Andrew J Sinnamon1,2, Yun Song3, Cimarron E Sharon3, Yu-Xiao Yang4, David E Elder5, Paul J Zhang5, Xiaowei Xu5, Robert E Roses3, Rachel R Kelz3, Douglas L Fraker3, Giorgos C Karakousis3. 1. Department of Surgery, Endocrine and Oncologic Surgery Division, Hospital of the University of Pennsylvania, Philadelphia, PA, USA. andrew.sinnamon@uphs.upenn.edu. 2. Department of Biostatistics, Epidemiology and Informatics, University of Pennsylvania, Philadelphia, PA, USA. andrew.sinnamon@uphs.upenn.edu. 3. Department of Surgery, Endocrine and Oncologic Surgery Division, Hospital of the University of Pennsylvania, Philadelphia, PA, USA. 4. Department of Biostatistics, Epidemiology and Informatics, University of Pennsylvania, Philadelphia, PA, USA. 5. Department of Pathology, Hospital of the University of Pennsylvania, Philadelphia, PA, USA.
Abstract
INTRODUCTION: While recent trial data have demonstrated no survival benefit to immediate completion lymph node dissection (CLND) for positive sentinel lymph node (SLN) disease in melanoma, prediction of non-SLN disease may help risk-stratify patients for more intensive observation of the nodal basin. PATIENTS AND METHODS: A retrospective cohort of patients with positive SLN biopsy (SLNB) who underwent CLND was identified (1996-2016). A risk score for likelihood of CLND-positive disease was developed based on factors associated with presence of CLND metastases identified on logistic regression. Survival outcomes were analyzed. RESULTS: Among 312 patients with positive SLN, 192 underwent CLND and had complete pathologic data for evaluation. The median age of the study cohort was 53 years [interquartile range (IQR) 43-66 years], and 112 (58%) were male. Thirty-one (16%) had non-SLN metastatic disease on CLND. The four factors independently associated with CLND positivity and thus included in the risk score were Breslow thickness ≥ 3 mm [odds ratio (OR) 2.56, p = 0.047], presence of primary tumor-infiltrating lymphocytes (OR 0.33, p = 0.013), ≥ 2/3 positive-to-total SLN ratio (OR 4.35, p = 0.003), and combined subcapsular and parenchymal metastatic SLN location or metastatic deposit ≥ 1 mm (OR 4.45, p = 0.013). The four-point risk score predicted CLND positivity well with area under the curve of 0.82 (0.80-0.85). Increasing risk score was independently associated with increasingly worse melanoma-specific survival [hazard ratio (HR) = 1.54, p = 0.001]. CONCLUSIONS: Likelihood of residual nodal disease after positive SLNB and survival can be predicted from primary tumor and SLN characteristics. High-risk patients may warrant more intensive surveillance of the nodal basin to reduce risk of loss of regional control.
INTRODUCTION: While recent trial data have demonstrated no survival benefit to immediate completion lymph node dissection (CLND) for positive sentinel lymph node (SLN) disease in melanoma, prediction of non-SLN disease may help risk-stratify patients for more intensive observation of the nodal basin. PATIENTS AND METHODS: A retrospective cohort of patients with positive SLN biopsy (SLNB) who underwent CLND was identified (1996-2016). A risk score for likelihood of CLND-positive disease was developed based on factors associated with presence of CLND metastases identified on logistic regression. Survival outcomes were analyzed. RESULTS: Among 312 patients with positive SLN, 192 underwent CLND and had complete pathologic data for evaluation. The median age of the study cohort was 53 years [interquartile range (IQR) 43-66 years], and 112 (58%) were male. Thirty-one (16%) had non-SLN metastatic disease on CLND. The four factors independently associated with CLND positivity and thus included in the risk score were Breslow thickness ≥ 3 mm [odds ratio (OR) 2.56, p = 0.047], presence of primary tumor-infiltrating lymphocytes (OR 0.33, p = 0.013), ≥ 2/3 positive-to-total SLN ratio (OR 4.35, p = 0.003), and combined subcapsular and parenchymal metastatic SLN location or metastatic deposit ≥ 1 mm (OR 4.45, p = 0.013). The four-point risk score predicted CLND positivity well with area under the curve of 0.82 (0.80-0.85). Increasing risk score was independently associated with increasingly worse melanoma-specific survival [hazard ratio (HR) = 1.54, p = 0.001]. CONCLUSIONS: Likelihood of residual nodal disease after positive SLNB and survival can be predicted from primary tumor and SLN characteristics. High-risk patients may warrant more intensive surveillance of the nodal basin to reduce risk of loss of regional control.