Diagnosis and initial staging for MCC

The diagnosis of MCC routinely involves immunohistochemistry (IHC) in addition to histopathological analysis of a tumor biopsy. The histological characteristics of MCC often resemble other small-cell neoplasms such as BCC, small-cell lung cancer, small-cell melanoma, and others;28 29 therefore, the differential diagnosis frequently involves IHC staining for MCC-specific markers. A useful marker for MCC is cytokeratin 20 (CK20),30 31 which is reported to be positive in 87.4% of MCCs.29 CK20 staining in a perinuclear dot-like pattern is highly suggestive of MCC. Positive IHC staining for additional neuroendocrine markers such as synaptophysin,32 neuron-specific enolase,33 neurofilament,34 35 and chromogranin A,36 37 and/or negative staining for thyroid transcription factor-1 (TTF-1)36 can also aid in the differential diagnoses, particularly for CK20-negative MCC.

In 2010, the American Joint Committee on Cancer (AJCC) adopted a consensus system to stage MCC,38 and the Tumor, Node, Metastasis (TNM) system forms the basis of modern MCC staging.22 39 Now in the 8th edition, the updated AJCC staging system was informed by an analysis of 9,387 MCC cases. Noted updates include a distinction between clinical and pathological assessment of nodal involvement, as well as new stage classifications for tumors with nodal disease and metastatic disease with an ‘unknown primary’ (discussed later in this section).22 A summary of the AJCC (8th edition) staging system for MCC can be found in table 1.

The presence or absence of a primary lesion is a clinical factor that affects staging and prognosis. Patients who present with nodal or distant metastases without a primary lesion (unknown primary MCC) have been shown to exhibit significantly improved survival compared with patients with similar extent of disease but with a known primary lesion. Patients with absent primary lesions have tumors with a higher median number of non-synonymous exome mutations, increased MCPyV oncoprotein antibody titers, and they are less likely to be immunocompromised, all factors suggesting a strong immunogenic response.40 For patients who present with a primary tumor, size and invasiveness of the tumor are factors that inform T staging. Another potential prognostic indicator when evaluating the primary tumor is the presence of lymphovascular invasion (LVI), which is associated with more aggressive tumor behavior.41 42

N staging requires an assessment of the lymph nodes for metastatic disease, which can be done clinically (physical exam or imaging) or pathologically. Physical exam and/or imaging studies may identify lesions that raise suspicion for in-transit, nodal, or distant metastases. LVI in the primary tumor is also associated with the presence of sentinel lymph node metastases.43 Generally, suspicious lesions or nodes are sampled to confirm staging pathologically. The identification of distant metastases, either clinically or pathologically, indicates stage IV disease.29

A number of imaging methods are used for staging MCC, including ultrasonography, computed tomography (CT), magnetic resonance imaging (MRI), brain MRI (can be considered for patients with stage IV or symptomatic disease), ⁶⁸Ga-DOTATOC or ⁶⁸Ga DOTATATE PET/CT/NMR (can be considered for patients on somatostatin analogs), and 2-deoxy-2-[¹⁸F] fluoro-D-glucose positron emission tomography (FDG-PET), as well as the combined FDG-PET/CT technique. A retrospective analysis including 492 patients with primary cutaneous MCC and no palpable lymph node enlargement or signs of disseminated disease found that baseline imaging led to upstaging in 13.2% of cases. In the same study, among 92 patients with suspected regional involvement based on palpable lymph nodes, 10 (10.8%) patients were found to have distant metastatic spread and were upstaged to stage IV. PET/CT was more sensitive than CT alone in this cohort, leading to upstaging in 16.8% vs 6.9% of patients who underwent CT alone (p=0.0006).44 The possibility of occult metastatic disease is more common in MCC than in other NMSCs,44 45 and therefore baseline FDG-PET imaging or CT imaging should be considered for all patients with MCC. When a patient is clinically suspected to have nodal or distant metastatic disease, a biopsy and histological confirmation is encouraged.

Sentinel lymph node biopsies should be considered in most patients with clinically localized MCC.46 47 In one analysis, as many as 30.6% of patients (126 of 412) had positive sentinel lymph node biopsies resulting in upstaging, despite no palpable lymph node enlargement, signs or symptoms of disseminated MCC, or evidence of metastases in baseline imaging.44 The results of sentinel lymph node biopsies may guide further management decisions (eg, surgery or radiation therapy). Further guidance on non-immunotherapy treatment decisions should generally follow the most updated treatment guidelines (eg, American Society of Clinical Oncology [ASCO] or National Comprehensive Cancer Network [NCCN]).

MCPyV status

MCPyV is thought to be one of the primary etiological agents responsible for MCC tumorigenesis in the US. Originally identified in 2008,48 MCPyV genomic sequences were present in 80% (8 of 10) of MCC tissue samples and in 16% of control skin tissue. Circulating antibodies against viral oncogenes (T-antigens)49 are specific to patients with MCC with approximate seroprevalence rates of 40.5% in patients with MCC, and as low as 0.9% in healthy controls.50 The presence of MCPyV T-antigen antibodies in serum at baseline and their titer trajectory over the course of treatment may have prognostic implications and utility for surveillance. In patients with detectable titers near the time of diagnosis, higher T-antigen antibody titers have been associated with higher tumor burden.50 Importantly for clinical management, anti-T-antigen antibody levels that decrease after initial treatment and remain low are associated with reduced risk of recurrence, and conversely, an increase in titers is predictive of recurrent disease.

The AMERK test, which measures serum anti-T-antigen antibody levels, is an emerging technology. Although evidence for its utility is not complete, the currently available evidence from prospective studies indicates that AMERK may be useful as an adjunct tool during surveillance. In one cohort of 219 newly diagnosed patients, seropositivity at baseline (blood draws taken within 90 days of diagnosis) was found to be associated with increased frequency of stage II or III disease versus stage I, but also a 42% reduced rate of recurrence (hazard ratio [HR] 0.58; 95% confidence interval [CI] 0.36 to 0.97) compared with seronegative patients.51 In addition, stable or decreasing T-antigen titers had a high negative predictive value for clinically evident recurrence, and therefore, a reduction in surveillance imaging frequency may be considered for patients with consistently stable or downtrending titers. Due to the implications for surveillance and disease recurrence, baseline AMERK testing should be considered in all patients with MCC (ideally within 3 months of diagnosis and prior to surgery or definitive radiation, if possible), as well as periodic serial testing to complement radiological surveillance in patients who are seropositive at baseline (for more information on MCPyV serology testing in surveillance, see the Response monitoring and surveillance section).

Testing MCC tumor biopsy samples for MCPyV is not routinely performed as there is currently no Clinical Laboratory Improvement Amendments (CLIA)-approved test and tumor viral status has yet to be proven relevant to clinical management. Research methods to determine tumor viral status include large T-antigen IHC (the CM2B4 antibody is most commonly used), quantitative PCR, and next-generation sequencing (NGS).21 A positive serum AMERK test likely indicates the tumor is MCPyV-positive.

Panel recommendations

• Histopathological diagnosis of MCC should include hematoxylin and eosin (H&E) and IHC with evaluation by a pathologist with experience in the diagnosis of skin cancers, when feasible.

• For patients with MCC, multidisciplinary care team management is optimal.

• Baseline imaging should be considered in all patients with MCC for the detection of metastatic disease (LE:4).

