Prediction of Non-sentinel Node Status in Patients with Melanoma and Positive Sentinel Node Biopsy: An Italian Melanoma Intergroup (IMI) Study

摘要

Abstract Background and Purpose Approximately 20% of melanoma patients harbor metastases in non-sentinel nodes (NSNs) after a positive sentinel node biopsy (SNB), and recent evidence questions the therapeutic benefit of completion lymph node dissection (CLND). We built a nomogram for prediction of NSN status in melanoma patients with positive SNB. Methods Data on anthropometric and clinicopathological features of patients with cutaneous melanoma who underwent CLND after a positive SNB were collected from nine Italian centers. Multivariate logistic regression was utilized to identify predictors of NSN status in a training set, while model efficiency was validated in a validation set. Results Data were available for 1220 patients treated from 2000 through 2016. In the training set ( n ?=?810), the risk of NSN involvement was higher when (1) the primary melanoma is thicker or (2) sited in the trunk/head and neck; (3) fewer nodes are excised and (4) more nodes are involved; and (5) the lymph node metastasis is larger or (6) is deeply located. The model showed high discrimination (area under the receiver operating characteristic curve 0.74, 95% confidence interval [CI] 0.70–0.79) and calibration (Brier score 0.16, 95% CI 0.15–0.17) performance in the validation set ( n ?=?410). The nomogram including these six clinicopathological variables performed significantly better than five other previously published models in terms of both discrimination and calibration. Conclusions Our nomogram could be useful for follow-up personalization in clinical practice, and for patient risk stratification while conducting clinical trials or analyzing their results.