Dirk Marwede MD, Presenter: Nothing to Disclose

Philipp Daumke MD, Abstract Co-Author: Nothing to Disclose

Kornel Marko MA, Abstract Co-Author: Nothing to Disclose

Donald Lobsien MD, Abstract Co-Author: Nothing to Disclose

Thomas Kurt Kahn MD, Abstract Co-Author: Nothing to Disclose

We developed a language-independent online term-browser for RadLex to facilitate term identification in the lexicon and to support the international use of RadLex for the indexing of teaching cases.

RadLex is a comprehensive and rapidly evolving radiological lexicon developed by the RSNA currently becoming available in other languages than english (e.g. german). The aim of the study was to develop a language-independent online term-browser for efficient term identification in RadLex and for comparison between terms in RadLex and other medical terminologies (MeSH, ICD 10).

RadLex, MeSH, and ICD 10 were imported in a term-browser and an online interface was developed to access those terminologies. A search engine was implemented using a method called morphosemantic segmentation which abstracts from variations in natural language by converting terms in a normalized language-independent format. This is achieved by splitting up words in small semantic units, comparing those units to language-specific subword-thesauri and aggregating similar units in language-independent equivalent classes (morphem-identifier). The algorithm supports different languages and synonym identification including latin phrases. Terms can be searched in RadLex, MeSh and ICD 10 and the complexity of resulting term lists can be weighted by relevance.

RadLex is a rapidly evolving radiological lexicon currently containing approx. 12000 terms and synonyms. To facilitate term identification in RadLex innovative search algorithms need to be developed. The implemented online term-browser supports search queries in different languages (currently english and german), displays the corresponding translation of terms and facilitates term identification on the basis of semantic associations (e.g. a search for liver tumor, a term not present in RadLex, returns hepatoblastoma and hepatocellular carcinoma and the corresponding translations). In addition, the term browser can be used to validate terms in RadLex against terms present in other medical terminologies (MeSH, ICD 10).

Marwede, D, Daumke, P, Marko, K, Lobsien, D, Kahn, T, A Language-independent Term-browser for Efficient Term Identification in RadLex.  Radiological Society of North America 2008 Scientific Assembly and Annual Meeting, February 18 - February 20, 2008 ,Chicago IL. http://archive.rsna.org/2008/6012346.html