The sentinel lymph node spread determines quantitatively melanoma seeding to non-sentinel lymph nodes and survival

INTRODUCTION: Complete lymph node dissection (CLND) after a positive sentinel node (SN) biopsy provides important prognostic information in melanoma patients but has been questioned for therapeutic use recently. We explored whether quantification of the tumour spread to SNs may replace histopathology of non-sentinel nodes (NSNs) for staging purposes. PATIENTS AND METHODS: We quantified melanoma spread in SNs and NSNs in 128 patients undergoing CLND for a positive SN. In addition to routine histopathology, one-half of each of all 1496 SNs and NSNs was disaggregated into a single cell suspension and stained immunocytochemically to determine the number of melanoma cells per 10(6) lymph node cells, i.e. the disseminated cancer cell density (DCCD). RESULTS: We uncovered melanoma spread to NSNs in the majority of patients; however, the tumour load and the proportion of positive nodes were significantly lower in NSNs than in SNs. The relation between SN and NSN spread could be described by a mathematical function with DCCDNSN = DCCDSN(c)/10(1)(-)(c) (c = 0.69; 95% confidence interval [CI]: 0.62-0.76). At a median follow-up of 67 months, multivariable Cox regression analyses revealed that DCCDSN (p = 0.02; HR 1.34, 95% CI: 1.05-1.71) and the total number of pathologically positive nodes (p = 0.02; HR 1.53, 95% CI: 1.07-2.22) were significant risk factors after controlling for age, gender, thickness of melanoma and ulceration status. A prognostic model based on DCCDSN and melanoma thickness predicted outcome as accurately as a model including pathological information of both SNs and NSNs. CONCLUSION: The assessment of DCCDSN renders CLND for staging purposes unnecessary.