The initial route of metastases in most patients with melanoma is via the lymphatics to the regional nodes. However, routine lymphadenectomy for patients with clinical stage I melanoma remains controversial because most of these patients do not have nodal metastases, are unlikely to benefit from the operation, and may suffer troublesome postoperative edema of the limbs. A new procedure was developed using vital dyes that permits intraoperative identification of the sentinel lymph node, the lymph node nearest the site of the primary melanoma, on the direct drainage pathway. The most likely site of early metastases, the sentinel node can be removed for immediate intraoperative study to identify clinically occult melanoma cells. We successfully identified the sentinel node(s) in 194 of 237 lymphatic basins and detected metastases in 40 specimens (21%) on examination of routine hematoxylin-eosin-stained slides (12%) or exclusively in immunohistochemically stained preparations (9%). Metastases were present in 47 (18%) of 259 sentinel nodes, while nonsentinel nodes were the sole site of metastasis in only two of 3079 nodes from 194 lymphadenectomy specimens that had an identifiable sentinel node, a false-negative rate of less than 1%. Thus, this technique identifies, with a high degree of accuracy, patients with early stage melanoma who have nodal metastases and are likely to benefit from radical lymphadenectomy.

Nevocytes in melanoma-draining lymph nodes can be mistaken for melanoma metastases and may possibly transform to melanoma. During the development of a new technique for managing high-risk primary melanomas, selective lymph node dissection, we examined 4,821 nodes from 208 melanoma patients by light microscopy and immunohistochemistry. Nodal nevi were identified in 49 of 226 lymphadenectomy specimens (22%), a frequency considerably higher than previously recorded (5-6%). Nevi occurred in 57 of 4,821 nodes (1.2%), in 84% of patients in one node, in 13% of patients in two nodes, and in 3% of patients in three nodes. Nevocytes were detected in hematoxylin and eosin-stained sections in 38 of 49 cases (78%) and exclusively by immunocytochemistry with an antibody to S-100 protein in 11 of 49 (22%). Nevi were in the peripheral capsule in 93% of cases and in internal trabecula in the remaining 7%. Nevocytes surrounded a small vessel in 33% of cases. Nevi were more frequent in axillary (37 of 140, 26%) and cervical nodes (seven of 40, 18%) than in inguinal nodes (five of 46, 11%). Nevi were more frequent in sentinel nodes, the first nodes on the lymphatics draining a primary melanoma (11 of 284, 3.9%), than in nonsentinel nodes (46 of 4,537, 1.01%; p < 0.0008). One of 1,071 nodes from 50 patients with breast cancer (0.1%) and none of 521 nodes from 50 patients with pelvic cancer contained nevocytes. That nodal nevi are selectively present in melanoma patients raises the possibility of their origin from nodal melanocytes influenced by tumor products. Alternatively, the association may indicate that the nevocytes of cutaneous nevi can be disrupted and displaced by the growth of an adjacent melanoma.

Intraoperative lymphatic mapping and SLND is a minimally invasive and highly accurate screening technique for determining which patients with CS-I head and neck melanomas have subclinical node metastases and therefore might benefit from cervical LND.