Sentinel lymph node (SLN) biopsy has emerged over the past 2 decades as a rational approach for staging regional lymph nodes in patients with clinically node-negative melanoma (stage I and II disease). Large multi-institutional studies have confirmed that when performed by experienced surgeons, it is an accurate, reliable technique for identifying occult regional nodal disease, and that SLN status is the most important prognostic factor in patients with stage I and II melanoma. However, the incidence of occult regional nodal metastasis in patients with thin melanoma (<or= 1.0 mm; approximately 70% of patients with newly diagnosed melanoma) is low, and whether to perform SLN biopsy in these patients remains controversial. Several predictors of SLN metastasis in patients with thin melanoma have been suggested, but none widely accepted. This article reviews current literature on these predictors in patients with thin melanoma. Although the ability to draw conclusions was limited by the size and design of the available studies, the authors tentatively conclude that SLN biopsy can be considered for patients with melanomas 0.75 mm or larger, those with T1b melanomas (i.e., <or= 1.0 mm; Clark level IV/V and/or ulcerated), and those with thin melanomas with an increased tumor mitotic rate (especially >or= 1 mitosis/mm2). Including younger age (e.g., <or= 40 years) in the decision also seems reasonable, particularly if the primary tumor is associated with a high tumor mitotic rate. Tumor regression does not seem to be associated with an increased risk for SLN metastasis. Firm conclusions on the predictive value of vertical growth phase, absence of tumor-infiltrating lymphocytes, or male gender were not possible, particularly if used as a sole criterion for offering this procedure. SLN biopsy should be discussed with all patients with newly diagnosed thin melanoma.