Continuous Path Radiation Therapy Treatment Planning for Gamma Knife Perfexion
K Ghobadi1*, D Aleman1, D Jaffray1,2, M Ruschin1,3, (1) University of Toronto, Toronto, Ontario, (2) Princess Margaret Hospital, Toronto, ON, (3) Odette Cancer Centre, Toronto, ONTH-A-137-4 Thursday 8:00AM - 9:55AM Room: 137
Purpose: The purpose of this work is to develop an inverse planning approach for Gamma Knife Perfexion that continuously delivers the dose over a path in the target.
Methods: Our approach consists of two steps: finding a path in the target, and optimizing shot shapes for that path. To obtain the path, a set of well-positioned isocentres are selected in the target, and then a Hamiltonian path that visits all the isocentres exactly once is found by graph-theory and spiral-based approaches.
Then, a linear optimization model is solved to obtain the optimal shot shapes and intensities by minimizing dose spillage from the target. The dose restrictions for the target and the organs-at-risk are constraints in the optimization. We additionally consider several criteria specific to continuous path, including machine speed constraints, delivery accuracy, preference for single/multiple paths, and smoothness of movement.
Results: We tested our approach on seven clinical cases and compared against inverse step-and-shoot and manually-generated forward plans. The mean difference in Paddick CI compared to forward and inverse step-and-shoot was 0.04 and -0.04, respectively, while the Classic CI mean difference was -0.05 and 0.05, respectively. The mean dose difference to 1mm^3 brainstem was -0.5Gy (range: -1.6Gy to 0.1Gy) and -0.24Gy (range: -1.9Gy to 0.9Gy) compared to forward and inverse step-and-shoot plans. However, the average beam-on time improved by 30min (range: -82.9min to -0.62min) and 103min (range: -304min to -9min) over forward and inverse step-and-shoot plans, respectively. The mean computational time for continuous path was 19.5min, a 198min improvement over inverse step-and-shoot plans.
Conclusion: The continuous path treatment plans showed comparable plan quality with forward and inverse step-and-shoot plans, while achieving better beam-on times. The computational time was also improved compared to the inverse step-and-shoot approach.
Funding Support, Disclosures, and Conflict of Interest: This work was by part funded by Elekta, Stockholm, Sweden.