survival favoring the patients who underwent adjuvant treatment for the involved node bed.51 Scwhartz et al. examined clinical and pathological features aiming to asses which patients would obtain benefit from SLN biopsy: among 97 tumor specimens, 45.7% showed SLN involvement. After a multivariate analysis, there was not a subgroup of patients or threshold for precluding the procedure, with a likelihood of at least 15-20% of SLN positivity.48 Finally, a metanalysis published by Sadeghi et al. showed that SLN biopsy predicted better disease free survival (DFS) and overall survival (OS) than nodal observation.52 This led to the recommendation from American and European guidelines to perform SLN mapping and biopsy for all clinically node negative patients.45 47

Given that patients with lymph node positive disease are at greater risk of recurrence and have a worse prognosis in terms of DFS and OS52, guidelines recommend treatment with regional lymph node dissection and/or RT to the nodal basin for those with clinically involved LN.45 47 Nonetheless, the available data to support this approach is limited and yet again based on retrospective studies and case series.

Based upon the well documented radiosensitivity of MCC, tumor bed RT has been used in order to improve local control. Several studies aimed to assess the benefit of adjuvant RT for local and locoregional disease. Veness et al published in 2005 a case series that evaluated the risk of local and nodal recurrence in 86 patients treated with surgical approach followed or not for adjuvant RT. Twenty two percent presented with clinically nodal disease and another 19% presented with lymph node involvement without a primary lesion. There was no difference in rates of local recurrence, but patients treated with surgery plus RT had a lower incidence of nodal recurrence, which translated into an improved DFS (10.5 months vs 4.0 months).37 In another paper, Lewis et al. performed a database analysis encompassing 1254 patients and found reductions in local and regional recurrences with similar rates of distant metastases, but no statistical difference in terms of overall and MCC specific survival.53 In the largest series to date aiming to assess whether adjuvant therapy was associated with better survival that included 6908 cases from the National Cancer Database (NCDB), Bathia et al. showed a reduced risk of death favoring the group receiving adjuvant RT; however, this benefit was limited to patients with negative nodal disease. Patients with node positive disease did not benefit in terms of survival from the addition of adjuvant RT.36 These findings suggest that the addition of radiotherapy may benefit the local control in localized disease; however, it is reasonable to believe that survival in patients with more advanced disease may be driven by the presence of subclinical distant metastasis.

Several other case series and retrospective studies addressed the risk of recurrence and the use of adjuvant RT.34 35 46 There are many limitations in these studies, such as selection bias, confounding factors, limited and missing data, lack of randomization, among others. As guideline recommendations from the European and American groups, patients with localized node negative disease are to be offered surgery followed by tumor bed RT.45 47

The role of adjuvant RT and chemotherapy (CT) has been assessed by Bathia et al and no difference in terms of survival following the addition of RT or CT to the treatment of stage III disease was observed; nevertheless, limitations to this retrospective study included missing data on CT and RT schemes, as well as status on lymph node dissection.36

A prospective trial was designed by the Trans-Tasman Radiation Oncology Group (TROG 96:07), aiming to assess the safety and tolerability of concurrent chemoradiotherapy as adjuvant or definitive treatment for MCC. Candidates were high risk patients, defined by primary tumors larger than 1 cm in size, recurrent disease, gross residual disease following initial surgery, or MCC of unknown primary site with nodal involvement. Fifty-three patients were accrued to receive CT with carboplatin and etoposide days 1-3 in weeks 1, 4, 7 and 10 and RT delivered to the primary site and regional nodes to a dose of 50 Gy in 25 fractions over 5 weeks. This regimen resulted in a 3-year OS of 76%, locoregional and distant control rates of 75% and 76%, respectively, suggesting that this strategy could provide better outcomes when compared with historical data.54 Subsequently, Poulsen et al. performed a retrospective comparison between patients with similar baseline characteristics in order to elucidate whether the CT arm would achieve better outcomes when compared to RT alone. In a multivariate analysis, the addition of CT did not provide additional benefit in terms of OS.55 Futhermore, in line with the aforementioned data, other series have suggested additional morbidity without improving survival for patients receiving adjuvant CT.56

Only one trial was able to show a survival benefit with the use of CT in the adjuvant setting. Chen et al. evaluated the use of adjuvant RT, adjuvant CRT and surgery alone. It showed a survival benefit for the use of adjuvant RT and adjuvant CRT over surgery alone and a survival benefit for CRT over adjuvant RT for a subset of patients with tumors > 3 cm, positive surgical margins or male sex.57 The data regarding the use of adjuvant CT are conflicting (with controversy over the real benefit on survival) and this modality is not routinely recommended by neither American nor European guidelines.45 42

To date, there is no defined role for adjuvant immunotherapy on MCC. Ongoing clinical trials are currently enrolling patients with completely resected tumors to evaluate the efficacy of nivolumab and ipilimumab as potential adjuvant therapies.58

