BACKGROUND: How to best risk-stratify patients with metastatic melanoma in the sentinel node (SN) is controversial. Not all node-positive disease is equivalent in terms of disease-free or overall survival, and some have suggested that submicrometastatic disease, characterized by <0.1 mm tumor burden, can represent a distinct classification not associated with a chance for recurrence or death. We hypothesize that all patients with metastatic melanoma cells in the sentinel node have potentially life-threatening disease. STUDY DESIGN: This is a retrospective review of an IRB-approved, prospectively maintained melanoma database of >1,100 patients. All invasive melanoma patients who had an SN biopsy and at least 1 year of followup were included. Patients with metastatic melanoma in the SN were divided into groups according to diameter of SN tumor burden: node-negative, <0.1 mm (submicrometastatic), 0.1 to 1.0 mm, and >1.0 mm. Statistical methods included the Jonckheere-Terpstra method, Fisher's exact tests, and Kaplan-Meier method. RESULTS: From July 1, 1998 to July 1, 2007, 578 patients with invasive melanoma underwent SN procedure. Median followup was 2.2 years. There was a statistically significant difference in the proportion of patients who experienced a recurrence between the node-negative group (11%) and the <0.1 mm group (24%) (p = 0.049). Patients in the submicrometastatic group have a statistically significant (p = 0.048) earlier recurrence than those in the node-negative group. CONCLUSIONS: These results suggest that patients with submicrometastatic SN disease should not be treated as node-negative, as it appears to represent a more biologically aggressive melanoma, associated with a substantially faster time to recurrence. We cannot agree with recent proposals that patients with very small tumor burden in the SN can be treated as node-negative and be spared completion node dissection.