Sentinel lymph node biopsy (SLNB) in melanoma and a simple prognostic score in a prospective cohort with long term follow-up

8058 Background: Outcome of early stage melanoma is related to pathological and clinical findings like SLNB status, but a systematic approach to risk stratification is lacking. Objectives: To assess outcome, and clinico-pathological criteria associated to recurrence and death of patients (pts) who underwent SLNB Methods: From 11/1994 to 5/2005, 286 clinical stages I/II melanoma pts underwent SLNB in our Institution. Median follow up was 38 months. Prognostic factors were analysed with Logranktest and proportional hazard regression. Survival curves with Kaplan-Meier method. A score was derived from coefficients of multivariate analysis and evaluated with ROC curves on 259 pts with complete data. Results: Median age: 48yr; male sex: 54%; median Breslow (Br): 1.8mm. SLNB +: 46 pts (16.1%). Ulcerated: 31%. 5-yr overall survival was 56% in SLNB+ vs 84% in SLNB-, p=0.0002. Five-yr relapse free survival was 52% in SLNB+ vs 73% in SLNB-, p=0.0017. SLNB status, Br, Clark level, Age>50yr, male sex, and ulceration were related to death by univariate analysis, but in multivariate analysis, only SLNB status (HR 2.36), Br (HR 1.66 for each T level of TNM staging) and ulceration (HR 2.35) remained. The score derived from the model was: 5 points (p) for SLNB +; 4p for ulceration; and 0p for Br4mm. Conclusions: SLNB status, Br, and ulceration were, in concordance with literature, statistically significant prognostic factors. This allowed us to build a simple score with good correlation with prognosis (c index: 0.79). This score could be a useful tool for clinical practice and for future clinical trials, but validation in different populations is required. [Table: see text] No significant financial relationships to disclose.