This article describes the development of SNQuery, a platform for defining subsets of clinical concepts from the SNOMED CT terminology using the SNOMED CT Expression Constraint Language (ECL). The platform uses a graph-oriented database to represent SNOMED CT content and methods to simplify and semantically validate ECL expressions to reduce execution time. In addition, the paper presents a method to visualize subsets of the terminology to aid understanding and validation. The goal of the platform is to support the linking of clinical information models to terminologies, which is necessary to achieve semantic interoperability between medical record systems.