• Baseline MCPyV oncoprotein serology testing (eg, AMERK) should be considered, as this has implications for surveillance (LE:3).

• If, by clinical examination or imaging, a patient is suspected of having nodal or distant metastatic disease, a biopsy should be performed for histological confirmation (LE:3).

• If imaging and clinical evaluation are negative for metastatic disease, patients should be considered for sentinel lymph node biopsy for staging (LE:1).

Approved anti-PD-(L)1 agents for MCC

Avelumab became FDA-approved in March 2017 for the treatment of metastatic MCC.6 The accelerated approval was based on the phase II, open-label, international study, JAVELIN Merkel 200 (NCT02155647), which enrolled a total of 204 patients into two arms: part A included patients with metastatic MCC whose disease had progressed after first-line chemotherapy, and part B included patients with metastatic or distally recurrent MCC who had not undergone prior systemic treatment (with the exception of adjuvant chemotherapy given with no clinically detectable or metastatic disease ≥6 months prior to study start). Patients in both arms received avelumab at 10 mg/kg intravenously (IV) every 2 weeks until disease progression or intolerable adverse events (AEs) occurred, with ORR being the primary endpoint measured by Response Evaluation Criteria in Solid Tumors v1.1 (RECIST v1.1). Approval was based on efficacy results from 88 patients in part A after the final patient enrolled had completed 12 months of follow-up (table 2). At longer follow-up (median 40.8 months; range 36.4─49.7), the ORR was maintained, with complete response (CR) achieved in 10 patients (11.4%) and partial response (PR) in 19 (21.6%). The median duration of response (DOR) was 40.5 months (95% CI 18 to not estimable [NE]), and at ≥44 months follow-up (as of May 2019) the OS was 12.6 months (95% CI 7.5 to 17.1).53 Responses were seen early, with 20 of 29 patients having confirmed responses by week 7, and 27 with a response by week 13.54 A rapid response was associated with improved survival. The 20 patients with an objective response at 7 weeks had a significantly improved 18-month survival probability of 90% (95% CI 65.6% to 97.4%), compared with 26.2% (95% CI 15.7% to 37.8%) for those without a rapid response. Significant differential survival probabilities were also seen in patients with objective responses at 13 weeks. PFS rates of the 88 patients in part A at 1 year, 2 years, and 3 years were 30% (95% CI 21% to 41%),55 26% (95% CI 17% to 36%), and 21% (95% CI 12% to 32%),53 respectively. In addition, patients treated with avelumab have reported experiencing improved QOL associated with less disease progression.56

In part B of the JAVELIN Merkel 200 study, the primary analyses in 116 total patients with a median follow-up of 21.2 months (range 14.9─36.6) reported an ORR of 39.7% (table 2) with 16.4% (n=19) experiencing a CR and 23.3% (n=27) a PR.57 The study also reported a durable response rate, defined as a CR or PR lasting ≥6 months, of 30.2% (95% CI 22.0% to 39.4%). Responses were observed by 3 months in a majority of patients, with a median time to response of 6.1 weeks (range 5–36). The median PFS was 4.1 months (95% CI 1.4 to 6.1). Median OS was 20.3 months (95% CI 12.4 to NE), and 6- and 12-month OS rates were 75% (95% CI 66% to 82%) and 60% (95% CI 50% to 68%), respectively.

Safety results from JAVELIN Merkel 200 demonstrate avelumab to be generally tolerable for patients with MCC. At follow-up analyses of ≥36 months, treatment-related adverse events (TRAEs) of any grade were reported in 68 of the patients in the part A arm (77.3%), with 10 patients (11.4%) experiencing a TRAE of grade ≥3. Immune-related adverse events (irAEs) were documented in 19 patients (21.6%), with 4 patients (4.5%) experiencing grade ≥3 irAEs. Overall, 8 patients (9.1%) had to cease treatment due to TRAEs, and no treatment-related deaths occurred.53 Similar safety profiles were seen in the part B arm, where 94 out of 116 total patients (81%) had any-grade TRAEs; 20 patients (17.2%) experienced a grade 3 TRAE, 1 patient experienced a grade 4 TRAE (dermatitis psoriasiform), and treatment was discontinued in 14 (12.1%) patients.57 Toxicities determined to be irAEs were seen in 35 (30.2%) patients, and 7 (6%) experienced a grade ≥3 irAE.

In December 2018, pembrolizumab was granted accelerated approval by the FDA for the treatment of adult and pediatric patients with recurrent, locally advanced, or metastatic MCC.8 Approval was based on tumor response rates and DORs in the KEYNOTE-017 trial (NCT02267603), which enrolled 50 subjects with locally advanced and metastatic MCC to receive 2 mg/kg IV pembrolizumab every 3 weeks as their first systemic treatment. All patients enrolled had unresectable stage IIIB (14%, n=7) or stage IV (86%, n=43) disease. Tumor responses were measured by RECIST v1.1 criteria. At a median follow-up of 14.9 months (range 0.4─36.4+), the ORR was 56% (95% CI 41.3% to 70.0%), with 12 CRs (24%; 95% CI 13.1% to 38.2%) and 16 PRs (32%; 95% CI 19.5% to 46.7%),58 and at a longer median follow-up of 31.8 months, the ORR was 58% (95% CI 43.2% to 71.8%).59 Responses were both rapid and durable; the median time to response was 2.8 months (range 1.5─9.7)58 and median DOR was not reached at 3 years (range 1.0+ to 51.8+ months).59 The 2- and 3- year OS rates were 68.7% and 59.4%, and PFS rates were 48.3% and 39.1%, respectively.58 59 Pembrolizumab was generally safe and well tolerated. TRAEs of any grade occurred in 96% of patients (n=48), with grade ≥ 3 experienced in 28% (n=14). Seven patients (14%) discontinued therapy due to TRAEs.58 A phase III confirmatory trial evaluating pembrolizumab for the treatment of patients with advanced MCC (KEYNOTE-913, NCT03783078) was ongoing at the time of guideline publication.

Locally advanced or metastatic MCC is often challenging to treat based on the anatomical location of disease, previous interventions, concomitant immunomodulation for autoimmune disease, a history of solid organ transplant, or the presence of hematological malignancies. Considerations for these challenging-to-treat populations are discussed further in the Patient selection for immunotherapy treatment of MCC section, and duration of treatment as well as surveillance are discussed in the Response monitoring and surveillance section. Generally, a multidisciplinary approach to care is optimal, with consideration of referral to an academic or high-volume center for best outcomes in particularly challenging cases. Enrollment in clinical trials for patients with contraindications to on-label indications for ICIs or with refractory/resistant disease is also important as trial settings may provide patients with access to potentially life-saving care that would be impossible to obtain in routine practice, in addition to continuing to make progress in identifying additional options in this heretofore largely untreatable disease60 61 (for more discussion of options for ICI-refractory or ICI-resistant disease, see the Anti-PD-(L)1 resistance section).

Immune biomarkers for response to ICIs for MCC

At the time of guideline publication, viral status does not have an established clinical utility for predicting benefit with ICIs. Neither KEYNOTE-017 nor JAVELIN Merkel 200 found a significant association between response rates to ICIs and MCPyV viral status. In KEYNOTE-017, the ORRs with pembrolizumab for patients with MCPyV-positive and MCPyV-negative tumors were 59% (19 of 32 patients) and 53% (9 of 17 patients), respectively (p=0.765).58 In cohort A (patients with metastatic MCC with disease progression after at least one prior line of chemotherapy) of JAVELIN Merkel 200, ORRs with avelumab in patients with MCPyV-positive tumors were 28.3% (95% CI 16.0% to 43.5%; n=46) and 35.5% (95% CI 19.2% to 54.6%; n=31) for patients with MCPyV-negative tumors.53 A similar trend was seen in cohort B (patients with metastatic MCC that was chemotherapy-naïve).57

Other immune-related biomarkers with well-characterized predictive power for benefit with checkpoint blockade in some solid tumor types include PD-L1 expression on tumor cells (TCs) and/or immune cells (ICs), high tumor mutational burden (TMB-H), high microsatellite instability (MSI-H), and IC infiltration. Although several of these immune biomarkers have been included in exploratory analyses, currently, no validated biomarkers exist to predict response to ICIs for patients with MCC. Efforts are ongoing, however, to develop reliable multifactor prediction tools.

PD-L1 expression in the tumor microenvironment has been correlated with increased response rates in multiple disease settings, and confirmed ‘positivity’ by an approved companion diagnostic is indicated for on-label use of ICIs in some cancers, such as breast cancer, lung cancer, urothelial cancer, and others.8 62 63 MCC is generally considered an immunologically inflamed tumor, with frequent PD-L1 expression and IC infiltration. One analysis reported PD-L1 positivity rates of 49% for TCs and 55% for ICs in MCC diagnostic pathology specimens across stages.17

In the registrational trials for ICIs in MCC, no association was seen between tumor PD-L1 expression and ORRs. In KEYNOTE-017, pretreatment PD-L1 expression was assayed by IHC using the anti-PD-L1 22C3 antibody, and ‘PD-L1-positive’ tumors were defined as tumors with positive PD-L1 staining on ≥1% of TCs or ICs. Trends were observed toward increased OS (p=0.057) and PFS (p=0.128) in patients with >1% PD-L1 TC expression; however, these analyses were exploratory in nature.58 The JAVELIN Merkel 200 trial also found that ORR was not significantly different between patients with PD-L1-positive (PD-L1 expression in ≥1% of TCs) and PD-L1-negative tumors, as scored by a proprietary PD-L1 IHC assay (using the 73-10 anti-PD-L1 antibody), although similar trends for better response and OS rates were seen in the small numbers of patients who had PD-L1-positive tumors. In cohort A of the study (patients with metastatic MCC with disease progression after at least one prior line of chemotherapy), of the 22 patients with long-term OS (defined as >36 months) after treatment, 81.8% had PD-L1-positive tumors, whereas only 18.2% had PD-L1-negative tumors.53 Although the study was not powered to compare PD-L1 expression subgroups, the data suggest that PD-L1 expression may be more prognostic than predictive in MCC. Further research is needed to identify and validate biomarkers for immunotherapy efficacy in MCC.

Several of the registrational trials for ICIs in MCC have included translational studies on tumor-infiltrating ICs,53 57 64–66 which generally portend a more favorable prognosis with higher levels of IC infiltration. However, these measures are not yet validated or readily available as predictive biomarkers for response.

Because the main etiological factors underlying tumorigenesis are either exposure to UV radiation or MCPyV infection, the mutational landscapes of MCC tumors, although heterogeneous, tend to cluster around two distinct signatures20 67: TMB-H MCC tumors (characteristic of UV exposure) tend not to harbor MCPyV genomes and vice versa. The mutation frequencies of MCPyV-positive tumors have been reported as 20- to 90-fold lower than for MCPyV-negative tumors.67 68 The mutational landscape of MCPyV-negative tumors shows strong UV signatures.20 68 69 TMB-H has also been correlated with improved response to immunotherapies in multiple types of solid tumors70 and is an approved tissue-agnostic eligibility criterion for treatment with pembrolizumab monotherapy8 based on improved ORR and DOR in patients with tumors harboring 10 or more mutations/megabase (Mb) in KEYNOTE-158 (for further discussion, see the Tissue-agnostic indications for ICIs section).71 The role of TMB as a predictor for efficacy is thought to be, in part, due to increased neoantigen presentation eliciting antitumor cytotoxic T cell responses.11 In MCC, however, data are still lacking on whether TMB-H predicts improved outcomes with ICI therapy, although trends for numerically higher responses have been reported.53 57 Despite the distinct genomic landscapes associated with individual etiologies, outcomes with ICIs have been similar across viral and non-viral MCCs in published trials.20 53 The published cohorts to date have been small, however, and further research is needed to understand determinants of response to ICIs in MCC.

Panel recommendations

• For patients with metastatic, recurrent, or locally advanced MCC that is not amenable to curative surgery or radiation and for whom no contraindications to immunotherapy are present, first-line therapy with an approved anti-PD-(L)1 ICI is recommended (LE:2).

• For patients with MCC who experience disease progression while on anti-PD-(L)1 immunotherapy, therapeutic options are limited and include clinical trials or chemotherapy. Switching treatments from one anti-PD-(L)1 antibody to another anti-PD-(L)1 antibody is unlikely to be beneficial.

• While responses to PD-(L)1 blockade are frequent and generally durable, disease progression may occur in a subset of responders, and hence, continued surveillance for MCC progression is warranted.

• For patients with MCC being treated with anti-PD-(L)1 ICIs, there are no validated biomarkers to predict benefit (including MCPyV viral status or PD-L1 expression).

Investigational immunotherapies for MCC in the neoadjuvant and adjuvant settings

The risk of recurrence in MCC is high, and thus, neoadjuvant and adjuvant therapies have been active areas of study, with adjuvant cytotoxic chemotherapy showing no discernable survival benefit at any stage of disease.72 73 The benefit of immunotherapy in earlier stages of MCC has not yet been established. Importantly, despite favorable reports of safety and efficacy, the potential for disease progression during neoadjuvant therapy, possibly to the point where the tumor becomes unresectable, is a major concern.

CheckMate 358 (NCT02488759), a multicohort phase I/II trial, demonstrated a pathologic complete response (pCR) rate of 47.2% (17 of 36) in patients with MCC who received 2 cycles of nivolumab followed by surgery.74 Median recurrence-free survival (RFS) and OS had not been reached at a median follow-up of 20.3 months. The safety profile was acceptable, with TRAEs occurring in 46.2% of patients and grade 3–4 TRAEs in 7.7% of patients. Three patients (7.7%) were unable to undergo surgery due to disease progression (n=1) or AEs (n=2). Further studies including comparator groups with adjuvant ICIs and resection alone are needed to understand if neoadjuvant therapy improves outcomes for patients with resectable MCC. Clinical trials are the most appropriate context for these approaches at this time.

Several trials are ongoing to investigate whether ICIs used in the adjuvant setting can reduce the risk of recurrence. No improvement in disease-free survival (DFS) (HR 1.8; 95% CI 0.3 to 10; p=0.48) and significantly increased incidence in AEs was seen in 40 patients with completely resected MCC who were treated with adjuvant anti-CTLA-4 ipilimumab in the ADMEC-O study (NCT02196961).75 An additional arm of ADMEC-O investigating adjuvant nivolumab for completely resected MCC tumors is ongoing (NCT02196961). Additional trials are comparing avelumab as an adjuvant therapy against placebo, such as the phase III ADAM trial (NCT03271372) and the phase II I-MAT trial (NCT04291885). Finally, several studies are combining adjuvant radiation therapy with ICIs, including the phase III STAMP study (NCT03712605), which is evaluating pembrolizumab plus optional radiation therapy versus observation plus optional radiation therapy after complete resection in patients with stage I–III MCC. Importantly, however, the risks of adjuvant therapy include increased AEs and treatment-related mortality, and there is no direct evidence to support the routine use of radiation with ICIs outside of a clinical trial setting at this time.

