Background: Sentinel lymph node (SLN) biopsy for head and neck (H&N) melanomas may be more technically challenging compared with other locations because of complex lymphatic drainage patterns. This analysis was performed to compare the results of SLN biopsy for H&N, truncal, and extremity melanomas.
Methods: The Sunbelt Melanoma Trial includes patients aged 18 to 70 with melanomas > or = 1.0 mm thick. Statistical comparison was performed by chi2 or analysis of variance test.
Results: A total of 2610 patients were evaluated with a median follow-up of 18 months. The mean number of SLN per nodal basin was 2.8, 2.7, and 2.1 for H&N, truncal, and extremity melanomas, respectively. Median Clark level, Breslow thickness, and percentage of ulceration were similar between the groups. Peri-parotid SLN was identified in 25% of cases; there were no facial nerve injuries. SLN biopsy for H&N melanoma had higher false-negative rates at 1.5% (vs. 0.5% for trunk or extremity) but less histologically positive SLN at 15% (vs. 23.4%, and 19.5%; P <.001) compared with truncal and extremity melanoma. Blue dye was visualized less frequently in SLN of H&N melanoma patients compared with those with trunk or extremity melanomas.
Conclusions: Preoperative lymphoscintigraphy and meticulous intraoperative search for blue/radioactive nodes may improve results in H&N melanomas.