Prognosis in human melanoma: PAR-1 expression is superior to other coagulation components and VEGF

Two hundred and four accessible cases of malignant melanoma from the Grampian region of Scotland, collected over a period of 4 years, with minimum follow-up of 2 years, were studied for coagulation factors and vascular endothelial growth factor (VEGF) expression as potential prognostic markers. The aim was to allow comparison with previous work using microvessel density on the same cases.

Immunohistochemistry for VEGF, tissue factor (TF), fibrin and protease-activated thrombin receptor (PAR)-1 in 204 cases of melanoma was performed, and intensity of expression scored. Chalkley microvessel counts (MVD) were obtained for the tumour edge. TF expression and presence of fibrin correlated well with Breslow thickness and ulceration, reaching statistical significance, but surprisingly not for metastatic recurrence. Fibrin was variably present in over half the cases, located at the invasive edge, ulcerated surface and between tumour cell surfaces. In a few cases fibrin was within tumour cells, typically co-located with melanin and confirmed by electron microscopy. In contrast, immunohistochemistry for PAR-1 produced statistically significant results, correlating expression with Breslow thickness (P < or = 0.001), ulceration (P = 0.001) and recurrence (P < or = 0.005). Intensity of reactivity of VEGF correlated significantly with Breslow thickness, Clark level, ulceration and MVD, but not for metastatic recurrence.

It appears paradoxical that VEGF expression is not more predictive of recurrence, but even low expression may be sufficient for tumour angiogenesis and other factors must govern tumour aggression. Antagonism of VEGF may still prove a successful adjunct in future therapeutic trials. Both MVD and PAR-1 can be used as adjuncts to Breslow thickness and ulceration as prognostic indicators for melanoma, as they appear to give independent information for all thicknesses. PAR-1 expression is the best antibody marker of recurrence risk from those studied. It remains to be seen whether this methodology can predict response to novel antiangiogenic therapies currently entering trial.