Constructing Thesaurus of Japanese Medical Term and Semantic Network in the area of Radiology Diagnosis
| Project/Area Number |
13670920
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Radiation science
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
ONOGI Yuzo The University of Tokyo, Graduate School of Medicine, Associate Professor, 医学部附属病院, 助教授 (90233593)
|
|---|
| Project Period (FY) |
2001 – 2002
|
| Project Status |
Completed (Fiscal Year 2002)
|
| Budget Amount *help |
¥3,400,000 (Direct Cost: ¥3,400,000)
Fiscal Year 2002: ¥900,000 (Direct Cost: ¥900,000)
Fiscal Year 2001: ¥2,500,000 (Direct Cost: ¥2,500,000)
|
|---|
| Keywords | Radiology report / Medical thesaurus / Natural language processing / MeSH / ICD10 / UMLS / SNOMED-CT / Morphological analysis / 放射線読影報告書 / 意味解析 / シソーラス |
|---|
| Research Abstract |
This research aimed to build a thesaurus of Japanese medical terms in the area of radiology diagnosis, and to build a semantic network among them so as to extract semantic information from radiology reports. As resources of Japanese medical terms, we used "Standard Disease Names in Japan corresponding to ICD-10 for electronic medical record" published by MEDIS-DC, "hesaurus for Medical and Health related Terms version 5" published in 2003 by JAMAS (Japan Medical Abstracts Society), and radiology specific terms extracted from radiology report being accumulated in the University of Tokyo Hospital. For each term we assigned semantic category such as 'disease name', 'anatomical structure', 'radiology finding', and 'other medical term'. At the same time each word is mapped to UMLS Metathesaurus by way of MeSH and ICD10. We intended to map these terms to SNOMED-CT, but we succeeded in only small part of it because it was difficult to build relationships between Japanese medical terms newly extracted from radiology report, stored in the thesaurus, and UMLS or SNOMED-CT. Next we chose frequent used verbs accompanied to disease name in radiology reports, and tried to extract semantic meaning from them in the following patterns ; "There exists disease", and "There are no disease" ; that is affirmative and denial statement which are regarded to appear frequently in radiology reports. Within 5000 reports, we succeeded to extract 1468 affirmative statements and 72 denial statements. This result was not good, and caused by the poor accomplishment of semantic network for these words. We conclude that we need to prepare mapping Japanese radiological terms to SNOMED-CT, and to analyze mechanisms to find appropriate semantic meaning of a radiology word when it is concatenated from more than two Japanese medical terms having specific semantic meanings already.
|
Report
(3 results)
Research Products
(4 results)
