| Project/Area Number |
18K15769
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 53010:Gastroenterology-related
|
|---|
| Research Institution | Aichi Cancer Center Research Institute |
|---|
Principal Investigator |
Takamichi Kuwahara 愛知県がんセンター(研究所), がん予防研究分野, 研究員 (10816408)
|
|---|
| Project Period (FY) |
2018-04-01 – 2021-03-31
|
| Project Status |
Completed (Fiscal Year 2020)
|
| Budget Amount *help |
¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000)
Fiscal Year 2020: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2019: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2018: ¥1,690,000 (Direct Cost: ¥1,300,000、Indirect Cost: ¥390,000)
|
|---|
| Keywords | 人工知能 / deep learning / 膵嚢胞 / IPMN / 膵腫瘍 / 膵管癌 / 膵神経内分泌腫瘍 / 慢性膵炎 / 膵臓 / 超音波内視鏡 |
|---|
| Outline of Final Research Achievements |
We developed AI system for the diagnosis of pancreatic malignancy which was difficult to be diagnosed by other image analysis. At first, we developed the AI system which diagnosed the IPMN malignancy and used EUS images of 50 IPMN patients. ResNet50 was used for this AI system and was developed by tensorflow and we evaluated this AI system using 10-fold cross validation. Using this AI system, we could diagnose IPMN malignancy (accuracy 94%).
Next, we developed the AI system for the diagnose of pancreatic tumor and used pancreatic ductal carcinoma (PDAC), pancreatic neuroendocrine tumor, autoimmune pancreatitis, and chronic pancreatitis patients (900 patients). Efficientnet-b4 was used for this AI system and developed by pytorch. Using super-computing resources, we evaluated this AI system by external validation. Using this AI system, we could diagnose PDAC or not (accuracy 90%)
|
| Academic Significance and Societal Importance of the Research Achievements |
AIによって超音波内視鏡（EUS）画像を解析することで他モダリティでは判定困難な膵疾患を高精度に診断することが可能であることを示した。今後薬事申請を踏まえた研究計画を立て一般的臨床で使用できるようにすることを目指す。それにより膵疾患の治療タイミングを逃さない、または不要な手術を減らすことができるなど膵疾患に対する治療判断の精度を向上することができるようになる。
|
