• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

Development of a real-time markerless tracking system for pancreatic cancer in carbon ion radiation therapy using deep learning

Research Project

Project/Area Number 22K15818
Research Category

Grant-in-Aid for Early-Career Scientists

Allocation TypeMulti-year Fund
Review Section Basic Section 52040:Radiological sciences-related
Research InstitutionGunma University

Principal Investigator

Varnava Maria  群馬大学, 重粒子線医学推進機構, 助教 (40913108)

Project Period (FY) 2022-04-01 – 2025-03-31
Project Status Granted (Fiscal Year 2023)
Budget Amount *help
¥4,420,000 (Direct Cost: ¥3,400,000、Indirect Cost: ¥1,020,000)
Fiscal Year 2024: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2023: ¥2,340,000 (Direct Cost: ¥1,800,000、Indirect Cost: ¥540,000)
Fiscal Year 2022: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Keywordstumor tracking / pancreatic cancer / deep learning / surface image guided / carbon ion radiotherapy
Outline of Research at the Start

Current tracking techniques for pancreatic cancer in carbon ion radiotherapy are invasive and require additional irradiation, or have poor robustness. This research will focus on the development of an accurate real-time markerless tracking system, which will be based on surface image guided radiation therapy and deep learning. The tracking system will detect the 3D surface of the patient using a camera and predict the target location at any time during treatment.

Outline of Annual Research Achievements

Pancreatic cancer is one of the most lethal cancers. One in three deaths in pancreatic cancer patients is related to tumor progression. Therefore, treatment methods that improve local tumor control may lead to better overall survival. The purpose for this year was to focus on completing a prediction model for the location of the target based on the surface of the patient, which is the basis for the research. The following steps were taken:
・Use combinations of recurrent and convolutional neural networks to create prediction models for the location of the tumor.
・Data collection/preprocessing/mining to match the nature of the network tested each time accordingly when necessary.
・Check network performance.

Current Status of Research Progress
Current Status of Research Progress

4: Progress in research has been delayed.

Reason

The progress was already slightly delayed from last year.
Also delayed in obtaining satisfactory performance from the model. Different approaches were explored that required the collection and preparation of more data, which took time.
The model is still under development, because of long calculation times and difficulties in solving technical difficulties during implementation.

Strategy for Future Research Activity

Complete the prediction model and continue as initially planned.

Report

(2 results)
  • 2023 Research-status Report
  • 2022 Research-status Report

URL: 

Published: 2022-04-19   Modified: 2024-12-25  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi