A study was done to determine whether the risk of cancer development can be calculated through the use of mammographic patterns. Hyperplasia, bland fibrocystic disease, and incident cancers were correlated with Wolfe's mammographic classification scheme. Intraobserver and interobserver consistency were measured in the 8,033 classified mammograms. Maximum observer agreement was achieved by combining high-risk and low-risk categories. The data presented do not support the contention that diffuse mammographic patterns are useful predictors for determining strategies of screening or patient management; large-scale studies are needed before mammographic classification is adopted.