Inverse Planning Algorithm In Leksell GammaplanKeywords: gamma knife, technique, dose planning, radiosurgery, Dose distributionInteractive Manuscript
Ask Questions of this Manuscript:
What is the background behind your study?
Treatment planning with Leksell Gamma Knife® Perfexion™ has so far been primarily a manual process.
What is the purpose of your study?
An efficient inverse planning method has been developed and implemented in Leksell GammaPlan™.
Describe your patient group.
Describe what you did.
The inverse planning is a two-stage process: the first step consists of determining the number of isocenters by performing a non-spherical packing, of the target(s), iteratively placing isocenters in the outlined target volume using dose templates being based on the geometrical shape of the composite shots at the planning isodose level,. Thus, after the first step the number of isocenters and their approximate positions are determined. The second step aims at minimizing a cost function based on conformity of the planning isodose to the target, thus considering coverage and selectivity. Optionally, the optimization can penalize plans with poor dose fall-off and long beam on times. The user controls the weighting between the different terms in the cost function. Because of the non-convexity of the cost function a fast simulated annealing algorithm was implemented. New states are sampled from a probabilistic function depending on distance between the present state and the new state and the system is cooled as 1/k, where k is the number of iterations. For each isocenter the position, sector combination and relative weight are determined. To decrease the simulation times an approximate dose algorithm was implemented on a GPU leading to a 20-30-fold speed-up of the dose calculation. At every step in the optimization the user may intervene and change control parameters and add or remove isocenters as well as adjusting their positions, sector combinations, and relative weights.
Describe your main findings.
The algorithm creates dose plans that can compete with the metrics of manual dose plans created by experienced users. Creating a plan for a typical radiosurgical target takes a few minutes (evaluation excluded). Also, the algorithm is also able to create good plans at high isodoses giving quite homogenous dose distributions in the target. This is a notoriously difficult problem in forward planning.
Describe the main limitation of this study.
This is a retrospective study.
Describe your main conclusion.
The implemented inverse planning tool is able to assist the user in the treatment planning process
Describe the importance of your findings and how they can be used by others.
This question was not answered by the author
Treatment planning with Leksell Gamma Knife® Perfexion™ has so far been primarily a manual process.
An efficient inverse planning method has been developed and implemented in Leksell GammaPlan™.
The inverse planning is a two-stage process: the first step consists of determining the number of isocenters by performing a non-spherical packing, of the target(s), iteratively placing isocenters in the outlined target volume using dose templates being based on the geometrical shape of the composite shots at the planning isodose level,. Thus, after the first step the number of isocenters and their approximate positions are determined. The second step aims at minimizing a cost function based on conformity of the planning isodose to the target, thus considering coverage and selectivity. Optionally, the optimization can penalize plans with poor dose fall-off and long beam on times. The user controls the weighting between the different terms in the cost function. Because of the non-convexity of the cost function a fast simulated annealing algorithm was implemented. New states are sampled from a probabilistic function depending on distance between the present state and the new state and the system is cooled as 1/k, where k is the number of iterations. For each isocenter the position, sector combination and relative weight are determined. To decrease the simulation times an approximate dose algorithm was implemented on a GPU leading to a 20-30-fold speed-up of the dose calculation. At every step in the optimization the user may intervene and change control parameters and add or remove isocenters as well as adjusting their positions, sector combinations, and relative weights.
The algorithm creates dose plans that can compete with the metrics of manual dose plans created by experienced users. Creating a plan for a typical radiosurgical target takes a few minutes (evaluation excluded). Also, the algorithm is also able to create good plans at high isodoses giving quite homogenous dose distributions in the target. This is a notoriously difficult problem in forward planning.
This is a retrospective study.
The implemented inverse planning tool is able to assist the user in the treatment planning process
Project Roles:
