Early-stage melanoma: staging criteria and prognostic modeling.

Accurate risk assessment is central to the process of making rational surgical and systemic treatment recommendations for melanoma patients and in establishing appropriate clinical trial stratification criteria. The current American Joint Commission on Cancer melanoma staging system incorporated relevant prognostic variables to provide a framework for the estimation of risk for recurrence; however, significant prognostic heterogeneity exists within the stage groupings. In the stage I/II group, survival rates range from 40% to 95% as defined by the combination of tumor thickness and ulceration. The use of novel prognostic factors, such as mitotic rate, sentinel node biopsy, and prognostic modeling using a variety of factors, can minimize this prognostic heterogeneity and provide a more accurate and individualized prognostic profile. Recent modifications in the stage III criteria include the number of positive nodes, whether the nodal disease is microscopic or clinically apparent, and the presence of an ulcerated primary. Through these factors, survival estimates can be provided, but like the stage I/II group, wide ranges in prognosis exist. The complexion of the stage III population is in evolution as a result of increasing numbers of patients being diagnosed as having microscopic sentinel node disease. Contemporary efforts are focused on defining the prognosis and natural history of this group. Through prognostic modeling using the number of nodes involved, ulceration status, and a measure of disease burden--disease in the sentinel node--relatively homogeneous subgroups can be identified. Long-term follow-up of patients staged with PCR molecular techniques on sentinel nodes shows conflicting value in assessing prognosis and therefore cannot be routinely used outside a clinical trial. The combination of genomic profiling using microarray analyses and the development of targeted therapy holds the future promise of individualizing prognosis and therapy.