Search

Objective To develop a nomogram to estimate the probability of positive sentinel lymph node (+SLN) for patients with thin melanoma and to characterize its potential impact on sentinel lymph node biopsy (SLNB) rates. Methods Patients diagnosed with thin (0.5-1.0 mm) melanoma were identified from the National Cancer Database 2012 to 2015. A multivariable logistic regression model was used to examine factors associated with +SLN, and a nomogram to predict +SLN was constructed. Nomogram performance was evaluated and diagnostic test statistics were calculated. Results Of the 21 971 patients included 10 108 (46.0%) underwent SLNB, with a 4.0% +SLN rate. On multivariable analysis, age, Breslow thickness, lymphovascular invasion, ulceration, and Clark level were significantly associated with SLN status. The area under the receiver operating curve was 0.67 (95% confidence interval, 0.65-0.70). While 15 249 (69.4%) patients had either T1b tumors or T1a tumors with at least one adverse feature, only 2846 (13.0%) had a nomogram predicted probability of a +SLN ≥5%. Using this cut-off, the indication for a SLNB in these patients would be reduced by 81.3% as compared to the American Joint Committee on Cancer 8th edition staging criteria. Conclusions The risk predictions obtained from the nomogram allow for more accurate selection of patients who could benefit from SLNB.