Prediction of nonsentinel lymph node metastasis in acral melanoma with positive sentinel lymph nodes

Background: Metastasis in a nonsentinel lymph node (non-SLN) is an unfavorable independent prognostic factor in cutaneous melanoma (CM). Recent data did suggest potential value of completion lymph node dissection (CLND) in CM patients with non-SLN metastasis. Prediction of non-SLN metastasis assists clinicians in deciding on adjuvant therapy without CLND. We analyzed risk factors and developed a prediction model for non-SLN status in acral melanoma (AM). Methods: This retrospective study enrolled 656 cases of melanoma who underwent sentinel lymph node biopsy at Fudan University Shanghai Cancer Center from 2009 to 2017. We identified 81 SLN + AM patients who underwent CLND. Clinicopathologic data, including SLN tumor burden and non-SLN status were examined with Cox and Logistics regression models. Results: Ulceration, Clark level, number of deposits in the SLN (NumDep) and maximum size of deposits (MaxSize) are independent risk factors associated with non-SLN metastases. We developed a scoring system that combines ulceration, the cutoff values of Clark level V, MaxSize of 2 mm, and NumDep of 5 to predict non- SLN metastasis with an efficiency of 85.2% and 100% positive predictive value in the high-rank group (scores of 17-24). Conclusions: A scoring system that included ulceration, Clark level, MaxSize, and NumDep is reliable and effective for predicting non-SLN metastasis in SLNpositive AM.

Automated summary

A prediction model for non-sentinel lymph node (non-SLN) metastasis in acral melanoma (AM) can assist clinicians in determining appropriate treatment options.