Understanding when uveal melanoma (UM) acquires the molecular features that drive metastasis has long been a challenge, particularly for small tumors that are often observed rather than biopsied. A large multicenter study published in Nature Communications provides new insight into the early genetic evolution of UM and highlights the strengths of gene expression profiling over mutation analysis for risk stratification.
UM is the most common primary intraocular malignancy in adults and carries a high risk of metastatic death, despite effective local tumor control. While several recurrent driver mutations are well described, their timing and prognostic relevance during early tumor development have remained unclear.
In the new study, investigators analyzed 1,140 primary UMs in patients enrolled in the Collaborative Ocular Oncology Group Study No. 2, including 131 small tumors measuring 2.5 mm or less in thickness. Using a targeted next-generation sequencing panel alongside a clinically validated 15-gene expression profile (15-GEP) and PRAME (preferentially expressed antigen in melanoma) expression status, the researchers compared mutational status with long-term clinical outcomes.
The analysis confirmed that mutations in BAP1, SF3B1, and EIF1AX are associated with high, intermediate, and low metastatic risk, respectively. However, after accounting for 15-GEP and PRAME status, individual driver mutations no longer predicted prognosis on their own. In contrast, the gene expression–based classifier remained strongly predictive of metastasis-free and overall survival.
Importantly, the study sheds light on early tumor evolution. Small tumors were more likely to be classified as low-risk (Class 1), lack BAP1 mutations, and show fewer high-risk genetic features than larger lesions. Many appeared to represent an early evolutionary stage before full malignant transformation. The authors also identified a subset of tumors with low 15-GEP discriminant scores that may be in transition from low-risk Class 1 to high-risk Class 2 status, even when classic high-risk mutations were not yet fully established.
These findings support a model in which most UMs begin as genetically simple, low-risk lesions and later acquire additional alterations that drive aggressive behavior. From a diagnostic and prognostic standpoint, the results reinforce the value of gene expression profiling as a functional readout of tumor biology, capturing both tumor-intrinsic changes and contributions from the immune microenvironment.
While the authors caution that longer follow-up is needed, the work moves the field closer to a molecular definition of malignant transformation in UM. The study also demonstrates the growing role of integrated molecular testing in distinguishing indolent lesions from those requiring early intervention.
Newsletters
Receive the latest pathologist news, personalities, education, and career development – weekly to your inbox.