Investigational immunotherapies for MCC in the advanced/metastatic setting

The incorporation of the approved anti-PD-(L)1 ICIs into the standard of care has improved outcomes for patients with advanced unresectable MCC. Building on these early successes, studies are under way to investigate additional immunotherapies for advanced unresectable MCC.

While avelumab and pembrolizumab are the only FDA-approved ICIs for the treatment of MCC at the time of guideline development, ongoing studies are evaluating other anti-PD-(L)1 agents. For example, early results (median follow-up of 26 weeks [range 5─35]) from the phase I/II CheckMate 358 trial evaluating systemic nivolumab for the treatment of advanced virus-associated tumors after two or fewer prior therapies, found an ORR of 68% (95% CI 48% to 86%; n=15) among the 22 patients with MCC, with ongoing responses in 13 patients.76 Although some guidelines recommend nivolumab or other anti-PD-(L)1 ICIs for advanced MCC,27 their use is off-label at the time of guideline preparation. Based on the known mechanisms of anti-PD-(L)1 therapies and response rates in published trials, however, there is no indication that nivolumab or other off-label anti-PD-(L)1 agents differs substantially from approved agents in efficacy and tolerability.

Several studies are investigating immune-modulatory agents targeting the innate and adaptive immune systems, with and without ICIs as a combination regimen, to enhance antitumor responses. Some of these strategies include local interleukin (IL)-12, IL-15 superagonists, IL-17, oncolytic viruses, radiation therapy, and various toll-like receptor (TLR) agonists.

Immune evasion mechanisms have been demonstrated to be active in patients with MCPyV-positive MCC tumors, including MCPyV-specific CD8⁺ T cell dysfunction due to high expression of the inhibitory receptors PD-1 and T cell immunoglobulin and mucin domain-containing protein 3 (TIM-3),77 and downregulation of class I human leukocyte antigen (HLA-I).78 To overcome these endogenous mechanisms, several clinical trials are using adoptive cell therapies to expand autologous MCPyV-specific cytotoxic T cells to treat patients in combination with recombinant interferon-β (IFN-β) and radiation therapy,79 80 both shown to upregulate HLA-I expression.78 81 Several allogenic cell therapies are also being studied using both virus-specific and other IC transfers.

Treatment options for patients with MCC whose disease has progressed on prior anti-PD-(L)1 therapies (discussed further in the Anti-PD-(L)1 resistance section) are also being studied. Of note, trials under way include the CARTA study (NCT04792073) assessing comprehensive ablative radiation therapy with or without avelumab, and the DUET-1; 02 study (NCT04590781) using a bispecific somatostatin receptor 2 (SSTR2) and CD3 binding antibody, tidutamab. In addition, two retrospective studies that treated patients’ MCC with salvage anti-PD-1 nivolumab in combination with anti-CTLA-4 ipilimumab in patients who have progressed on prior anti-PD-(L)1 therapy have reported responses in 3 of 582 and 4 of 1383 patients. Additionally, clinical trials assessing cell therapy combinations for patients with anti-PD-(L)1 refractory disease are also under way, though most are in early phases (NCT03747484).

Panel recommendations

• For patients with resectable MCC at high risk of recurrence, enrollment in clinical trials of neoadjuvant or adjuvant therapy should be offered where available.

• For patients with unresectable, locally advanced, metastatic, or recurrent MCC, enrollment in clinical trials of novel agents and novel combinations should be offered where available.

• Patient selection for immunotherapy treatment of MCC

Immunotherapy should be considered within the broader context of available therapies for MCC. Both of the currently FDA-approved immunotherapies for MCC, avelumab and pembrolizumab, are indicated specifically for patients with advanced forms of the disease.6 8

For those patients with unresectable tumors or with contraindications to surgery or radiation, systemic therapy should be considered. Reasons to avoid surgery and/or radiation include inability to achieve resection with microscopically negative margins, medically unfit patients for surgical or radiation intervention, or treatment-related morbidity that would significantly impact the patient’s QOL. Of the systemic therapy options, chemotherapy (as discussed in the Recommended immunotherapies for MCC section) has not demonstrated durable benefit as an initial treatment for advanced MCC, with high initial response rates but poor PFS,84 nor as an adjuvant therapy at any disease stage of MCC.73 ICIs have become the new standard for patients with advanced MCC, often with long-lasting responses.

Special patient populations for MCC

In addition to the general label indications to determine which patients with MCC are candidates for the approved ICIs, special considerations need to be taken for certain patient populations, especially those with altered or suppressed immune systems at baseline. Patients with HIV have typically been excluded from clinical trials of ICIs due to a lack of data on how ICIs would interact with altered immune systems. This is an important population to consider, however, as HIV-positive individuals are at elevated risk of developing MCC, presumably due to their immunocompromised state.85 Although data on the safety of ICIs specifically in HIV-positive patients with MCC are lacking, a systematic review of 73 patients with HIV and a variety of solid tumors demonstrated that ICIs were well tolerated and efficacious, with ORRs and frequencies of grade 3–4 irAEs similar to those found in patients without HIV. Furthermore, of 28 patients with no detectable HIV load prior to ICI therapy, HIV remained controlled in 26 (93%) after treatment with checkpoint inhibition.86 Overall, patients with HIV can be treated with ICIs, ideally with the viral status well controlled (ie, CD4 count ≥100 cells/µL, HIV viral load <200 copies/mL, and on antiretroviral therapy), however, treatment for patients with viral status that does not meet these parameters can still be considered.

Other patient populations that have frequently been excluded from clinical trials of ICIs include patients with pre-existing autoimmune disease and solid organ transplant recipients, due to the possibility that ICI-mediated immune activation could lead to flares of autoimmunity or graft rejection, respectively. A retrospective study of 30 patients with melanoma and an existing autoimmune disease treated with ipilimumab found an ORR of 20%, with grade 3–5 irAEs in 33% of patients. Although 27% of patients experienced a flare of their autoimmune disease, all of the flares were manageable with corticosteroids.87 A subsequent systematic review that included 123 patients with pre-existing autoimmune disease and a variety of tumor types treated with anti-PD-(L)1 and/or anti-CTLA-4 ICIs found that 41% of patients had an exacerbation of their autoimmune disease, 25% developed de novo irAEs, and 9% experienced both.88

ICI therapy administered to solid organ transplant recipients may be efficacious but carries a high risk of allograft rejection, as demonstrated in retrospective analyses. One report describing 39 solid organ transplant recipients with various solid tumors who received ICIs found that 41% (n=16) experienced allograft rejection, leading to graft loss in 81% (n=13). Among the 22 patients with melanoma and 5 with CSCC in the study, allograft rejection occurred in 8 patients with melanoma and 2 patients with CSCC. Favorable antitumor responses were seen in 36% of patients with melanoma, and all patients with CSCC achieved a CR or a PR.89 A more recent pooled analysis described similar results, with a rate of allograft rejection at 41% in a total of 64 patients.90 Currently, NMSCs such as MCC lack tumor-specific data on the risk of graft rejection and successful management strategies. Therefore, ongoing and future clinical trials, such as NCT03816332, which is investigating the use of ICIs in combination with standardized transplant immunosuppression in kidney transplant recipients diagnosed with unresectable or metastatic cutaneous malignancies, will be critical in better understanding how to optimize the safety and efficacy of immunotherapy in these patient populations.

Patients with chronic lymphocytic leukemia (CLL) are immunocompromised in addition to being at increased risk of developing MCC.91 Patients with some NMSC and concurrent hematological malignancies may have worse outcomes with ICI treatment;92 however, prospective data are lacking at this time and trials are ongoing. A small number of reports have demonstrated general safety using ICIs alone or in combination with ibrutinib.93 94

Some patients who are being considered for immunotherapy are on immunosuppressive drugs, such as corticosteroids, to treat pre-existing autoimmune conditions. Corticosteroids suppress the immune system, and steroid exposure is independently associated with worse survival outcomes in advanced solid tumors.95 However, corticosteroids are also a cornerstone of irAE management, and studies to date have not conclusively demonstrated that short-term (ie, limited duration for the purposes of toxicity management and promptly weaned when AEs resolve) immunosuppression affects clinical benefit with ICI therapy. A systematic review of 4,045 patients with a variety of tumor types found that patients taking high-dose steroids (>10 mg prednisone or equivalent) were at increased risk of death or disease progression compared with patients who were not. This increased risk was greatest in subgroups of patients who were taking steroids for supportive care or for brain metastases. Notably, patients who received steroids to treat irAEs stemming from ICI therapy did not exhibit reduced OS.96 Similar results were found in a study of patients with non-small-cell lung cancer, where patients who received ≥10 mg prednisone at the time of beginning immunotherapy had worse median OS and PFS if receiving the prednisone for palliative care, but those receiving ≥10 mg prednisone for cancer-unrelated reasons (such as prior autoimmune conditions) experienced OS and PFS similar to those who received <10 mg prednisone.97 While these studies demonstrate that the context of steroid use is important, in general, for patients with immunological comorbidities (other than solid organ transplant preservation) who are taking high-dose corticosteroids and/or other immunosuppresives, a reduction to the minimum necessary dose is best practice prior to treatment with ICIs, and enrollment in clinical trials is encouraged if available. A skin cancer multidisciplinary tumor board or referral to a tertiary care center with a tumor board for evaluation of clinical trial eligibility may be beneficial for these difficult-to-treat patient populations.

Currently, there are limited data on the safety of pregnancy and ICI therapies (currently considered category D by FDA for pregnancy). In a review of available case studies (n=7) on the safety of ICI use in pregnant patients with melanoma, complications were observed in 71.4% of patients, and the average gestational age of delivery (n=9) was 30.4 weeks (range 24–38).98 In addition to considering the effects of ICIs on pregnancy, it is also important to consider the potential need for immunosuppressants for the treatment of irAEs, some of which, such as mycophenolate mofetil, have been shown to cause fetal malformation in the context of solid organ transplant.99 100 In addition, data are lacking on the effects of ICIs on fertility in women of childbearing age. Because endocrine irAEs may arise and persist long term after cessation of therapy, a referral to an onco-fertility specialist may be considered.

Panel recommendations

• For solid organ transplant recipients with MCC who are potential candidates for ICI therapy, healthcare professionals should have a careful risk–benefit discussion with the patient that includes the potential risk of allograft loss, which may result in the need for dialysis (ie, for kidney transplant recipients) or death (eg, recipients of heart, lung, or liver transplants) (LE:1). Chemotherapy is an acceptable alternative first-line therapy in this population (LE:3). Enrollment in clinical trials should be encouraged.

• For patients with MCC on therapeutic immune suppression, a discussion should be initiated on reducing or modifying immune suppression, if appropriate, before proceeding with ICI therapy.

• For patients with MCC who are receiving high-dose corticosteroids for reasons other than solid organ allograft preservation, the dose of corticosteroids should be reduced to ≤10 mg prednisone (or equivalent) per day, if possible, prior to initiation of ICI therapy.

• Caution is advised when treating with ICIs in patients with MCC who have a history of autoimmune disease, as they may experience a flare of their autoimmune disease and/or a distinct irAE. Healthcare professionals should have a careful risk–benefit discussion with the patient, which includes the potential risks associated with an autoimmune flare (LE:1).

Panel recommendations

• Histopathological diagnosis of advanced CSCC should include H&E evaluation by a pathologist with experience in the diagnosis of skin cancers, when feasible.

• For patients with advanced CSCC, multidisciplinary care team management is optimal.

• Patients with CSCC who are being considered for systemic therapy should be evaluated by appropriate radiological imaging.

Approved anti-PD-1 agents for CSCC

Two anti-PD-1 ICIs have been approved by the FDA for the treatment of CSCC: pembrolizumab and cemiplimab (figure 1). Key outcomes from the landmark trials leading to the FDA-approved ICI indications for CSCC are summarized in table 5. Importantly, there is no evidence of clinical benefit thus far to favor one ICI over the other.

In September 2018, the FDA approved cemiplimab for the treatment of patients with metastatic or locally advanced CSCC who are not candidates for curative surgery or radiation.7 Cemiplimab was the first approved ICI for CSCC, with a regulatory decision based on data from the phase I dose-escalation Study 1423 (NCT02383212) and the ongoing phase II open-label EMPOWER CSCC 1/Study 1540 (NCT02760498). The trials initially enrolled a total of 108 patients with metastatic, recurrent, or locally advanced CSCC who were not candidates for curative surgery or radiation for cemiplimab treatment at 3 mg/kg IV every 2 weeks (an additional 56-patient cohort received cemiplimab at 350 mg IV every 3 weeks [the approved dosage]). Patients in both trials had commonly received prior surgery (96% of patients) or radiotherapy (79% of patients), and 50% of patients had received prior systemic therapy. The approval was based on ORR and DOR data at a median follow-up of 8.9 months in the combined metastatic and locally advanced cohorts.

Follow-up analyses of EMPOWER CSCC 1/Study 1540 found an ORR of 50.8% (95% CI 37.5% to 64.1%) with a median duration of follow-up of 18.5 months in the metastatic cohort, an ORR of 44.9% (95% CI 33.6% to 56.6%) with a median duration of follow-up of 15.5 months in the locally advanced cohort, and an ORR of 42.9% (95% CI 29.7% to 56.8%) with a median follow-up of 17.3 months in the fixed dosage cohort, for a combined ORR from all three groups (n=193) of 46.1% (95% CI 38.9% to 53.4%, median duration of follow-up 15.7 months).110 CRs were achieved in 16.1% of patients, and PRs in 30.1% of patients. The median observed time to response across cohorts was 2.1 months (IQR 1.9–3.7). Although median DORs were not reached at the time of reporting, the estimated rates of patients with observed DORs of ≥12 months and ≥24 months were 89.5% and 68.8% in the metastatic cohort and 83.2% and 62.5% in the locally advanced cohort, respectively.

The most recent safety data available at the time of guideline publication reported that grade ≥3 TRAEs occurred in 17.1% of patients.110 Sponsor identified irAEs were reported in 29.5% of patients and at least one grade ≥3 irAE in 9.3%.

Pembrolizumab was approved by the FDA in June 2020 for the treatment of patients with recurrent or metastatic CSCC not curable by surgery or radiation.8 Approval was based on ORR and DOR in the phase II, open-label KEYNOTE-629 trial (NCT03284424). Of the 105 patients enrolled and treated with pembrolizumab at 200 mg IV every 3 weeks for up to a maximum of 24 months, 74.3% had prior radiation therapy, 86.7% had previously received at least one systemic therapy, and only 13.3% were receiving pembrolizumab as their first line of treatment.111 With an encouraging ORR reported at the time of approval (median follow-up of 9.5 months, table 5), the median DOR was not reached and 69% of patients had a median DOR of ≥6 months. In July 2021, the approval for pembrolizumab was expanded to also include patients with locally advanced CSCC not curable by surgery or radiation, based on an interim analysis of this cohort of patients in KEYNOTE-629.8

In an interim analysis of KEYNOTE-629 with longer follow-up (14.9 months [IQR 12.6–17.2] for the locally advanced cohort and 27.2 months [IQR 25.6─29.2] for the recurrent/metastatic cohort),112 the ORR for patients with locally advanced CSCC was 50.0% (95% CI 36.1% to 63.9%) with a CR rate of 16.7% and a PR rate of 33.3%. Median time to response was 2.6 months (IQR 1.4–3.6), with responses reported in 14 of the 27 responders by week 6. Median PFS and median DOR were not reached with an estimated 88.1% and 84.1% of responders having responses that lasted at least 6 and 12 months, respectively. Of the patients with baseline and postbaseline imaging available (n=48), 83.3% had a reduction in target lesion size, including 66.7% with at least a 30% reduction. Median OS was not reached, and the estimated OS rate at 18 months was 73.6% (95% CI 59.5% to 83.4%).

In the recurrent/metastatic cohort of KEYNOTE-629, the ORR was 35.2% (95% CI 26.2% to 45.2%), with a CR rate of 10.5% and a PR rate of 24.8%. The ORR was higher in patients who received pembrolizumab as first-line therapy (n=14) compared with patients who received pembrolizumab as second-line or later therapy (n=91; 50.0% vs 33.0%, respectively). The overall median time to response was 1.6 months (IQR 1.4–13.1), with responses reported in 20 of the 37 responders by week 6. Median PFS was 5.7 months (95% CI 3.1 to 8.5) and median DOR was not reached (95% CI 22.4 months to not reached) with estimates of 80.7% of responses lasting at least 6 months and 77.8% lasting at least 12 months. Among the 95 patients who had baseline and postbaseline evaluable images, 77.9% demonstrated a reduction in target lesion size, and 58.9% had at least a 30% reduction in target lesion size. Median OS was 23.8 months (95% CI 13.4 to 29.8), and the OS rates at 12 and 24 months were 61.0% (95% CI 50.9% to 69.5%) and 48.4% (95% CI 38.5% to 57.6%), respectively. Biomarker data from this study reported response to pembrolizumab occurred independent of PD-L1 status of tumors (discussed in the Immune biomarkers for response to ICIs for CSCC section). Additionally, the phase II, open-label CARSKIN trial (NCT02883556) is examining the efficacy and safety of pembrolizumab as a first-line therapy for unresectable locally or regionally advanced or metastatic CSCC. Results from 39 patients found a median PFS of 6.7 months and median OS was 25.3 months (95% CI 14.2 months to NE) at a median follow-up of 22.4 months. At 15 weeks, the response rate was 41% (95% CI 26% to 58%), with 3 CRs and 13 PRs. TRAEs occurred in 71% of patients, and 7% of patients experienced serious TRAEs.113

Pembrolizumab was found to be generally tolerable in KEYNOTE-629. TRAEs were observed in 69.2% of patients, with 11.9% experiencing grade ≥3 and 8.8% requiring a discontinuation of treatment. Infusion reactions or irAEs occurred in 22.6% of patients and were mostly grade ≤2, but 8.2% of patients experienced a grade ≥3 irAE or infusion reaction.112

While there are differences in the patient populations evaluated in the registrational studies (described in table 5), either cemiplimab or pembrolizumab is appropriate first-line therapy for patients with metastatic or locally advanced CSCC who are not candidates for curative surgery or radiation. In the published trials, tumor regression typically occurred within the first 2 months, but in some cases, delayed responses can occur and it may be appropriate to continue treatment if the patient is clinically stable and experiencing clinical benefit. Duration of therapy in both studies was not longer than 2 years, by design, and optimal duration of treatment is not known.

Similar to MCC, patients with locally advanced or metastatic CSCC may present with comorbidities that complicate an immunotherapy treatment plan. Considerations for challenging populations are discussed in the Patient selection for immunotherapy treatment of CSCC section, and duration of treatment and surveillance are discussed in the Response monitoring and surveillance section. Generally, a multidisciplinary approach to care is essential, with consideration for referral to an academic or high-volume center for best outcomes in particularly challenging cases. Enrollment in clinical trials for patients with contraindications or with refractory/resistant disease to ICI treatment is also important (see the Anti-PD-(L)1 resistance section)—not only because a trial setting may provide patients with access to potentially life-saving care that would be impossible to obtain in routine practice, but also to continue to make progress in identifying additional options in this heretofore largely untreatable disease.60 61

Immune biomarkers for response to ICIs for CSCC

Currently, there are no validated biomarkers to predict immunotherapy-related treatment response or inform clinical decisions for patients with CSCC, including PD-L1 expression or TMB. Tumor PD-L1 expression did not significantly correlate with response to PD-1 blockade with pembrolizumab in KEYNOTE-629, although there was a trend towards better ORR in tumors with a PD-L1 combined positive score (CPS, the number of PD-L1 staining cells [TCs, lymphocytes, and macrophages] divided by the total number of viable TCs, multiplied by 100) ≥1 in the combined analysis of patients with metastatic and locally advanced CSCC.112 Patients with tumors with PD-L1 CPS ≥1 (n=115) had an ORR of 42.6% (95% CI 33.4% to 52.2%), whereas for those with PD-L1 CPS <1 tumors (n=15), the ORR was 20% (95% CI 4.3% to 48.1%). In a separate analysis of response rates stratified by tumor proportion score (TPS, the number of positive TCs divided by the total number of viable TCs, multiplied by 100), similar ORRs in the PD-L1-positive group were seen when the threshold for positivity was 50%. In addition, PD-L1 expression in CSCC has been shown to correlate with risk for metastasis.15 114 115 Overall, the evidence at this time indicates that individuals with PD-L1-negative tumors may still derive benefit from the approved immunotherapies, and therefore all patients eligible for systemic immunotherapy should receive an ICI regardless of PD-L1 expression.

The value of TMB as a predictive or prognostic biomarker for patients with CSCC is an ongoing area of study. CSCC tumors have among the highest levels of TMB in solid malignancies, with a damage pattern characterized by UV signature mutations. One study including patients with a variety of squamous cell carcinomas from several tissue types found that both cutaneous origin and increasing TMB were associated with better outcomes with anti-PD-(L)1 blockade (monotherapy or in combination regimens). The link between TMB and clinical benefit with immunotherapy for CSCC specifically was not statistically significant in multivariate analysis; however, these data are difficult to interpret due to the small sample size.116

Owing to the high TMB of CSCC, identifying individual driver genes for prognostic use has been a challenge, and single-gene tests have had little utility thus far. However, the use of gene expression panels to profile transcriptional activity in tumors is an ongoing area of research to attempt to identify predictive or prognostic signatures. A commercially available gene expression panel (DecisionDx-SCC) uses a panel of 40 genes (including six control loci) to stratify primary tumors into low-risk, high-risk, and highest-risk categories for prediction of metastasis. The assay has a reported positive predictive value of 60% for patients with class IIB (highest risk) disease117; however, there are no prospective studies of its clinical utility, and thus, its use is not recommended at this time.

Panel recommendations

• For patients with metastatic, recurrent, or locally advanced CSCC that is not amenable to curative surgery or radiation and for whom no contraindications to immunotherapy are present, first-line therapy with an approved anti-PD-1 ICI is recommended (LE:2).

• While responses to PD-1 blockade are frequent and generally durable, disease progression may eventually occur in a subset of responders, and hence continued surveillance for CSCC progression is warranted.

• For patients with CSCC being treated with anti-PD-(L)1 ICIs, there are no validated biomarkers that predict benefit (including PD-L1 expression).

Investigational immunotherapies for CSCC in the neoadjuvant and adjuvant settings

For patients with resectable CSCC tumors, there are currently no approved immunotherapy agents for neoadjuvant or adjuvant treatment. At the time of guideline preparation, several trials are investigating neoadjuvant ICIs, both as monotherapy and in combination with other ICIs, for the treatment of CSCC. The considerations and risk–benefit analyses for neoadjuvant use of ICIs in CSCC include the possibility for undue delay in surgery to the point of becoming unresectable and increased toxicity.

Although not current practice, one strategy that has reported encouraging data from non-randomized studies is a short course of neoadjuvant PD-(L)1 blockade followed by curative surgery. In a pilot study (NCT03565783) including 20 patients with CSCC of the head and neck, two preoperative doses of cemiplimab led to 11 pCRs and 3 major pathological responses in the resection specimens. Inflamed tumor microenvironments and enrichment for CD8⁺ tumor-specific T cells were observed in samples from tumors with pCR. DSS, DFS, and OS rates at 12 months were 95% (95% CI 85.9% to 100%), 89.5% (95% CI 76.7% to 100%), and 95% (95% CI 85.9% to 100%), respectively.118 A larger clinical trial (NCT04154943) for patients with CSCC is in progress to validate these findings. Both cemiplimab and pembrolizumab are being investigated in phase III studies (NCT03969004 and NCT03833167, respectively) for efficacy as adjuvant treatments in patients with high-risk or locally advanced CSCC tumors that have undergone complete/gross macroscopic resection and postoperative radiation therapy. However, as results for NMSCs, including CSCC, have yet to be reported, administration of adjuvant immunotherapies should only be done within a clinical trial.

Investigational immunotherapies for CSCC in the advanced/metastatic setting

Additional anti-PD-1 monotherapy studies are also ongoing. Two phase II trials are also currently investigating nivolumab for patients with advanced CSCC, NCT04204837 and NCT03834233. In the latter study, CA209-9JC, a phase II, open-label trial evaluating nivolumab for the first-line treatment of advanced or metastatic CSCC, the best ORR was 54.5% (n=22), and the median DOR, PFS, and OS had not been reached at the time of follow-up. Grade ≥3 TRAEs were reported in 21% of patients, with one discontinuation due to toxicity.119

ICI combination strategies for the treatment of advanced CSCC that have reached early-phase clinical trials include ICIs combined with radiation therapy, IL-17, targeted therapies (such as EGFR and MEK inhibition), and oncolytic viruses. Early data have been reported on the intratumoral use of the oncolytic herpes simplex type-1 virus RP1, which is engineered to increase immunogenic responses via delivery of granulocyte-macrophage colony-stimulating factor (GM-CSF) to the tumor microenvironment as well as via a fusogenic GALV-GP R-protein that causes TCs to form large multinucleated syncytia inducing cytotoxicity and promoting antigen presentation by dendritic cells. The agent has advanced through phase II trials for solid tumors, including CSCC. In the phase II IGNYTE trial (NCT03767348) of RP1 in combination with nivolumab that enrolled patients with either melanoma or NMSCs, interim results reported that 5 of 6 patients with CSCC responded within 1─7 months of treatment, with 3 achieving CR.120 RP1 is also being studied in combination with cemiplimab for treatment of advanced CSCC in the CERPASS study (NCT04050436).

Panel recommendations

• For patients with resectable CSCC at high-risk of recurrence, enrollment in clinical trials of neoadjuvant or adjuvant therapy should be offered where available.

• For patients with unresectable locally advanced or metastatic CSCC, enrollment in clinical trials of novel agents and novel combinations should be offered where available.

Special patient populations for CSCC

Similar to MCC, individuals who are immunosuppressed are more likely to develop CSCC,121 122 including those who have received solid organ transplants121 123 (who are more likely to develop aggressive CSCC),124 are HIV-positive,121 125 or have been diagnosed with CLL.126 These and other factors require special considerations for patients with CSCC when administering ICIs, including advanced patient age, high-dose corticosteroid use, and pre-existing autoimmune disease, which generally overlap with and are discussed in detail in the Special patient populations for MCC section. One phase II single arm study for immune-compromised patients—defined as having a history of HIV or a history of treated or active hematologic malignancies—with locally recurrent and/or metastatic CSCC (NCT04242173) is ongoing, evaluating cemiplimab. This study, and others, highlight the importance of collecting data on ICI safety and efficacy in patients considered ‘not typical’ candidates for immunotherapy, as well as providing access to care for these populations. Clinical trial enrollment is encouraged, if available, and remains a high priority for special patient populations.

Panel recommendations

• For solid organ transplant recipients with CSCC who are potential candidates for ICI therapy, healthcare professionals should have a careful risk–benefit discussion with the patient that includes the potential risk of allograft loss, which may result in the need for dialysis (ie, for kidney transplant recipients) or death (eg, recipients of heart, lung, or liver transplants) (LE:1). Chemotherapy (LE:1) and/or EGFR inhibitors (LE:3) are acceptable alternative first-line therapies in this population. Enrollment in clinical trials should be encouraged.

• For patients with CSCC who are receiving high-dose corticosteroids for reasons other than solid organ allograft preservation, the dose of corticosteroids should be reduced to ≤10 mg prednisone (or equivalent) per day, if possible, prior to initiation of ICI therapy.

• Caution is advised when treating with ICIs in patients with CSCC who have a history of autoimmune disease, as they may experience a flare of their autoimmune disease. Healthcare professionals should have a careful risk–benefit discussion with the patient, which includes the potential risks associated with an autoimmune flare (LE:1).

• For patients with CSCC on therapeutic immune suppression, a discussion should be initiated on reducing or modifying immune suppression, if appropriate, before proceeding with ICI therapy.

