Neoadjuvant Nivolumab for Patients With Resectable Merkel Cell Carcinoma in the CheckMate 358 Trial

Abstract

Purpose: Merkel cell carcinoma (MCC) is a rare, aggressive skin cancer commonly driven by the Merkel cell polyomavirus (MCPyV). The programmed death-1 (PD-1)/programmed death-ligand 1 (PD-L1) immunosuppressive pathway is often upregulated in MCC, and advanced metastatic MCC frequently responds to PD-1 blockade. We report what we believe to be the first trial of anti-PD-1 in the neoadjuvant setting for resectable MCC.

Methods: In the phase I/II CheckMate 358 study of virus-associated cancer types, patients with resectable MCC received nivolumab 240 mg intravenously on days 1 and 15. Surgery was planned on day 29. Tumor regression was assessed radiographically and microscopically. Tumor MCPyV status, PD-L1 expression, and tumor mutational burden (TMB) were assessed in pretreatment tumor biopsies.

Results: Thirty-nine patients with American Joint Committee on Cancer stage IIA-IV resectable MCC received ≥ 1 nivolumab dose. Three patients (7.7%) did not undergo surgery because of tumor progression (n = 1) or adverse events (n = 2). Any-grade treatment-related adverse events occurred in 18 patients (46.2%), and grade 3-4 events in 3 patients (7.7%), with no unexpected toxicities. Among 36 patients who underwent surgery, 17 (47.2%) achieved a pathologic complete response (pCR). Among 33 radiographically evaluable patients who underwent surgery, 18 (54.5%) had tumor reductions ≥ 30%. Responses were observed regardless of tumor MCPyV, PD-L1, or TMB status. At a median follow-up of 20.3 months, median recurrence-free survival (RFS) and overall survival were not reached. RFS significantly correlated with pCR and radiographic response at the time of surgery. No patient with a pCR had tumor relapse during observation.

Conclusion: Nivolumab administered approximately 4 weeks before surgery in MCC was generally tolerable and induced pCRs and radiographic tumor regressions in approximately one half of treated patients. These early markers of response significantly predicted improved RFS. Additional investigation of these promising findings is warranted.

Trial registration: ClinicalTrials.gov NCT02488759.

Figures

FIG 1.

Characteristics of treatment response. Change from baseline in the sum of target lesion diameters according to modified RECIST v1.1 in 33 evaluable patients who underwent surgery, by (A) tumor Merkel cell polyomavirus (MCPyV) status and pathologic response and (B) tumor cell programmed death-ligand 1 (PD-L1) expression and pathologic response. Pathologic response according to both site investigator review and central review is shown. Dashed horizontal lines indicate 30% target lesion reduction (consistent with a partial response in the absence of new lesions) and 20% increase (consistent with progressive disease). Note that radiographic responses per modified RECIST v1.1 were based on a single imaging scan before surgery, with no confirmatory scan performed. MPR, major pathologic response; pCR, pathologic complete response.

FIG 2.

Treatment response to neoadjuvant nivolumab in 2 patients with Merkel cell carcinoma (MCC). (A) Complete response in a 53-year-old woman with an advanced facial primary MCC, T3N0 (patient No.13 per Data Supplement). The tumor was Merkel cell polyomavirus (MCPyV) positive and programmed death-ligand 1 (PD-L1) negative. This patient received 2 doses of nivolumab preoperatively, with evidence of rapid tumor regression on physical examination (top row) and CT scans (bottom row, red arrows) at day 17. On day 20, she underwent surgery as originally planned (radical cheek resection, parotidectomy, and cervical lymph node dissection), revealing a pathologic complete response by site and central pathology reviews. Adjuvant radiotherapy totaling 50 Gy was administered to the primary tumor site as standard of care at the investigator’s discretion. At 3.5 years of follow-up, this patient remains tumor-free per the investigator. (B) Near-complete response of stage III MCC in a 67-year-old man with bulky left axillary lymph node metastases and an unknown primary tumor site (patient No. 21 per Data Supplement). The axillary tumor was MCPyV negative and PD-L1 negative. This patient received 2 doses of nivolumab on days 1 and 15, with a CT scan at day 18 showing partial tumor regression (top row, red circles). He underwent a complete left axillary lymph node dissection on day 23, with 2 of 47 lymph nodes showing residual microscopic tumor deposits (major pathologic response by central review, non-pCR by site review). No postoperative therapy was administered. This patient remains tumor-free 3 years later per the investigator. Features of immune-mediated pathologic response in the surgical specimen were evident with hematoxylin and eosin staining (bottom left panel: yellow arrow, infiltrating lymphocytes; red arrows, prominent plasma cell infiltrates; green arrow, proliferative fibrosis) and with multispectral immunofluorescence staining (bottom right panel: red stain, ERG+ neovasculature; green, CD79a+ B-lineage cells; yellow, CD3+ T cells; purple, programmed death-1 (PD-1)+ cells; blue, CD163+ myeloid cells). Note individual cells with both purple and yellow staining, indicating PD-1+ T cells (white arrow).

FIG 3.

Recurrence-free survival (RFS) in (A) all patients with Merkel cell carcinoma who underwent surgery (n = 36), and for patients who underwent surgery, on the basis of (B) pathologic response by site review (n = 36) or (C) radiographic response per RECIST v1.1 (n = 33). Panel B shows that a single event occurred among patients with a pathologic complete response (pCR); this was a death unrelated to the study drug, in the absence of tumor relapse. After database lock, 1 case scored as non-pCR by site review was found to be not evaluable for pathologic response, because a planned subtotal tumor resection was performed. Note that radiographic responses per modified RECIST v1.1 were based on a single imaging scan before surgery, with no confirmatory scan performed. Median follow-up for 36 patients from date of surgery was 19.3 months (range, 1.0-38.7 months). Analyses were exploratory, and all reported P values are nominal. HR, hazard ratio.

FIG 4.

Overall survival (OS) in (A) all patients with Merkel cell carcinoma who received ≥ 1 dose of nivolumab (n = 39) and for subgroups of patients who underwent surgery, on the basis of (B) pathologic response by site review (n = 36) or (C) radiographic response per RECIST v1.1 (n = 33). After database lock, 1 case scored as non-pCR by site review was found to be not evaluable for pathologic response, because a planned subtotal tumor resection was performed. Note that radiographic responses per modified RECIST v1.1 were based on a single imaging scan before surgery, with no confirmatory scan performed. Analyses were exploratory, and all reported P values are nominal. (*) Because of the absence of events among patients with tumor reduction ≥ 30%, hazard ratios and P values are not available from statistical analyses. HR, hazard ratio; NA, not available; pCR, pathologic complete response.

FIG 5.

Immunologic and pathologic changes associated with neoadjuvant nivolumab therapy in paired pretreatment and post-treatment tumor biopsy specimens from 8 patients with Merkel cell carcinoma. Each pair of biopsy specimens was taken from the same tumor site. The pathologic response status of each patient by central review, and tumor Merkel cell polyomavirus (MCPyV) status, are displayed. Proportions of cell types in selected high-power fields indicate that patients experiencing a response to therapy have a decrease in tumor cell proportion, with measurable increases in CD79a+ cell populations (pan–B-cell marker including plasma cells) in 3 of 5 responders. Among the 3 patients with non–pathologic complete response (pCR)/major pathologic response (MPR), 80%-100% residual viable tumor was observed in the complete specimen. CD3, T cells; CD163, macrophages; CD79a, B-lineage cells; Other, additional cell types such as fibroblasts and endothelial cells.