Keywords:
Education and training, eLearning, Computer Applications-Detection, diagnosis, Conventional radiography, MR, CT, Neuroradiology spine, Neuroradiology brain, Computer applications
Authors:
D. BARUAH, C. E. Kahn; Milwaukee, WI/US
DOI:
10.1594/ecr2013/C-1088
Methods and Materials
"Gamut" means differential diagnosis in radiology; meaning a list of conditions that can cause a given radiological finding.
Neuroradiology comprises a vast array of diseases and their corresponding radiological manifestations.
We created a relational database of neurological diseases (such as “Craniopharyngioma”) and imaging findings (such as “Erosion of sphenoid wing”). The database defined a knowledge model of disease entities,
imaging findings,
synonyms,
and hierarchical relationships between terms.
In our Gamuts,
we defined relationships to link the ontology's terms in the following ways -
- Causality
- The causes relation indicates that one entity may be a cause of another; for example,
"hydrocephalus causes macrocephaly." The inverse relation is is_caused_by.
- Causal relations can be specified as common,
uncommon,
or rare.
- Synonymy
- A preferred term is linked to its synonyms using the has_synonym relation; for example,
"tuberous sclerosis has_synonym Bourneville disease".
The inverse relation is is_synonym_for.
- Subsumption
- The is_a indicates superclass-subclass relationships; for example,
"tuberous sclerosis is_a phakomatosis." The inverse relation is has_subtype.
Imaging manifestations were categorized as “common,” “uncommon,” or “rare” when appropriate. Several well-recognized radiology texts were used as references.
The system was built using open-source tools such as the MySQL relational database system and PHP programming language.