The authors evaluated the prognostic value of clinically assessed variables for predicting length of survival until death from metastatic disease in 237 patient with a primary choroidal or ciliary body melanoma. Using multivariate Cox proportional hazards modeling, the authors identified the largest linear basal tumor diameter (mm), estimated by indirect ophthalmoscopy and fundus drawing, the location of the anterior margin of the tumor relative to the ocular equator and ora serrata, and the age of the patient at the time of treatment as the best combination of the clinically assessed variables for predicting survival. For each patient in this group, the authors computed a prognostic index based on the best multivariate Cox model. They showed that patients with low, intermediate, and high values of prognostic index had low, intermediate, and high melanoma-related mortality rates, respectively, during the first 5 to 8 posttreatment years. These results suggest that: (1) clinically assessed variables evaluated according to a standardized protocol are useful for predicting the survival of treated patients with posterior uveal melanoma and (2) ophthalmologists who assess potential clinical prognostic variables consistently may identify subgroups of patients with comparable mortality risk on the basis of noninvasive testing.