Realtime Motion Compensation

Organ motion is one of the main problems in minimally invasive therapies in the abdomen. It is a recurring problem for almost all therapies, such as radiation therapy, and focused ultrasound treatment. Motion is arising from physiological processes like the respiratory and cardiac cycles, but also due to patient motion, or as a result of the treatment itself.

Fraunhofer MEVIS offers a software solution that can solve the motion problem for minimally invasive therapies. Based on an image stream showing the motion (for example from magnetic resonance or ultrasound imaging) the software analyses and predicts the motion ahead of time for real time application.

If you are vendor of a treatment system that does not yet compensate for motion and wonder how this would work in your treatment system: the toolbox can work as a plugin. Route your data stream first through our plugin and fetch data as if there was no motion! All complexity is hidden inside, like concurrent executions, temporal relations between data as well as logging and data persistence in background.   

• Motion compensation for real time use within your therapy solutions
• Combination of image registration, sensor data fusion, and temporal predictions
• Easy to use: all complexity like concurrent executions is hidden inside
• Can be used as a plugin or a standalone application

The toolbox keeps track of the relation between the moving organ state and a reference motion state which is defined by a planning image. Data from the moving state can then be mapped to the reference motion state.

Data which is defined in the reference state can be mapped to the moving state, for example organ locations and therapy plans.

This process allows for more accurate treatment monitoring, for example to map MR thermometry data from the moving object to the planning image. Additionally, this process can be applied to map estimated radiotherapy doses from the moving organ to the planning state.

For real time control of treatments delays introduced by image acquisition, transfer, and processing need to be compensated.
The analyzed motion will be up to half a second old depending on image modality and available hardware system. Temporal predictions compensate these delays and allow for real time treatment control. Sensor data like respiratory belt signals can be used to improve the robustness of the prediction.

The use of the toolbox is shown here for focused ultrasound in the moving liver. Using the temporal predictions, a control loop is established that predicts the target position ahead of time to configure the treatment device in advance.