Purpose: To illustrate an approach to evidence-based medical practice by reporting the Collaborative Ocular Melanoma Study (COMS) randomised clinical trials and cohort studies of choroidal melanoma. Methods: COMS randomised clinical trials of Iodine-125 (I-125) brachytherapy, adjunctive cohort study of visual acuity in eyes treated with brachytherapy and adjunctive natural history study. COMS randomised clinical trial of pre-enucleation radiation. Results: The COMS I-125 brachytherapy trial (N = 1,317 patients) of medium-sized choroidal melanoma showed 5-year all-cause mortality of 18% [95% Confidence Interval (CI), 16-20%] and no statistically significant difference in mortality following 1-125 brachytherapy or enucleation. Adjunctive cohort natural history study (N-42 patients) of patients eligible for the I-125 brachytherapy trial who deferred treatment or had no melanoma treatment had a 5-year all-cause mortality of 30% (95% CI, 18-47%). The COMS pre-enucleation radiation trial (N = 1,003 patients) of large-sized choroidal melanoma showed 5-year all-cause mortality of 40% (95% CI, 37-44%). Conclusions: Evidence derived from randomised clinical trials and cohort studies shows the need for longterm (ž 5 years) follow-up to determine the efficacy of treatment for choroidal melanoma by any modality. The rather similar 5-year mortality for treated and untreated medium melanoma patients suggests that metastatic dissemination may occur at an early stage of choroidal melanoma. To increase longterm survival, ocular treatment of choroidal melanoma must strive for diagnosis and treatment of melanoma at an early stage when metastasis is less likely and be combined with measures to detect and treat micrometastasis