Program Information
Analysis of Spatial Trade-Offs for a Spinal SRS Case
T Viulet1 , O Blanck2 , A Schlaefer1,3*, (1) Universitaet zu Luebeck, Luebeck, Germany, (2) CyberKnife Zentrum Norddeutschland, Guestrow, Germany, (3) Hamburg University of Technology, Hamburg, Germany
Presentations
SU-E-T-42 Sunday 3:00PM - 6:00PM Room: Exhibit HallPurpose: To study whether the spatially limited relaxation of dose constraints could improve the target coverage for spinal lesions. Particularly, to illustrate that the localized relaxation of areas identified as restricting can be used to consider critical trade-offs.
Methods: We used an in-house planning system based on constrained optimization to explore potential spatial dose distributions. A three-dimensional visualization of the iso-dose surfaces and the dose bounds allowed studying the location of limiting constraints. An interactive tool was used to set optimization objectives and to selectively relax constraints. Considering a spinal tumor treatment which resulted in a relapse close to the spinal cord we studied trade-offs that could increase the coverage of the original GTV. Starting from a plan similar to the original treatment our goal was to cover the GTV facing the spinal cord and including the area of recurrence. Subsequently, different trade-offs are analyzed. For comparison, a global relaxation of the bounds was studied.
Results: We identified that the upper bounds on spinal cord, esophagus and PTV are partially limiting for the optimization. Relaxing limiting constraints in lateral directions, the upper bound in part of the GTV, and on the surface of the spine increased the coverage of the selected region from 5.6 to 87.4 % with 97.3% coverage of the recurrence. Only the explicitly relaxed areas received higher doses. In comparison, relaxing the upper bounds globally yielded 94.5% and 98.5% coverage for the selected and the recurrence regions, respectively. However, hot spots extended even outside the PTV and the 8.5 Gy iso-dose cut substantially deeper into the spinal cord.
Conclusion: Highlighting the constraining regions helped to quickly identify possible trade-offs. Spatially localized relaxation allows more fine grained compromises when planning cases where conflicting goals prevent solutions close to the original goals.
Contact Email: