Discussion

Immune checkpoint blockade is an exciting treatment option in mMCC, but a small proportion of patients do not respond or experience disease progression despite initial response. Herein, we describe two case examples whereby the addition of short-course RT was associated with subsequent improved immunotherapy response.

Both patients developed imaging evidence of disease progression with new and enlarging metastatic lesions within the first 3 months of initiating pembrolizumab. Pseudoprogression was considered in both cases. Pseudoprogression is a transient increased size of baseline tumor lesions due to edema and immune cell infiltration, and is occasionally used to describe delayed clinical responses in which initial increases in tumor burden are followed by tumor regression in the setting of immunotherapy [14, 15]. While pseudoprogression/delayed response has only been observed in two of 114 patients (1.8%) in the two phase II trials investigating pembrolizumab in mMCC, it has been reported to range from 2% in head and neck squamous cell carcinomas [16] to 6–7% in non-small cell lung cancer and melanoma [15, 17]. In patient A, pseudoprogression was unlikely due to the increasingly infiltrative nature of his growing metastasis, and a repeat scan in 3 months to rule out pseudoprogression was felt to be impractical given the high level of pain and need for prompt intervention. Patient B, who was found to have multiple new metastatic lesions increasing in size and number on serial imaging evaluations, was ruled out for pseudoprogression. Therefore, both patients were considered to have true disease progression.

In the setting of disease progression, standard treatment options are unfortunately limited and include cytotoxic chemotherapy or investigational trials. For the two cases described, single fraction RT was attempted for local control and to potentially enhance the opportunity for systemic response to continued pembrolizumab. These were the first two patients managed with this exact sequence at our institution and both achieved long-term local and systemic disease control.

We expected excellent local control and tolerability with single-fraction RT in mMCC. In a retrospective review of 26 mMCC patients treated with single-fraction RT of 8 Gy, 94% of patients had objective responses with a median PFS of 6.4 months [18]. Both patients tolerated their treatment course very well without any notable RT-associated toxicities. Consistent with prior reports of single-fraction RT, the treated sites of disease had marked response and durable local control.

Remarkably, both of our patients showed an out-of-field effect following immunotherapy and RT, with augmented response at unirradiated sites of disease. The timing of response, 2 months after RT for patient A and 4 months after RT for patient B, is consistent with what has been reported in the literature on abscopal effect, with a median time to response of 5 months [19]. Patient A’s untreated superior tumor demonstrated a complete response after 10 months, leading to pembrolizumab discontinuation at 12 months and a subsequent 17-month disease-free interval. In total, following RT patient A experienced a remarkable disease response for 29 months including a long treatment-free interval before developing a very recent 2 cm isolated thyroid recurrence. Patient B, who had diffuse LN involvement that was progressing prior to RT, now has a partial or complete response at multiple unirradiated LN regions at 12 months after RT. He remains on pembrolizumab with stable disease and continues to do very well clinically.

This is the first case series demonstrating an abscopal effect from limited-field, single-fraction RT in mMCC patients receiving immunotherapy. Two important clinical trials will investigate the interaction between immunotherapy and short-course RT in mMCC (NCT03071406 and Alliance A091605). Our case series supports the rationale of these trials. We believe there was synergism where the radiation may have acted as an in situ vaccine, enhancing the efficacy of the systemically administered immunotherapy [20, 21]. Interestingly, Patient A was an initial immunotherapy non-responder and Patient B was an initial responder with rapid re-progression, suggesting RT may have synergistic effects with immunotherapy in both patient populations.

We acknowledge that the findings of this case series require validation in larger, better-controlled cohorts and ongoing clinical trials. However, given that short-course RT is exceptionally well-tolerated, highly effective for local control, and potentially synergistic with immunotherapy, it is a compelling strategy that should be considered as a means to augment response in patients with mMCC who are progressing on immune checkpoint blockade. Although cure cannot be expected, our report demonstrates that a long disease-free and treatment-free interval is achievable with the addition of short-course RT in the setting of mMCC progression on immunotherapy.