Background: In patients with melanoma, lymph node staging information is obtainable by the surgical techniques of lymphatic mapping and sentinel lymph node (SLN) biopsy. Although no survival benefit has been proven for the procedure, the staging information is useful in identifying patients who may benefit from further surgery or adjuvant therapy. Currently, however, it is not being recommended for patients with thick melanomas (>3-4 mm). The risk of hematogenous dissemination is considered too great in these patients. Recent studies indicate, however, that a surprising number of patients with thick melanomas become long-term survivors, and the lymph node status may be predictive. None of the conventional microscopic features used to gauge prognosis in patients with melanoma have proven helpful in distinguishing the survivors with thick melanoma from those who will die of their disease. Objective: Our purpose was to evaluate the influence of SLN histology and other microscopic parameters on survival of patients with thick melanomas. Methods: A computerized patient database at the Cutaneous Oncology Clinic at H. Lee Moffitt Cancer Center was accessed to obtain records on patients with melanomas thicker than 3.0 mm (AJCC T3b). A retrospective analysis was conducted with attention paid to histologic variables, sentinel node status, and survival. Survival curves were constructed with the Kaplan-Meier method, and a Cox-Mantel rank testing was used to establish statistical significance. Results: Between 1991 and 1999, 201 patients were diagnosed with melanoma thicker than 3.0 mm, and 180 were alive at an average follow-up of 51 months. Of these, 166 were alive without disease. The mean overall and disease-free survival rates were 78% and 66%, respectively. There was a statistically significant difference in disease-free survival (3-year) between SLN-positive and SLN-negative patients (37% vs 73%, respectively; P = .02). The overall survival (3-year) for the SLN-positive patients was less than the node-negative patients (70% vs 82%), but it was not statistically significant (P = .08). The disease-free survival for patients with ulcerated lesions was less than for nonulcerated lesions (77% vs 93%, P = .05). None of the other histologic parameters studied, including Breslow thickness, Clark level, mitotic rate, or regression, had an influence on the overall or disease-free survival in this group of patients with thick tumors. Conclusions: The results indicate that the SLN node status is predictive of disease-free survival for patients with thick melanomas. A surprising number of patients in the study were free of disease after prolonged follow-up. None of the histologic features of the primary tumor were helpful in predicting outcome, except for ulceration. SLN biopsy appears to be justified for prognostic purposes in patients with thick melanomas. (J Am Acad Dermatol 2001;44:762-6.)