Rapid Semi-Automated Planning for VMAT Craniospinal Irradiation Using Raystation
Abstract
Purpose
For craniospinal irradiation (CSI), VMAT offers superior dose conformality and normal tissue sparing compared to traditional 3D techniques. However, generating these plans is cumbersome. It requires multiple isocenters, optimization of a large number of organs-at-risk (OARs), and complexities regarding how to junction the beam sets together. We aimed to demonstrate the feasibility of semi-automating, rapid VMAT CSI planning in Raystation.
Methods
Raystation's scripting environment was used to create a planning framework and GUI to rapidly plan VMAT CSI. The required inputs were: brain and spine CTVs, OAR contours and prescription. Users also need to generate patient's FeasibilityDVH using Sun Nuclear's PlanIQ software. Our framework can accommodate patients with a craniospinal axis up to 85cm. All PTV creation, isocenter placement, PRV generation, tuning/optimization structure generation, plan setup, and optimization can be performed automatically. Two to three beam sets are generated implementing 1 or 2 corresponding dose gradient junction regions based on the technique described by Myers et al. Beam sets are planned sequentially from superior to inferior with subsequent beam sets using the prior beam set as a background dose in generating gradient junctions to minimize the dosimetric impact of patient setup inaccuracies.
Results
This framework was used to re-plan 7 patients using a prescription of 3000cGy in 10 fractions. PTV lengths range from 52-77cm. Plan generation took an average of 34 minutes (8.8 for plan setup, 25.2 for optimization) with a range of 19.2 to 41.5 minutes. The average prescription coverage of the PTV was 94.3% with an average volume of 105% of prescription of only 0.73%. The average conformity index at prescription was 0.96. Our methodology generated plans with favorable target homogeneity, conformality, and OAR sparing compared to current literature.
Conclusion
Semi-automated rapid (~35min) VMAT CSI planning is feasible and can produce high quality plans.