Models for Predicting Melanoma Outcome

Abstract

Clinical and pathological features that impact melanoma patient survival have been studied extensively for decades at major melanoma centers around the world. With the aid of powerful statistical techniques and computational methods, remarkable progress has been made in the identification of dominant factors that are linked to the natural history of melanoma and associated clinical outcome. A wide array of clinical prediction tools have been promulgated, primarily focused on forecasting survival outcomes across the melanoma continuum, with the exception of distant metastatic (Stage IV) melanoma. Recent changes in melanoma clinical practice resulting from the availability of new targeted and immune therapies that are effective in both metastatic and adjuvant settings, as well as level I evidence demonstrating no survival benefit for completion lymph node dissection after a positive sentinel lymph node biopsy, have together changed the melanoma landscape and will no doubt impact on approaches to outcome prediction. Against this contemporary and ever-evolving backdrop, we present clinical applications, criteria, challenges, and opportunities for interpreting and building tools for predicting melanoma outcomes.