Purpose To establish and evaluate a visual score focused on entorhinal cortex atrophy (ERICA), as the entorhinal cortex is one of the first brain structures affected in Alzheimer disease (AD). Materials and Methods In this retrospective study, ERICA was visually evaluated with magnetic resonance imaging (2009-2016). First, a four-point ERICA score was developed by using data in 48 consecutive subjects (20 patients with AD and 28 control subjects). Then, in the main analysis, ERICA and the standard medial temporal lobe atrophy (MTA) scores were determined in an independent cohort of 60 patients suspected of having AD (mean age, 69.4 years; range, 46-86 years) and in 60 age-matched patients with subjective cognitive decline (SCD) (mean age, 72.4 years; range 50-87 years). Score performances were evaluated with κ statistics, receiver operating characteristic analysis, t tests, and analysis of variance according to the Standards for Reporting of Diagnostic Accuracy Studies. Results Patients with AD had higher MTA scores (mean, 2.13) and ERICA scores (mean, 2.05) than patients with SCD (P < .001). An ERICA score of 2 or greater achieved a higher diagnostic accuracy (91%) than the MTA score (74%), with a sensitivity of 83% versus 57% and a specificity of 98% versus 92% in discriminating dementia caused by AD from SCD (P < .001). The ERICA score was correlated with amyloid β 42/40 ratio (ρ = -0.54, P < .001) and with cerebrospinal fluid tau (ρ = 0.35, P = .001) and p-tau (ρ = 0.31, P = .004). In multivariable linear regression analysis, ERICA was associated with verbal learning and recall (β = -.40 and -.41), nonverbal recall (β = -.28), and cued recall (β = -.41, P ≤ .002 for all). Conclusion An ERICA score of 2 or greater indicates probable AD with high diagnostic accuracy. © RSNA, 2018 Online supplemental material is available for this article.