Michael Schwenke1, Joachim Georgii2, Tobias Preusser2. 1. Fraunhofer Institute for Medical Image Computing MEVIS, Bremen, Germany. 2. Fraunhofer Institute for Medical Image Computing MEVIS.
Abstract
OBJECTIVE: Focused ultrasound (FUS) is rapidly gaining clinical acceptance for several target tissues in the human body. Yet, treating liver targets is not clinically applied due to a high complexity of the procedure (noninvasiveness, target motion, complex anatomy, blood cooling effects, shielding by ribs, and limited image-based monitoring). To reduce the complexity, numerical FUS simulations can be utilized for both treatment planning and execution. These use-cases demand highly accurate and computationally efficient simulations. METHODS: We propose a numerical method for the simulation of abdominal FUS treatments during respiratory motion of the organs and target. Especially, a novel approach is proposed to simulate the heating during motion by solving Pennes' bioheat equation in a computational reference space, i.e., the equation is mathematically transformed to the reference. The approach allows for motion discontinuities, e.g., the sliding of the liver along the abdominal wall. RESULTS: Implementing the solver completely on the graphics processing unit and combining it with an atlas-based ultrasound simulation approach yields a simulation performance faster than real time (less than 50-s computing time for 100 s of treatment time) on a modern off-the-shelf laptop. The simulation method is incorporated into a treatment planning demonstration application that allows to simulate real patient cases including respiratory motion. CONCLUSION: The high performance of the presented simulation method opens the door to clinical applications. SIGNIFICANCE: The methods bear the potential to enable the application of FUS for moving organs.
OBJECTIVE: Focused ultrasound (FUS) is rapidly gaining clinical acceptance for several target tissues in the human body. Yet, treating liver targets is not clinically applied due to a high complexity of the procedure (noninvasiveness, target motion, complex anatomy, blood cooling effects, shielding by ribs, and limited image-based monitoring). To reduce the complexity, numerical FUS simulations can be utilized for both treatment planning and execution. These use-cases demand highly accurate and computationally efficient simulations. METHODS: We propose a numerical method for the simulation of abdominal FUS treatments during respiratory motion of the organs and target. Especially, a novel approach is proposed to simulate the heating during motion by solving Pennes' bioheat equation in a computational reference space, i.e., the equation is mathematically transformed to the reference. The approach allows for motion discontinuities, e.g., the sliding of the liver along the abdominal wall. RESULTS: Implementing the solver completely on the graphics processing unit and combining it with an atlas-based ultrasound simulation approach yields a simulation performance faster than real time (less than 50-s computing time for 100 s of treatment time) on a modern off-the-shelf laptop. The simulation method is incorporated into a treatment planning demonstration application that allows to simulate real patient cases including respiratory motion. CONCLUSION: The high performance of the presented simulation method opens the door to clinical applications. SIGNIFICANCE: The methods bear the potential to enable the application of FUS for moving organs.
Authors: Michael Schwenke; Jan Strehlow; Daniel Demedts; Sabrina Haase; Diego Barrios Romero; Sven Rothlübbers; Caroline von Dresky; Stephan Zidowitz; Joachim Georgii; Senay Mihcin; Mario Bezzi; Christine Tanner; Giora Sat; Yoav Levy; Jürgen Jenne; Matthias Günther; Andreas Melzer; Tobias Preusser Journal: J Ther Ultrasound Date: 2017-07-24