Patterns of distant metastases in 215 Merkel cell carcinoma patients: Implications for prognosis and surveillance

Abstract

Approximately one-third of Merkel cell carcinoma (MCC) patients eventually develop distant metastatic disease. Little is known about whether the location of the primary lesion is predictive of initial distant metastatic site, or if survival likelihood differs depending on the metastatic site. Such data could inform imaging/surveillance practices and improve prognostic accuracy. Multivariate and competing-risk analyses were performed on a cohort of 215 MCC patients with distant metastases, 31% of whom had two or more initial sites of distant metastasis. At time of initial distant metastasis in the 215 patients, metastatic sites (n = 305) included non-regional lymph nodes (present in 41% of patients), skin/body wall (25%), liver (23%), bone (21%), pancreas (8%), lung (7%), and brain (5%). Among the 194 patients who presented with MCC limited to local or regional sites (stage I-III) but who ultimately developed distant metastases, distant progression occurred in 49% by 1 year and in 80% by 2 years following initial diagnosis. Primary MCC locations differed in how likely they were to metastasize to specific organs/sites (P < .001). For example, liver metastases were far more likely from a head/neck primary (43% of 58 patients) versus a lower limb primary (5% of 39 patients; P < .0001). Skin-only distant metastasis was associated with lower MCC-specific mortality as compared to metastases in multiple organs/sites (HR 2.7; P = .003), in the liver (HR 2.1; P = .05), or in distant lymph nodes (HR 2.0; P = .045). These data reflect outcomes before PD1-pathway inhibitor availability, which may positively impact survival. In conclusion, primary MCC location is associated with a pattern of distant spread, which may assist in optimizing surveillance. Because it is linked to survival, the site of initial distant metastasis should be considered when assessing prognosis.

Keywords: Merkel cell; carcinoma; dermatology; medical oncology; neoplasm metastasis; neoplasm staging; prognosis; radiology.

Figure 1

Flowchart of Merkel cell carcinoma (MCC) patient selection for Analysis Cohort. Patients in the Analysis Cohort either presented with stage IV MCC or developed distant metastases during follow‐up and had sufficient data to identify the location and timing of their distant metastases. The 49 patients excluded for delayed entry enrolled either greater than 180 days after their initial distant metastasis or were enrolled after death (13 patients were enrolled by family members or legal representative after their death). ^†Dates required for analysis were date of initial distant metastasis, date of death or last follow‐up, and date of initial consent

Figure 2

Timing of initial distant metastases after diagnosis (n = 194 patients). Patients in the Analysis Cohort (n = 215) who presented with distant metastases at initial diagnosis (n = 21) were excluded from this analysis. Patients were grouped by the quarter year in which they developed their first distant metastasis. The height of each bar is the percentage of patients who developed distant metastases during that quarter. On the bottom of the figure is listed the percentage of patients who have yet to develop their initial distant metastasis at the beginning of each year

Figure 3

A, Comparison of sites of initial distant metastasis between Merkel cell carcinoma (MCC) and melanoma. Percentages shown indicate the fraction of patients (with MCC or melanoma as indicated) whose first distant metastatic disease presented at each listed anatomic site. Given that some patients' initial metastatic disease was apparent at multiple sites, percentages totaled to greater than 100%. Melanoma data were obtained from multiple publications19, 29 with Samlowski et al reporting central nervous system (CNS) metastases and Tas et al reporting all others. Data for MCC metastases are from the 215 patients included in this study. Metastatic frequency by anatomic site was not compared for statistical significance as the melanoma data were aggregated from multiple cohorts. ^*There were not ample data describing the rate of initial metastatic involvement of the pancreas in patients with metastatic melanoma. B, Pattern of initial site of distant metastasis by site of primary lesion (215 patients; 305 sites of metastasis). Patients were grouped by their primary Merkel cell carcinoma site. Each bar represents the percentage of patients with a given primary site who developed an initial distant metastasis in the indicated distant site. 66 patients had multiple sites of initial distant metastasis (157 sites) and could contribute to multiple bars in one row (the median number of sites among patients with multiple sites was 2). As a result, the sum of each row is greater than 100%. A patient could not contribute to multiple bars in any column, as each patient had only one primary site. The metastatic site frequencies for each primary lesion site group (each row) were compared with the metastatic site frequency of all other primary site groups with the corresponding p‐value representing the significance of this comparison. ^†The “other” category includes brain, adrenal gland, bladder, bowel, gonad, heart, kidney, peritoneum/retroperitoneum, oral, and spleen

Figure 4

Merkel cell carcinoma (MCC) specific survival comparison by site of first distant metastasis (n = 215 patients). Five aggregate categories for initial metastatic site were created for statistical analysis. Twenty‐nine patients developed initial distant metastases limited to the skin/body wall (dashed line). The MCC‐specific survival of all other sites of initial distant metastasis (solid lines) was compared to the survival of patients with skin‐only distant metastases. Tick marks indicate patient censoring. Hazard ratios and corresponding p‐values were calculated using univariate competing‐risks regression models. Complete univariate and multivariate competing‐risks estimates are described in Table 2. ^†The “other” category includes brain, adrenal gland, bladder, bowel, gonad, heart, kidney, peritoneum/retroperitoneum, oral, and spleen