Weighted Four-Dimensional IMRT Planning for Dynamic MLC Tracking Using a Practical and Simple Framework
H Tachibana*, Y Cheung, S Jain, A Sawant, UT Southwestern Medical Center, DALLAS, TXSU-E-T-555 Sunday 3:00:00 PM - 6:00:00 PM Room: Exhibit Hall
Purpose: We present a simple, practical framework for truly 4D lung IMRT planning based on a weighted individual-phase optimization paradigm. This strategy is specifically developed for use in real-time tumor tracking delivery systems so as to utilize respiratory motion as an additional degree of freedom rather than a constraint.
Methods: A 4D-CT scan from a lung SBRT patient was loaded into the Eclipse treatment planning system. The target and normal structures were manually contoured on each of the ten phases. For each phase, the total dose prescription was scaled by the number of phases and a seven-field plan was developed. An open-source deformable image and dose registration engine (DIRART) was used to deform the dose map at each phase to a reference phase. DVH data from the individually optimized phase plans were input into an in-house linear programming-based optimizer implemented in MATLAB, in order to determine dose-weighting factors for each phase. The objective function aimed to maintain PTV coverage while keeping normal structure dose as low as possible. This weighted-4D plan (W-4D) was compared to an ITV-based plan and a 4D plan with equal dose-weights to individual phases (E-4D).
Results: The W-4D dose fractions were determined to be 0.33, 0.01, 0.65 and 0.02 at phase 0%, 30%, 40%, and 90%, respectively (and zero elsewhere). PTV coverage (V95) was close to identical for all three strategies. The W-4D plan exhibited mean lung dose 18.8% and 8.5% lower and mean liver dose 23.3% and 5.7% lower than corresponding values from ITV-based and E-4D plans, respectively.
Conclusions: By significantly improving normal structure sparing while maintaining PTV coverage, weighted 4D planning represents a more attractive solution than ITV-based planning for (currently investigational) real-time tumor tracking-based delivery systems.