| Project/Area Number |
18K07629
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 52040:Radiological sciences-related
|
|---|
| Research Institution | International University of Health and Welfare (2021-2022)
The University of Tokyo (2018-2020) |
|---|
Principal Investigator |
Kunimatsu Akira 国際医療福祉大学, 国際医療福祉大学三田病院, 教授 (20323553)
|
|---|
| Project Period (FY) |
2018-04-01 – 2023-03-31
|
| Project Status |
Completed (Fiscal Year 2022)
|
| Budget Amount *help |
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2020: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2019: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2018: ¥2,210,000 (Direct Cost: ¥1,700,000、Indirect Cost: ¥510,000)
|
|---|
| Keywords | MRI / 脳腫瘍 / 画像診断支援 / 機械学習 / 深層学習 |
|---|
| Outline of Final Research Achievements |
Although machine learning technology has shown great promise in assisting the diagnosis of medical images, its evaluation has yet to be established. The main objective of this study was to construct a diagnostic support model to discriminate between multiple types of brain tumor images using deep learning and to evaluate its performance for brain tumors that are of high clinical importance and occur relatively frequently. Input images were created by extracting T1-weighted, T2-weighted, and contrast-enhanced T1-weighted images from brain MRI images of glioblastoma, primary central nervous system lymphoma, and meningioma, and trained on a publicly available deep learning architecture using the fine-tuning method. When the trained model was allowed to identify the images for validation, correct answers were obtained with more than 90% accuracy.
|
| Academic Significance and Societal Importance of the Research Achievements |
脳腫瘍の画像識別を行う特化型の人工知能アプリケーションの臨床導入はまだなされておらず、本研究の成果は同種の技術の臨床展開が進歩する上での土台となることが期待される。異なる種類の脳腫瘍においても人間の目にとって類似した画像パターンを示すことがしばしばあり、本研究の成果を用いることで従来に比べ医師の診断を保管する客観的指標として統合することでより質の高い画像診断が可能となることが期待される。
|
