Patient Population Characteristics
533 patients (315 male, 218 female) treated for MCC at PMCC between 1980 and 2018 were included in this study. Median follow-up for patients was 5.3 years. The distribution of patient characteristics is as listed in Table 1. The median age at diagnosis was 78 years (range, 19 – 98). Most primary tumours were located on the head and neck (50%) and 69/533 patients (13%) presented with regional nodal disease without an identified primary tumour. The median primary tumour size was 15.0mm (range, 1.0 – 180.0mm).
Merkel Polyomavirus testing was performed in 98 patients, of which 39/98 patients (40%) were virus positive. A history of synchronous or previous cutaneous skin malignancy was reported in 272/353 (77%) documented cases and 77/533 patients (14%) were recorded to be immunosuppressed.
Staging
524 patients had some or all of their staging at PMCC. PET scanning has become more prevalent for diagnostic work up over the last ten years. Between 2010 – 2019, 196/251 patients (78%) received PET; an absolute increase by 25% compared to the previous decade. N1b stage (image-detected lymphadenopathy) also increased during that time (16% vs 26%) (Supplementary online data – Table e1).
The majority of patients (43%) had Stage I disease (no more than 2cm across and not spread to lymph nodes). Only 12 patients (2%) presented with distant metastatic disease as this was routinely managed at their local centre.
Sentinel lymph node biopsy
Sentinel lymph node biopsy (SLNB) was performed in 66 patients at the PMCC. For patients with T1 disease, 60 SLNB were performed which returned 16/60 positive (27%) and 42/60 negative (70%) for disease with sentinel node status unknown for 2 cases. For T2 patients, 8 SLNB were performed, and of those 3/8 were positive (38%) and 5/8 were negative (63%). (Supplementary online data – Table e1).
Surgery
As first modality, surgery with therapeutic excision was performed on 454/533 patients (85%) with histopathological margin status known for 433 patients. Median pathological margin was 2.0mm [0.0 - 40.0mm] with positive pathological margins reported in 145/433 patients (33%), 154 patients underwent nodal surgery, of these 70/154 were biopsies (45%), 56/154 were nodal level dissections (36%) and 28/154 were local node excisions (18%). Overall, only 34/533 patients (6%) were treated with surgery as a single modality of which 26/34 (76%) were stage I or II.
Radiotherapy
79/533 patients (15%) were initially treated with definitive RT (11%) or combined chemo- RT (4%) with carboplatin and etoposide. 18 patients had Stage I disease, 15 had Stage II disease, 44 had Stage III disease and 2 had Stage IV disease. Following excision, 393/454 patients (87%) received post-operative RT to the primary and nodal regions if indicated.
Immunotherapy
Of the 533 patients identified, 26/533 (5%) received immunotherapy during their treatment course for recurrence.
Recurrence and Survival
The cumulative incidence of local recurrence at one, two and five years were 8%, 10% and 10% respectively. Locoregional recurrence occurring at one, two and five years was 24%, 31% and 32%, respectively. Distant recurrence occurring at one, two and five years was 16%, 25% and 30% respectively. See supplementary online data – Table e2 for more information.
Figure 1 demonstrates the relationship between locoregional and distal recurrence and death. The estimated five-year OS was 46% (95% CI = 41-51%). The estimated five-year DFS was 34% (95% CI = 30-39%).
292/533 patients (55%) died during follow up. There was information on cause of death available in 161/292 deaths of which 106/161 (66%) were attributed to MCC and 55/161 (34%) were from other causes.
Difference in DFS and OS by modality of treatment is outlined in Figure 2. The one and five year overall DFS estimates with surgery alone were 56% (95% CI =38 -71%) and 30% (95% CI =14 -48%) respectively. Those who were treated with definitive radiotherapy or chemo-radiotherapy had one and five-year DFS of 65% (95% CI =52 -75%) and 42% (95% CI =28 -55%). Patients who were treated with surgery followed by adjuvant radiotherapy had one and five- year DFS of 60% (95% CI =55 -65%) and 35% (95% CI 30 -40%).
The one and five- year OS estimates with 95% confidence intervals with surgery alone were 73% (95% CI = 54-85%) and 35% (95% CI = 17-54%). Definitive radiotherapy or chemo-radiotherapy had one- and five-year OS of 76% (95% CI = 63 – 85%) and 44% (95% CI = 30-58%). Patients who were treated with surgery followed by adjuvant radiotherapy had one and five- year OS of 85% (95% CI = 81 – 88%) and 48% (95% CI = 43 – 54%).
Those with MCC of unknown primary had improved DFS but not OS compared to a known cutaneous MCC primary (see supplementary online data Figure e1).
Table 2 demonstrates univariable and multivariable analysis on the prognostic variables of age, sex, immunosuppression, tumour size and histopathological margins. Positive histopathological margins were not significant for DFS (HR = 0.99, 95% CI: 0.74 - 1.34, p = 0.964) or OS (HR = 0.9, 95% CI: 0.68 - 1.18, p = 0.442). Increasing tumour size and age demonstrated an association with OS and DFS, with tumour size the most significant adverse factor for DFS (HR = 1.22, 95% CI: 1.09 - 1.37, p = 0.001) and OS (HR = 1.14, 95% CI: 1.03 - 1.26, p = 0.021). Figure 3 also demonstrates that earlier staging results in improved outcomes however over the last four decades, we have not observed any difference in DFS or OS (see supplementary online data Figure e2, Table e3-5).