2019 Fiscal Year Final Research Report
Investigation of multimodality diagnosis assistant system for breast cancer
Project/Area Number |
17K09061
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Multi-year Fund |
Section | 一般 |
Research Field |
Medical Physics and Radiological Technology
|
Research Institution | Shiga University (2019) Gifu University (2017-2018) |
Principal Investigator |
|
Project Period (FY) |
2017-04-01 – 2020-03-31
|
Keywords | breast cancer / computer aided diagnosis / image retrieval / multimodality imaging / mammography / breast ultrasound |
Outline of Final Research Achievements |
The purpose of this study was to develop a computerized system for assisting multimodality image diagnosis by collating image findings and abnormality locations on multimodality exams and retrieving reference images using multimodality information. For matching the locations of abnormalities, automatic detection of breast masses on automatic breast ultrasound volume scanner (ABVS) was investigated. Using deep leaning technique, relatively high detection rate was obtained, although improvement is needed for finding corresponding locations on a 3D model. For assisting differential diagnosis, reference image retrieval system was developed for mammograms and breast ultrasound images. High precision was obtained using multimodality information compared with those using single modality information, indicating the potential usefulness of the proposed study.
|
Free Research Field |
Medical image analysis
|
Academic Significance and Societal Importance of the Research Achievements |
乳がんによる死亡率低下には早期発見が最も重要である.乳がんの診断には,マンモグラフィや超音波検査など複数の画像モダリティが使用される.また,乳房を自動的にスキャンする超音波自動ボリュームスキャナ(ABVS)の有用性も認識されている.しかし読影データの増加により医師の負担も増加している.本研究ではマルチモダリティ診断が効率的かつ正確に行われるように,画像診断支援システムの開発を試みた.本研究成果で得られたABVSにおける病変の自動検出システムは今後マンモグラフィとの照合システムの開発に役立てられる.また,マルチモダリティ情報を用いた参照画像検索システムは,鑑別診断に有用であることが示唆された.
|