Lymphatic invasion and the Shields index in predicting melanoma metastases.

Background: Findings of the prognostic significance of lymphatic invasion are contradictory. To determine an as efficient cutaneous melanoma metastasis predictor as possible, Shields et al. created a new prognostic index. This study aimed to examine whether the lymphatic invasion analysis and the Shields index calculation can be used in predicting lymph node status in patients with cutaneous melanoma.

Methods: Lymphatic invasion of 100 melanoma specimens was detected by dual immunohistochemistry staining for the lymphatic endothelial marker D2-40 and melanoma cell S-100 protein. The Shields index was calculated as a logarithm by multiplying the melanoma thickness, square of peritumoural lymphatic vessel density and the number "2" for the present lymphatic invasion.

Results: No statistically significant difference was observed between lymph node metastatic and nonmetastatic melanomas regarding the lymphatic invasion. Metastatic melanomas showed a significantly higher Shields index value than nonmetastatic melanomas (p = 0.00). Area under the receiver operator characteristic (ROC) curve (AUC) proved that the Shields index (AUC = 0.86, 95% confidence interval (CI) 0.79-0.93, p = 0.00) was the most accurate predictor of lymph node status, followed by the melanoma thickness (AUC = 0.76, 95% CI 0.67-0.86, p = 0.00) and American Joint Committee on Cancer (AJCC) staging (AUC = 0.75, 95% CI 0.66-0.85, p = 0.00), while lymphatic invasion was not successful in predicting (AUC = 0.56, 95% CI 0.45-0.67, p = 0.31). The Shields index achieved 81.3% sensitivity and 75% specificity (cut-off mean value).

Conclusions: Our findings show that D2-40/S-100 immunohistochemical analysis of lymphatic invasion cannot be used for predicting the lymph node status, while the Shields index calculation predicts disease outcome more accurately than the melanoma thickness and AJCC staging.