Approved immunotherapies for BCC

Immunotherapy for advanced BCC

In February 2021, the FDA granted full approval to cemiplimab for the treatment of patients with locally advanced BCC previously treated with a HHI or for whom HHI treatment is not appropriate, and accelerated approval for metastatic BCC previously treated with or not appropriate for treatment with an HHI.7 Approval was based on ORR (evaluated by RECIST v1.1 criteria or by a composite endpoint that integrated assessments of radiological data and digital medical photography for externally visible lesions) and DOR results from Study 1620 (NCT03132636), an open-label non-randomized phase II trial. Patients enrolled had experienced disease progression following prior treatment with HHI therapy, had not experienced an objective response after 9 months of HHI therapy, or were not suitable for HHI therapy and were not candidates for curative surgery or radiation therapy. Cemiplimab was administered at 350 mg IV every 3 weeks for up to 93 weeks or until disease progression or unacceptable toxicity. Results have been reported from a total of 112 patients, 28 with metastatic BCC7 and 84 with locally advanced BCC.138 For patients with locally advanced BCC, at a median follow-up of 15 months (IQR 8─18), the ORR was 31% (95% CI 21% to 42%), with 5 CRs (6%) and 21 PRs (25%). The median time to response was 4.3 months (95% CI 4.2 to 7.2). The median DOR was not reached (range 2.1─24.1+ months), and Kaplan-Meier estimates for DOR at 6 and 12 months were 91% (95% CI 68% to 98%) and 85% (95% CI 61% to 95%), respectively. Exploratory biomarker analyses in the cohort of patients with locally advanced BCC showed no clinically meaningful associations with response and PD-L1 expression, TMB status, or major histocompatibility complex (MHC)-I expression. Among the 28 patients with metastatic BCC, the ORR was 21% (n=6; 95% CI 8% to 41%) at a median follow-up of 9.5 months.7 All 6 of the responses observed were PRs (no CRs). DOR was ≥6 months for all 6 patients, and the median DOR was not reached (range 9─23 months). Notably, delayed responses were observed for all treated patients, with a median time to response of 4.3 months for patients with locally advanced BCC, and 3.2 months for those with metastatic BCC.

Serious AEs were experienced in 32% of all patients and discontinuation of treatment was required in 13% of patients.7 In the cohort of patients with locally advanced BCC, 48% reported grade 3–4 TRAEs, and 11% of patients had to discontinue treatment due to a TRAE. Any-grade irAEs were reported in 25% of patients, and grade 3 irAEs in 10% of patients (no grade 4 or 5 irAEs reported).138

Immunotherapies in development for BCC

In addition to imiquimod and cemiplimab as immunotherapy treatments for BCC, several novel agents and ICI combination strategies are also under investigation. Nivolumab monotherapy (for anti-PD-(L)1 naïve patients) and in combination with ipilimumab or relatlimab (for patients with progressive disease after anti-PD-(L)1 therapy) is being assessed in patients with metastatic or locally advanced, unresectable BCC (NCT03521830). The SONIB trial (NCT03534947) is investigating neoadjuvant strategies in resectable BCC tumors, where patients are treated with the HHI sonidegib followed by topical imiquimod, or sonidegib followed by surgery. Finally, in the phase II, open-label NCT03714529 trial, treatment with the PD-L1 peptide vaccine IO103 led to one PR in a target lesion and two CRs in non-target lesions among 10 patients, with no AEs greater than grade 2 (all reversible).139

Angiosarcoma

Angiosarcoma is a highly aggressive endothelial malignancy that can occur anywhere in the body; however, as many as 60% of cases are cutaneous.140 Being a rare tumor, there is limited consensus on standard of care, but current treatments tend to involve chemotherapy, surgery, radiation therapy, and anti-vascular endothelial growth factor (VEGF) and tyrosine kinase inhibitor (TKI) targeted therapies. There have been limited reports of immunotherapies being used to treat angiosarcoma to date, and only one case series has provided clinical data on the use of ICIs in this setting. Of 7 patients (5 with cutaneous angiosarcoma and 2 with breast angiosarcoma) treated with a variety of ICIs, 1 patient experienced a CR with extended treatment; 5 experienced PRs; and no patients developed grade ≥3 irAEs.141 Additional data are being gathered on the effectiveness of ICI combinations (with other ICIs, targeted therapies, and oncolytic viruses) for angiosarcoma in clinical trials.

Adnexal tumors

Cutaneous adnexal carcinomas are a group of rare NMSCs that are morphologically similar to primary adnexal structures present in normal skin (hair follicles, sebaceous glands, apocrine glands, and eccrine glands). Overall, the prognosis for patients with metastatic cutaneous adnexal carcinoma is quite poor, and these cancers lack approved systemic chemotherapies and standardized treatment guidelines.142 Owing to their rarity, data on the use of immunotherapies to treat these tumors primarily stem from case studies. Two case reports have demonstrated antitumor responses to pembrolizumab monotherapy143 and in combination with carboplatin.144 In the case report using pembrolizumab monotherapy, the tumor was TMB-H (17 mutations/Mb), illustrating the potential clinical utility of the tissue-agnostic indications discussed in the Tissue-agnostic indications for ICIs section.

Panel recommendations

• For patients with locally advanced or metastatic BCC that is relapsed/refractory to HHIs, or for patients who are intolerant to or not appropriate for HHIs, an approved anti-PD-1 therapy is recommended (LE:3).

• For patients with resectable BCC at high risk of recurrence, enrollment in clinical trials of neoadjuvant or adjuvant therapy should be offered where available.

• For patients with unresectable, locally advanced or metastatic BCC, clinical trial enrollment should be offered where available.

• For patients with advanced rare skin cancers (eg, cutaneous sarcomas, adnexal carcinomas, or sebaceous carcinomas) with progressive disease after first-line therapies or for whom no standard first-line therapies exist, tumors should be analyzed for TMB-H and dMMR/MSI-H. Off-label use of ICIs may be considered, and clinical trial enrollment should be encouraged.

Panel recommendations

• For patients with NMSCs who are receiving ICIs, optimal duration is unknown at this time. The panel recommends that therapy should continue for at least 1─2 years. If a patient experiences a confirmed CR, ICI therapy may be discontinued earlier.

• Patients with MCC, advanced CSCC, or advanced BCC should perform monthly self-skin and lymph node exams. They should also have clinical skin exams by a dermatologist and lymph node exams every 3─6 months for the first 24─36 months after diagnosis, and every 6─12 months thereafter.

• Post-treatment surveillance imaging for advanced MCC, CSCC, and BCC should be conducted using diagnostic CT with contrast or FDG-PET/CT scans. If indicated by symptoms, MRI may be helpful to image the head and extremities.

• The frequency of surveillance imaging for MCC should be based on recurrence risk. Surveillance imaging for high-risk MCC is typically conducted every 3─6 months during the first 2 years after diagnosis, and every 6─12 months thereafter until 5 years.

• Monitoring MCPyV oncoprotein serology (eg, AMERK) every 3─6 months may be used for surveillance of recurrent MCC in seropositive patients as an adjunct to physical exams and scans (LE:3).

• The frequency of surveillance imaging for advanced CSCC should be based on recurrence risk. Surveillance imaging for high-risk CSCC is typically conducted every 3─6 months for 3 years after diagnosis.

• For patients with NMSCs being treated with immunotherapy, physical functioning and symptoms should be assessed throughout treatment and post-treatment surveillance.

• Prior to being treated with immunotherapy, patients and caregivers should be educated about the treatment plan and potential AEs. Explicit call parameters for AEs should be provided. Listing practical information on pamphlets, wallet cards, medical alert bracelets, or other easy-to-carry media is encouraged.

• For patients undergoing treatment for advanced NMSCs, a multidisciplinary care approach is encouraged with attention to survivorship and consideration for referrals as needed to physical therapy, nutrition, pain and symptoms management, wound care, mental healthcare, financial counseling, fertility counseling, and end of life planning, as needed.