PROJECT 4 SUMMARY Despite recent therapeutic advances, prognosis for metastatic melanoma remains poor. Patients with primary melanomas that are clinically and histologically similar at diagnosis often have vastly different outcomes: whereas some are cured after initial surgical resection, others develop loco-regional recurrence(s) and metastases, and eventually die. Such highly variable outcomes suggest underlying biological differences in tumors (cell-intrinsic) and/or the patients themselves (host, cell-extrinsic, e.g. immune response). Molecular alterations in tumors that can be robustly measured at diagnosis could be useful prognostic markers. Moreover, given that some of these markers may also drive disease progression, their study may yield novel insights into melanoma biology and generate new therapeutic targets. Recent trials have demonstrated that adjuvant treatments for advanced melanoma (stage III and IV) reduce rates of melanoma recurrence and metastasis(1-3). The success of adjuvant immune and small molecule inhibitor therapies has opened the possibility of extending their use to stage II patients, for whom adjuvant therapy is yet not part of standard care. However, these therapies have a significant toxicity, monetary cost, and unclear long-term benefit. Companion assays that might accurately assign a patient’s risk of recurrence and even predict a patient’s benefit from adjuvant therapy—measured as increased relapse-free survival (RFS)—could transform clinical management, reduce unnecessary morbidity and toxicity, and dramatically improve patient outcomes. MicroRNAs (miRNAs) are promising biomarkers because of their stability in tissues and fluids, and their demonstrated roles in cancer biology, including in melanoma. We hypothesize that a set of candidate miRNAs can be integrated into a relapse-prediction model that can predict stage II patient outcomes and benefits from adjuvant therapy, and that some prognostic miRNAs functionally modulate melanoma progression. We identified a tumor tissue-based miRNA signature highly prognostic of outcome for stage II melanoma patients and used an independent cohort of patients to demonstrate its excellent discriminatory accuracy for identifying patients with short (<3 years) versus long (>3 years) RFS. Here we propose to transform melanoma clinical practice and research paradigms by: 1) using NanoString, a state-of-the-art technology currently employed in clinical labs, to develop a relapse-prediction model for stage II melanoma patients based on miRNA expression in tumor samples (Aim 1); 2) identifying clinically relevant miRNA-regulated mechanisms (e.g., cell proliferation, immune evasion) that drive metastatic spread of melanoma cells from the primary tumor (Aim 2); and 3) testing the clinical validity of the relapse-prediction model in a randomized, prospective trial, the gold standard for clinical validation of biomarkers (Aim 3). Successful completion of this project promises to demonstr...