Sentinel node biopsy for head and neck melanoma: a systematic review.

OBJECTIVE: This systematic review was conducted to examine the test performance of sentinel node biopsy in head and neck melanoma, including the identification rate and false-negative rate. DATA SOURCES: PubMed, EMBASE, ASCO, and SSO database searches were conducted to identify studies fulfilling the following inclusion criteria: sentinel node biopsy was performed, lesions were located on the head and neck, and recurrence data for both metastatic and nonmetastatic patients were reported. REVIEW METHODS: Dual-blind data extraction was conducted. Primary outcomes included identification rate and test performance based on completion neck dissection or nodal recurrence. RESULTS: A total of 3442 patients from 32 studies published between 1990 and 2009 were reviewed. Seventy-eight percent of studies were retrospective and 22% were prospective. Trials varied from 9 to 755 patients (median 55). Mean Breslow depth was 2.53 mm. Median sentinel node biopsy identification rate was 95.2%. More than 1 basin was reported in 33.1% of patients. A median of 2.56 sentinel nodes per patient were excised. Sentinel node biopsy was positive in 15% of patients. Subsequent completion neck dissection was performed in almost all of these patients and revealed additional positive nodes in 13.67%. Median follow-up was 31 months. Across all studies, predictive value positive for nodal recurrence was 13.1% and posttest probability negative was 5%. Median false-negative rate for nodal recurrence was 20.4%. CONCLUSION: Sentinel node biopsy of head and neck melanoma is associated with an increased false-negative rate compared with studies of non-head and neck lesions. Positive sentinel node status is highly predictive of recurrence.