Complete lymph node dissection versus no dissection in patients with sentinel lymph node biopsy positive melanoma (DeCOG-SLT): a multicentre, randomised, phase 3 trial

Background: Complete lymph node dissection is recommended in patients with positive sentinel lymph node biopsy results. To date, the effect of complete lymph node dissection on prognosis is controversial. In the DeCOG-SLT trial, we assessed whether complete lymph node dissection resulted in increased survival compared with observation.

Methods: In this multicentre, randomised, phase 3 trial, we enrolled patients with cutaneous melanoma of the torso, arms, or legs from 41 German skin cancer centres. Patients with positive sentinel lymph node biopsy results were eligible. Patients were randomly assigned (1:1) to undergo complete lymph node dissection or observation with permuted blocks of variable size and stratified by primary tumour thickness, ulceration of primary tumour, and intended adjuvant interferon therapy. Treatment assignment was not masked. The primary endpoint was distant metastasis-free survival and analysed by intention to treat. All patients in the intention-to-treat population of the complete lymph node dissection group were included in the safety analysis. This trial is registered with ClinicalTrials.gov, number NCT02434107. Follow-up is ongoing, but the trial no longer recruiting patients.

Findings: Between Jan 1, 2006, and Dec 1, 2014, 5547 patients were screened with sentinel lymph node biopsy and 1269 (23%) patients were positive for micrometastasis. Of these, 483 (39%) agreed to randomisation into the clinical trial; due to difficulties enrolling and a low event rate the trial closed early on Dec 1, 2014. 241 patients were randomly assigned to the observation group and 242 to the complete lymph node dissection group. Ten patients did not meet the inclusion criteria, so 233 patients were analysed in the observation group and 240 patients were analysed in the complete lymph node dissection group, as the intention-to-treat population. 311 (66%) patients (158 in the observation group and 153 in the dissection group) had sentinel lymph node metastases of 1 mm or less. Median follow-up was 35 months (IQR 20-54). Distant metastasis-free survival at 3 years was 77·0% (90% CI 71·9-82·1; 55 events) in the observation group and 74·9% (69·5-80·3; 54 events) in the complete lymph node dissection group. In the complete lymph node dissection group, grade 3 and 4 events occurred in 15 patients (6%) and 19 patients (8%) patients, respectively. Adverse events included lymph oedema (grade 3 in seven patients, grade 4 in 13 patients), lymph fistula (grade 3 in one patient, grade 4 in two patients), seroma (grade 3 in three patients, no grade 4), infection (grade 3 in three patients, no grade 4), and delayed wound healing (grade 3 in one patient, grade 4 in four patients); no serious adverse events were reported.

Interpretation: Although we did not achieve the required number of events, leading to the trial being underpowered, our results showed no difference in survival in patients treated with complete lymph node dissection compared with observation only. Consequently, complete lymph node dissection should not be recommended in patients with melanoma with lymph node micrometastases of at least a diameter of 1 mm or smaller.

Funding: German Cancer Aid.