Terminologia anatomica: considered from the perspective of next-generation knowledge sources.

This report examines the semantic structure of Terminologia Anatomica, taking one page as an example. The focus of analysis is the meaning imparted to an anatomical term by virtue of its location within the structured list. Terminologia's structure, expressed through hierarchies of headings, varied typographical styles, indentations, and an alphanumeric code, implies specific relationships among the terms embedded in the list. Together, terms and relationships can potentially capture essential elements of anatomical knowledge. The analysis focuses on these knowledge elements and evaluates the consistency and logic in their representation. Most critical of these elements are class inclusion and part-whole relationships. Since these are implied, rather than explicitly modeled, by Terminologia, the use of the term list is limited to those who have some knowledge of anatomy; computer programs are excluded from navigating through the terminology. Assuring consistency in the explicit representation of anatomical relationships would facilitate adoption of Terminologia as the anatomical standard by the various controlled medical terminology (CMT) projects. These projects are motivated by the need to computerize the patient record, and their aim is to generate machine-understandable representations of biomedical concepts, including anatomy. Because of the lack of a consistent and explicit representation of anatomy, each of these CMTs has generated its own anatomy model. None of these models is compatible with any other, yet each is consistent with textbook descriptions of anatomy. The analysis of the semantic structure of Terminologia Anatomica leads to some suggestions for enhancing the term list in ways that would facilitate its adoption as the standard for anatomical knowledge representation in biomedical informatics.