RadLex exhibits excellent performance in indexing radiology-centric content

Despite RadLex’s fewer number of terms compared with other vocabularies, it showcases excellent performance in indexing radiology-centric content, according to a study published in the March issue of Academic Radiology.

The radiology lexicon, RadLex, offers a standardized means of categorizing journal articles and reviewers, encoding radiology result information, searching the content of radiology reports, analyzing queries to web-based search engines, and standardizing imaging procedure names for the American College of Radiology’s Dose Index Registry. Author Charles E. Kahn, MD, MS, of the Medical College of Wisconsin in Milwaukee, investigated the extent to which RadLex may be used to annotate and index a large database of biomedical images, as well as understand the scope and coverage of RadLex in comparison to other biomedical ontologies.

RadLex was evaluated during the study, as well as five other biomedical ontologies. Kahn utilized an automated web service to determine the vocabularies’ terms in a collection of 385,018 figure captions from 613 peer-reviewed biomedical journals. The annotations were analyzed by vocabulary and RadLex annotations were analyzed by journal and RadLex term class.

Results indicated that RadLex had the greatest number of annotations per term out of the six ontologies studied. There were an average of 10.1 RadLex annotations per figure, and 98.8 percent of figures were annotated with at least one RadLex term. Six or more RadLex terms were annotated in 74.8 percent of the figures. Out of the 39,218 RadLex terms, 8,504 were mapped to images in the collection. This was the highest percentage of any of the vocabularies.

“The rapid growth of biomedical information has posed new challenges for physicians and researchers. By helping to bring structure to that information, ontologies such as RadLex can aid clinical practice, medical education, and biomedical research,” wrote Kahn.