New data presented at the 2025 Association for Research in Vision and Ophthalmology (ARVO) meeting supports the combined use of PRAME status and the DecisionDx-UM gene expression profile test to refine metastatic risk stratification in uveal melanoma.

Castle Biosciences, the maker of DecisionDx-UM, reported results from an analysis of 1,297 patients, marking the first independent real-world validation of the 2023 COOG2.1 study. The findings confirm that the expression of the PRAME gene can further distinguish metastatic risk when used in combination with the 15-gene test, which classifies tumors as Class 1 (low risk) or Class 2 (high risk).

The data showed that PRAME status provided significant additional stratification within each class. Three-year melanoma-specific survival rates were 98.3% for Class 1/PRAME-negative tumors, 91.7% for Class 1/PRAME-positive, 87.9% for Class 2/PRAME-negative, and 66.7% for Class 2/PRAME-positive. Patients with Class 2/PRAME-positive tumors had a hazard ratio of 25.3 for melanoma-specific death compared to the Class 1/PRAME-negative group.

The results were derived through a collaboration between Castle Biosciences and the National Cancer Institute’s SEER Program, linking clinical testing data with cancer registry records for a population-based cohort.

“This real-world validation affirms both the long-term performance of DecisionDx-UM and the value of PRAME as an adjunct biomarker,” said Rebecca Critchley-Thorne, PhD, vice president of research and development at Castle Biosciences.

DecisionDx-UM is already widely used in US ocular oncology practices and is recommended by NCCN guidelines and the AJCC staging manual. Castle estimates that nearly 80 percent of newly diagnosed uveal melanoma patients in the United States receive the test. The company said in a press release that it plans to continue evaluating of the combined biomarker approach as more follow-up data become available. RP