Abstract Archives of the RSNA, 2013
David Joseph Vining MD, Presenter: Royalties, Bracco Group
Usama Ibrahim Salem MBBCh, MSC, Abstract Co-Author: Nothing to Disclose
Cihan Duran MD, Abstract Co-Author: Nothing to Disclose
Cristian Popovici, Abstract Co-Author: Employee, Eloquentix
Andreea Pitici, Abstract Co-Author: Employee, Eloquentix
Liming Jiang MD, Abstract Co-Author: Nothing to Disclose
Chengqian Xuanzi MD, Abstract Co-Author: Nothing to Disclose
ViSion is a multimedia structured reporting system that has the potential to become a global solution with its ability to operate in multiple languages and to provide an inexpensive method for coding radiologic observations/diagnoses for use in medical informatics applications.
We describe an integrated ontology authoring tool that is used in a multimedia structured reporting solution, called ViSion, which is capable of automatically translating radiology reports to any language and coding reports for billing and data mining.
We developed a multimedia structured reporting solution, called ViSion, which allows a radiologist to capture key images, tag those images with metadata describing anatomical locations and radiological observations/diagnoses, and assemble a multimedia structured report with image findings organized by anatomical categories. The metadata used to tag image findings has been developed and maintained with an integrated ontology authoring tool. The anatomical terms in the ViSion ontology are organized in a hierarchy for each body section, and each anatomical term in this structure is associated with a pathology tree containing radiologic observations and diagnoses for that anatomical site. The pairing of an anatomy location with a radiologic observation/diagnosis comprises a specific diagnosis. Each diagnosis can be further described with secondary characteristics that provide granular detail. This ViSion ontology and its tree structures were assembled in English, but the ontology has been translated to multiple foreign languages including Chinese and Arabic. Furthermore, all of the diagnoses contained in the ViSion ontology have been crossed-referenced to other standardized medical ontologies (e.g., RadLex, SNOMED, ICD-10-CM) to facilitate data mining and electronic billing operations.
The ViSion ontology has been created and is maintained by an authoring tool integrated with the system. The ontology currently consists of 918 anatomy terms and 1424 pathology terms that combined form 12,046 unique observations/diagnoses. Each of the terms has been translated to foreign languages and cross-referenced to other standardized ontologies.
Development of the ViSion Ontology for Structured Reporting. Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13023813.html