| Project/Area Number |
17H02110
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Medical systems
|
|---|
| Research Institution | Osaka University (2019)
Yamaguchi University (2017-2018) |
|---|
Principal Investigator |
KIDO SHOJI 大阪大学, 医学系研究科, 特任教授(常勤) (90314814)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
岡田 宗正 山口大学, 医学部附属病院, 准教授 (70380003)
間普 真吾 山口大学, 大学院創成科学研究科, 准教授 (70434321)
金 亨燮 九州工業大学, 大学院工学研究院, 教授 (80295005)
平野 靖 山口大学, 大学院創成科学研究科, 准教授 (90324459)
岩野 信吾 名古屋大学, 医学系研究科, 准教授 (90335034)
|
|---|
| Project Period (FY) |
2017-04-01 – 2020-03-31
|
| Project Status |
Completed (Fiscal Year 2019)
|
| Budget Amount *help |
¥15,600,000 (Direct Cost: ¥12,000,000、Indirect Cost: ¥3,600,000)
Fiscal Year 2019: ¥3,770,000 (Direct Cost: ¥2,900,000、Indirect Cost: ¥870,000)
Fiscal Year 2018: ¥5,330,000 (Direct Cost: ¥4,100,000、Indirect Cost: ¥1,230,000)
Fiscal Year 2017: ¥6,500,000 (Direct Cost: ¥5,000,000、Indirect Cost: ¥1,500,000)
|
|---|
| Keywords | コンピュータ支援診断 / ディープラーニング / ディープランニング / 類似画像検索 |
|---|
| Outline of Final Research Achievements |
We have developed computer-aided diagnosis (CAD) system that is more accurate and robust than the conventional CAD methods on high-resooution 3D images obtained from multi-detector row CT system for various lung diseases by use of deep learning technology.
For diffuse lung diseases, we extracted abnormal regions from each opacity pattern using U-Net and Residual U-Net. Nd, also we classified diffuse lung opacity patterns by use of unsupervised learning which does not require annotations by radiologists. For lung nodules, region extraction was performed three-dimensionally using DeconvNet and V-Net. In all cases, good results were obtained, which were in good agreement with the annotations by the radiologists.
|
| Academic Significance and Societal Importance of the Research Achievements |
これまでのCADの研究開発は，肺癌検診のために肺結節の検出や鑑別をするCADやびまん性肺疾患診断のための陰影パターンの分類をおこなうCADといった単一病変の検出や鑑別が目的とされてきたが，これは日常臨床業務のニーズにはマッチしてない．またCAD開発のために多くの画像症例が必要とされた．本研究では，実際の臨床現場で放射線科医が必要とする多様な肺疾患を統合的に診断支援するために放射線科医の負担を軽減して開発可能なCADを目指したところに学術的・社会的意義がある．
|
