Prediction of outcome for patients with cutaneous melanoma.

Most patients with primary melanoma are cured by local surgery, but a significant minority develop fatal metastases. The ability to identify patients with progressive disease is central to efficient management: permitting optimal deployment of adjunctive therapy and sparing the non-progressing majority the morbidity of aggressive therapy. Accurate prediction on an individual patient basis is the ideal, but the best current prognosticators permit only assignment to risk categories. Formulaic combinations of well tried correlates of outcome (gender, ulceration, depth, thickness, and mitotic rate increase accuracy of prediction, but not to personalised level. The use of large data bases against which the attributes of individual patients may be compared is useful and amalgamation of data bases will increase the availability of this approach. The development of markers of proliferation fraction (PCNA and MIB-1) and of the metastatic phenotype (PNA-receptor status) will further refine the process. Staging of disease is critical. Accuracy of staging is improved by mapping the (sentinel) lymph nodes likely to contain early tumor by lymphoscintigraphy and dye/radiomarker localisation. The application of exquisitely sensitive immunohistochemical and molecular biological techniques to biopsies from tissues likely to be the site of metastases permit assessment of clinical stage with a previously impossible degree of accuracy (ultrastaging).