Servicenavigation

Project Details

Back

Projekt Print View

Robotic ultrasound guidance for real-time motion compensated radiation therapy (RobUST), Phase II

Subject Area Medical Physics, Biomedical Technology
Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing
Epidemiology and Medical Biometry/Statistics
Term from 2016 to 2024
Project identifier Deutsche Forschungsgemeinschaft (DFG) - Project number 288979502
 
Final Report Year 2024

Final Report Abstract

The research project focuses on the integration of robotic ultrasound (US) systems in medical applications, particularly in the context of radiation therapy treatment. The project encompasses multiple phases and explores various aspects, from data acquisition and analysis to the development of novel algorithms and methodologies for treatment planning. In the initial project phase, first, a long-term volunteer study was conducted using a robotic arm holding a transabdominal probe and an intercostal liver probe. The study revealed challenges with the accuracy of the force estimation model, leading to the implementation of a deep neural network to determine contact forces. This novel approach significantly reduced errors in measuring contact forces. Motion analysis of US probes in different anatomical viewports highlighted varying motion patterns in the liver and prostate. To optimize robotic US image acquisition, the project explored factors affecting tracking accuracy, such as imaging resolution and probe orientation. Also a neural network was trained to adapt the probe orientation based on 3D US images and force values, demonstrating a low error in a phantom study. Real-time processing of volumetric US data was addressed, and machine learning-based approaches were developed for target tracking and deformation estimation. These methods proved to be real-time capable with high accuracy, offering promising solutions for tissue tracking during medical procedures. The project also delved into calibration algorithms for robot-US probe alignment and investigated the impact of the US probe's weight and cable on joint torque measurements. The second part of the project focused on fundamental aspects of integrating robotic ultrasound with a robotic radiation therapy system, i.e., considering the robot carrying the ultrasound probe (US robot) during treatment planning. Furthermore, we studied how motion and deformation detected by real-time ultrasound imaging could be used for online treatment optimization. First, we studied the impact of robotic ultrasound on the plan quality, illustrating that particularly the US robot’s kinematic redundancy allows minimizing adverse effects. Second, we showed that a careful patient specific optimization of the setup is feasible and relevant and we demonstrated that coordinating the US robot and the treatment robot results in reduced treatment time. Third, we proposed a novel approach using machine learning methods for direct aperture optimization, significantly improving target coverage while reducing computational runtime. Forth, we investigated the impact of organ motion and deformation on the treatment plan quality, leading to the development of a new method for fast constraint re-planning after partial delivery of a treatment plan. Overall, this research project contributes valuable insights into the challenges and solutions associated with integrating robotic ultrasound systems into medical applications, particularly in the context of radiation therapy treatment planning. The study showcases advancements in technology, algorithm development, and optimization techniques, laying the foundation for enhanced medical procedures and patient-specific treatment planning.

Publications

 
 

Additional Information

© 2025 DFG Disclaimer / Copyright / Privacy Policy

Textvergrößerung und Kontrastanpassung