Background: Individuals diagnosed as having mild cognitive impairment (MCI) have a high likelihood of progressing to dementia within 3 to 5 years, but not all individuals with MCI progress to dementia. Prognostic uncertainty suggests the need for additional measures to assist the clinician. Objective: To assess the added value of qualitative measures of medial temporal atrophy (MTA) to estimate the relative risk of progressing from MCI to dementia. Design: A 3-year, double-blind, placebo-controlled Alzheimer's Disease Cooperative Study initially designed to evaluate the efficacy of donepezil hydrochloride or vitamin E vs placebo to delay progression of MCI to dementia. Setting: Memory assessment centers. Patients: A total of 190 individuals with MCI. Main Outcome Measures: Ratings of MTA performed using magnetic resonance images obtained at baseline. Log-rank tests and Cox proportional hazards ratios examining the significance of MTA estimates in predicting progression of MCI to dementia. Results: A mean MTA score greater than 2.0 was associated with a greater than 2-fold increased likelihood of progression to dementia during the observation period (hazards ratio, 2.30; 95% confidence interval, 1.09-4.92; P=.03) after controlling for age, education, sex, and baseline Mini-Mental State Examination score. Conclusions: Adjusted estimates of MTA were associated with significantly increased risk of developing dementia within 3 years, suggesting that obtaining a magnetic resonance image during the evaluation of MCI may offer additional independent information about the risk of progression to dementia. Given the relatively high prevalence of MCI in the general population, use of this method as part of routine clinical evaluation may help identify individuals who might benefit from increased surveillance and future treatment. Trial Registration: clinicaltrials.gov Identifier: NCT00000173.