CT has historically been the standard approach for patients with advanced (unresectable or metastatic) MCC, either at presentation or after prior definitive therapy. However, given the rarity of this tumor, there have been no definitive prospective clinical trials of CT for patients with this tumor, and most of the evidence and rationale has been extrapolated from regimens applied to neuroendocrine tumors of distinct primary sites. Treatment options are based in classical schemes used to treat small-cell lung cancer, and platinum doublets (including irinotecan or etoposide) have been long regarded as the historical standard approaches for those with advanced MCC. Topotecan and cyclophosphamide/doxorubicin/vincristine (CAV) have also been used in distinct series.59 Despite overall response rates (ORR) of up to 60%, the benefit from these strategies has been usually short-lived, with most patients developing disease progression in less than 6 months. As an example, Voog et al analyzed 107 published cases between 1980 and 1995. The median age at diagnosis was 66 years old, 70% of the patients underwent CT due to distant metastasis and 42 different regimens were employed. The most prescribed regimens were: cyclophosphamide/ifosfamide-based (56%), anthracyclin- based (49%) and platinum-based ones (25%); ORR were 57%, 45% and 20% in the first, second and third-line, respectively. The median OS was 9 months for patients with metastatic disease and no specific regimen was associated with significantly superior survival. Myelotoxicity was the most commonly reported toxicity and nine treatment associated deaths were reported (five due to febrile neutropenia and septic shock).60 Another case series published by Tai et al analyzed 204 cases, in which the most commonly used chemotherapy regimens were CAV, cyclophosphamide/epirubicin/vincristine (CEV) or etoposide/cisplatin (EP). CAV or CEV had an overall response rate of 75.7% and etoposide-platinum regimens led to an ORR of 60%. The median OS for the 204 patients was 21.5 months (including local, locoregional and advanced disease in the analysis). There were 6 deaths certainly related to treatment toxicity. [61] Becker et al analyzed retrospectively refractory patients beyond second-line therapies; CT achieved an ORR of 8.8%, median duration of response (DOR) of 1.9 months, median PFS of 3.0 months and median OS of 5.3 months, in line with previous reports of the ineffectiveness of second-line or subsequent treatments in patients with previously-treated metastatic MCC.62

The limitations to those studies are particularly the lack of standardized chemotherapy regimens, heterogeneous patient population, the limited number of patients, and their retrospective nature.45 63 64 Until the approval of immune checkpoint blockade, some guidelines would recommend the inclusion in clinical trials or best support of care as the standard options for patients with advanced MCC.45 59

The role of immune system in the development of MCC has been long suggested. Higher rates of MCC in immunosuppressed patients and better outcomes in patients with unknown primary lesion, suggesting immunological response and regression, were the first evidence of the association between the immune system and MCC. More recently, the discovery of MCPyV positive tumors associated with antibodies production as well as high TMB and neoantigens in MCPyV negative tumors confirmed the immunogenic potential of MCC and the rationale for the use of immunotherapy with PD1/PDL1 inhibitors. Indeed, half of MCC express PD-1 on tumor infiltrating lymphocytes and express PD-L1 on tumor cells.65 The performance of different treatment regimens in unresectable or metastatic MCC is shown intable 2.

Nghiem et al evaluated the efficacy of first-line pembrolizumab, in a phase II trial that included 26 patients with stage IIIB or IV MCC. The median age of the patients was 68 years, 35% of patients were classified as MCPyV negative, and 65% were MCPyV positive tumors. The median TMB was 1121 muts/Mb e 12.5 muts/Mb in MCPyV negative and MCPyV positive MCC, respectively. Pembrolizumab was administered intravenously at a dose of 2 mg per kilogram of body weight every 3 weeks for a maximum of 2 years or until a complete response, dose-limiting toxic effects, or progressive disease occurred. The median follow-up was 33 weeks. The ORR was 56% and complete response (CR) rate was of 15%. The ORR was 62% among MCPyV positive tumors and 44% in MCPyV negative. Responses with pembrolizumab were durable, and 14 of 26 patients were still on treatment by the time of the final analysis. The estimated progression-free survival rate at 6 months was 67% and 77% of the patients experienced any grade adverse events (AE) during treatment (15% of grade 3 or 4). The most common AE were fatigue and laboratory abnormalities.66 Recently, updates reported on the complete cohort of 50 patients have confirmed the benefit of pembrolizumab in treatment- naive patients with advanced disease. After a median follow-up of 14.9 months, a 56% ORR was observed, regardless of MCPyV status, with 24% achieving a CR and a median PFS of 16.8 months for the entire cohort.67

The efficacy of nivolumab has also been evaluated in metastatic MCC. In the phase I/II CheckMate-358 trial, patients with 5 subtypes of virus-associated advanced cancers (including MCC), who had received =2 prior therapies, were treated with nivolumab 240 mg every 2 weeks until disease progression or unacceptable toxicity. Among 25 treated patients, 60% were treatment naïve; 22 patients were available for radiological response assessment, with ORR of 68%. The ORR were 71% in treatment-naïve patients and 63%. in the pretreated subgroup (1-2 prior systemic therapies), with similar activity in both MCPyV positive and MCPyV negative tumors. At 3 months, PFS and OS rates were 82% and 92%, respectively. Treatment-related AE of any grade occured in 68% of patients and grade 3/4 occurred in 20%.68. In the cohort of 25 patients with stage IIA to IV resectable MCC, neoadjuvant nivolumab was administered intravenously at a dose of 240 mg on D1 and D15 and surgery was planned on D29. Radiological response was observed in 80% of the 20 evaluable patients. Surgery occurred without significant delay in 88% of the patients and, among 17 surgical specimens evaluated for pathological response, 47% presented CR.69

In March of 2017, the Food and Drug Administration approved avelumab to treatment of metastatic MCC based on data published by Kaufman et al, part A of JAVELIN Merkel 200 trial.70 This phase II trial enrolled 88 patients with stage IV MCC chemotherapy- refractory disease. The median age of the patients was 72.5 years, 59% received one previous systemic anticancer treatment, 66% had tumor PD-L1 expression and 52% of tumors were MCPyV positive. Avelumab was given intravenously at a dose of 10 mg/kg every 2 weeks. After median follow up of 16.4 months the ORR was 33.0% with 11.4% reaching CR. Responses to avelumab were observed irrespective of PD-L1 expression or MCPyV status, but ORR was higher in PD-L1 positive (36.2% vs 18.8%) and in MCPyV negative tumors (35.5% vs 28.3%). The 12 month PFS and OS rate were 30% and 52%, respectively. 72.4% of were ongoing at data cutoff, and a median duration of response (DOR) was not yet reached.71 The treatment was well tolerated, with an incidence of grade 3 or higher treatment-related AE of 5%. The most common AE were fatigue and infusion-related reactions.72 Preliminary results from part B of Javelin Merkel 200 trial were presented at the American Society of Clinical Oncology Annual Meeting in 2017.73 This separate cohort included chemotherapy-naïve patients that received first-line avelumab 10 mg/Kg every 2 weeks. In this prespecified interim analysis, cut off on March 2017, 39 of 112 patients were enrolled in safety analysis and 29 of 112 patients in efficacy analysis. The median of follow-up was 5.1 months, the ORR and CR were 62.1% and 13.8%, respectively. Responses were ongoing in 21/29 patients (72.4%) at last report. Among the 14 patients with at least 6 months of follow-up, the confirmed ORR was 71.4% and CR was 28.6%. Among 39 patients evaluable for safety analysis, 71.8% had a treatment-related AE, with 20.5% of grade 3. No grade 4 or treatment-related deaths occurred.74

Unlike other tumors, for which high PD-L1 expression and TMB are well-known predictive biomarkers of response, no biomarker has been proven useful in predicting response to checkpoint inhibitors (CPI) in advanced MCC so far. Georges et al evaluated the association of TMB, PD-L1 expression, MCPyV status and CD8+ tumor-infiltrating T-cell density with ORR and survival. No biomarker alone was predictive of response.75 Therefore, CPI are becoming the new standardof-care for treating patients with metastatic or unresectable MCC. The efficacy of anti-PD-1 and anti-PD-L1 is higher and duration of response is longer than with chemotherapy and new dates of immunotherapy safety have been reported not only in the treatment of MCC but also in the management of other tumors.76 77 78

In the metastatic setting, there are newly designed approaches that are currently in progress. These includes the open label phase II study NCT01758458 [79] that tests the use of viral oncoprotein targeted laboratory treated autologous T-cell therapy together with aldesleukin (IL-2). In this trial, these T-cells comprehend a polyclonal autologous CD8+ cells that recognize MCPyV T-antigen. Combined modality therapies are also under investigation in NCT03071406 [80]; in this regard, the open label, randomized, phase 2 study of nivolumab + ipilimumab +/- stereotactic body radiation therapy (SBRT) at the start of week 2 of treatment (in both arms, nivolumab is administered every 2 weeks and ipilimumab every 6 weeks until progression or unacceptable toxicity) is underway. We eagerly await for the results, given the bad prognosis intrinsically associated to this disease.

In the adjuvant scenario, efforts have been made to address the potential role of immunotherapy. An ongoing phase 2, open-label, randomized trial (NCT02196961) of adjuvant nivolumab or ipilimumab as monotherapy in complete resected MCC, with PFS at 12 months as the primary endpoint81. To date, despite immature data, in this study, ipilimumab arm is already closed due to worst outcomes. There is also high expectancy on an ongoing phase 3, randomized, double blind, placebo controlled study that investigates the role of avelumab in the setting of clinically positive nodal disease after definitive therapy (surgery with/without adjuvant radiation therapy). The results of both trials are expected after 2021.82

Merkel cell carcinoma remains a challenging disease marked by low awareness and numerous diagnostic challenges. Despite optimal upfront therapies that include surgery and radiation therapy, a significant proportion of patients will develop disease recurrence. In this setting, the characterization of the immunogenic potential of MCC has paved the way for the development of immune checkpoint blockade with anti-PD-1 or anti-PD-L1 agents, that now represent the standard of care for patients with irresectable or metastatic disease, leaving the use of CT to those who are not candidates or are refractory to immunotherapy. Ongoing studies addressing the role of these strategies at earlier stages of MCC are under development, and may lead to additional changes in the management of this disease in the near future, translating into even better outcomes for these patients.