Management decisions for nodal metastasis from an unknown primary melanoma

Case Study

A 55 year-old man who was in otherwise good health, presented with a palpable mass in his right axilla. Because the mass failed to resolve with conservative measures, an excisional biopsy was performed. Immunocytochemical staining of the biopsy specimen with antibodies to S-100, HMB-45, and Melan-A revealed cells consistent with malignant melanoma. The patient denied a personal or family history of melanoma. Thirty years earlier, he had undergone excision of a suspicious lesion from his back, which proved to be a pathologically benign, compound nevus. The patient acknowledged considerable sun exposure throughout his life.

Physical exam demonstrated a solitary, mobile palpable mass measuring approximately 2.5 cm in the right axillary region. Thorough physical evaluation including dermatologic exam did not identify a primary source of melanoma. Staging evaluation with whole-body positron emission tomography (PET)/computed tomography (CT), CT of chest/abdomen/pelvis, and brain magnetic resonance imaging (MRI) demonstrated an enlarged and PET-avid right axillary node (Figure 1) but no other evidence of metastatic disease outside the axilla.

Figure 1.

PET-avid right axillary node (arrow)

The patient underwent a radical axillary lymph node dissection. Of the 22 lymph nodes removed, none was positive for tumor. After an uncomplicated recovery from surgery, the patient received an allogeneic whole-cell melanoma vaccine (Canvaxin, CancerVax Corp., Carlsbad, CA) for one year under protocol. The patient remained disease-free for approximately 5 years postoperatively. At this juncture, routine imaging with PET/CT demonstrated a solitary PET-avid lesion within the left lateral lobe of the liver, which was highly suspicious for metastatic recurrence. Contrast CT confirmed a minimally enhancing hepatic mass (Figure 2). The patient subsequently underwent laparotomy and a left lateral hepatectomy. On exploration, there was no evidence of extrahepatic disease. Pathologic diagnosis of the hepatic lesion was consistent with metastatic malignant melanoma; the tumor was 2.5 cm with negative margins. The patient has remained alive and without clinical evidence of disease for 4 years since the liver resection and 9 years since the axillary lymphadenectomy.

Figure 2.

CT of abdomen with contrast showing hepatic metastasis (broken circle) and normal vascular landmarks, inferior vena cava (black arrow) and hepatic veins (white arrows).

Discussion

Nodal metastasis from an unknown primary melanoma (MUP) often presents as a single palpable mass, as described here. The prognostic significance of MUP has been unclear in prior studies.^1-9 However, our recent report demonstrated that among patients who undergo lymphadenectomy for palpable nodal metastasis in a single nodal basin, those with MUP have a better survival than those with known primary melanoma (MKP).¹⁰ In this large comparative analysis, survival rates were significantly higher in the MUP group. Even when MUP and MKP patients were matched by relevant prognostic variables, 5-year survival rates were strikingly different: 58% for MUP patients versus 40% for MKP patients (P<0.001); median survival was 13.8 years for MUP patients versus 2.8 years for MKP patients. The survival advantage for MUP persisted when prognostic matching was limited to the subgroup of patients who met stringent criteria for adequate lymphadenectomy. Multivariate analysis of the entire study population identified age, sex, nodal tumor burden, and primary status as significant prognostic factors. Worse prognosis was associated with age ≥60 years, male sex, greater number of tumor-involved nodes, and MKP.

The demographics of our previously reported MUP study group¹⁰ are consistent with those of prior studies.^2,4-7 Most of our MUP patients were <60 years old and male. The axilla was the most common site of nodal metastasis from MUP. The larger number of male than female patients is surprising because male sex is considered an adverse prognostic factor in patients with primary melanoma.¹¹ The large number of younger patients might be expected if MUP is the result of regression in response to a strong and durable antimelanoma immune response. Site of nodal metastasis was not significant in our multivariate analysis¹⁰ and is not a known prognostic factor for melanoma. Although not proven, a higher occurrence of regression in upper extremity and truncal primaries would favor an axillary site of nodal metastasis in patients with MUP.

Few studies have evaluated nodal tumor burden in MUP.^4-6 Anbari et al⁴ found that nodal involvement in patients with MUP was more likely to be limited to a single node. However, two other studies^5,6 reported that more than half of patients with MUP had more than one tumor-involved node. Similarly, our report linked MUP to a higher rate of multiple (>1) tumor-involved nodes.¹⁰ As expected, survival of both MUP and MKP patients consistently decreased with an increase in nodal tumor burden, but survival rates at each level of nodal tumor burden (1 vs 2-3 vs >3 tumor-involved nodes) were better for MUP patients than for MKP patients.

Lymphadenectomy is the standard initial therapy for most patients with stage III MKP and also is beneficial for patients with MUP and nodal metastasis in a single basin. As yet, no postoperative adjuvant therapy has been shown to prolong overall survival of patients in either group.¹⁰

Theories to explain MUP include spontaneous regression of the primary melanoma, malignant transformation of ectopic nodal nevi, prior removal of the primary without a pathologic diagnosis, and failure to recognize the primary. We believe that the most plausible explanation is immune-mediated regression induced by a strong antimelanoma immune response. As illustrated by the current example, the nodal metastasis involving the right axilla may have originated from a cutaneous melanoma at an unknown site, which subsequently underwent immune-induced complete regression. Complete regression is much less common than partial regression but has been documented and reported.^12-14 Such a process may involve a complex interplay of cellular and humoral immune-mediated responses.

In conclusion, palpable and histologic evidence of melanoma in a lymph node should prompt a careful search for the primary lesion and for potential signs of metastatic disease in other nodal or non-nodal sites. Staging evaluation should include complete imaging to rule out distant disease. If clinical evidence of disease is confined to a single nodal basin, regional lymphadenectomy is the initial treatment of choice, as it is for patients with MKP. Our study suggests a more favorable outcome in patients with MUP¹⁰ and confirms the important role of complete lymphadenectomy for MUP.