Clark Level Risk Stratifies Patients with Mitogenic Thin Melanomas for Sentinel Lymph Node Biopsy

Background. The role for sentinel lymph node biopsy (SLNB) in patients with thin melanoma (<= 1 mm) remains controversial. We examined a large cohort of patients with thin melanoma to better define predictors of SLN positivity. Methods. From 1995 to 2011, 781 patients with thin primary melanoma and evaluable clinicopathologic data underwent SLNB at our institution. Predictors of SLN positivity were determined using univariate and multivariate regression analyses, and patients were risk-stratified using a classification and regression tree (CART) analysis. Results. In the study cohort (n = 781), 29 patients (3.7 %) had nodal metastases. In the univariate analysis, mitotic rate [odds ratio (OR) = 8.11, p = 0.005], Clark level (OR 4.04, p = 0.003), and thickness (OR 3.33, p = 0.011) were significantly associated with SLN positivity. In the multivariate analysis, MR (OR 7.01) and level IV-V (OR 3.45) remained significant predictors of SLN positivity. CART analysis initially stratified lesions by mitotic rate; nonmitogenic lesions (n = 273) had a 0.7 % SLN positivity rate versus 5.6 % in mitogenic lesions (n = 425). Mitogenic lesions were further stratified by Clark level; patients with level II-III had a 2.9 % SLN positivity rate (n = 205) versus 8.2 % with level IV-V (n = 220). With median follow-up of 6.3 years, five SLN-negative patients developed nodal recurrence and four SLN-positive patients died of disease. Conclusions. SLN positivity is low in patients with thin melanoma (3.7 %) and exceedingly so in nonmitogenic lesions (0.7 %). Appreciable rates of SLN positivity can be identified in patients with mitogenic lesions, particularly with concurrent level IV-V regardless of thickness. These factors may guide appropriate selection of patients with thin melanoma for SLNB.