Char and associates1 provided a retrospective review of 145 patients with uveal melanoma in whom serial assessments of tumor growth were made prior to treatment. In their article, they provided a careful application of the Cox2 proportional hazards regression model in which the time to metastasis is regressed on a number of covariates, including largest tumor diameter, patient age, distance of the tumor to the nerve or fovea, and ciliary body involvement.
Woolson RF, Folberg R. A Response to Uveal Melanoma: Growth Rate and Prognosis. Arch Ophthalmol. 1998;116(3):399-400. doi: