Risk Factors for Clinically Aggressive Melanoma in Veterans

Va Boston Health Care System, Boston MA

Investigators

Linked publications & trials

Abstract

Malignant melanoma is one of the top five cancers in Veterans. If diagnosed early, like the majority of melanomas in the U.S., it is highly curable with local surgery alone in the outpatient setting. In contrast, clinically aggressive melanoma, defined in this proposal as tumor stage pT1b, requires sentinel lymph node biopsy under general anesthesia and has a worse prognosis, accounting for a double the number of melanoma deaths compared to early-stage disease. There are no melanoma screening guidelines, and thus, there is a critical need for personalized easy-to-use approaches for melanoma screening to improve health outcomes and reduce treatment costs. Dr. Rebecca Hartman is a clinical dermatologist and researcher in epidemiology and health services research who seeks to improve melanoma screening approaches through personalization of risk assessment in a clinically applicable manner. The aims of this CDA-2 proposal are to (1) create and validate a clinical risk prediction model for clinically aggressive melanoma in Veterans and (2) create a genetic risk prediction model for clinically aggressive melanoma in Veterans and validate a combined genetic and clinical risk prediction model in an external civilian cohort. Aim 1 will conduct logistic regression analyses to examine for associations between potential phenotypic risk factors and clinically aggressive melanoma, including demographics, co-morbidities, disease history, immunosuppression, military history, environmental exposures, and healthcare utilization. This aim will produce an easy-to-use clinical risk prediction model for clinically aggressive melanoma that is specific to Veterans and can be integrated into the electronic health record. Aim 2 will conduct a candidate gene analysis of previously established cutaneous melanoma SNP risk factors as well as GWAS to examine for novel genetic risk factors for clinically aggressive melanoma. Cox proportional hazards regression will be used to create a polygenic risk score for clinically aggressive melanoma and model performance will be evaluated using decile percentiles of risk and ROC curve. This aim will produce a polygenic risk score for clinically aggressive melanoma. Subsequently, the combined clinical and genetic risk prediction model will be validated in an independent civilian cohort using the Mass General Brigham Biobank. Future directions include a clinical trial of the clinical risk prediction model, integrated into the electronic health record, to examine prospective real-world model performance as well as examination of any novel genetic associations found in GWAS for potential therapeutic applications. This clinical risk prediction tool will help VA clinicians select patients to triage to dermatology for melanoma screening. The mentorship, research, and training programs described in this CDA-2 application will propel Dr. Hartman’s career as an independent VA-based dermato-epidemiologic researcher. Her mentorship team includes Drs. Mary Brophy, Maryam Asgari, Michael Gaziano, Nathanael Fillmore, and Luc Djousse, who are experts in oncology, dermato-epidemiology, epidemiology, genetic epidemiology, and big data analysis from the Massachusetts Veterans Epidemiology Research and Information Center (MAVERIC) and the Harvard Dermatology Program. Through this CDA-2 application, Dr. Hartman will develop skills in big data analysis, risk prediction modeling, and genetic epidemiology as well as undergoing professional development to prepare her to apply for a VA Merit Award and continue toward becoming an independent VA-based clinician researcher.