Exhibit Hall | Forum 4
Purpose: To establish a framework to automate SIMT treatment plans that incorporates plan checks and different levels of decision-making.
Methods: A script was developed using the Varian Eclipse v15.6 scripting API that includes the following units: 1.Initial check for contours and CT sim resolution; 2.Gross tumor volume (GTV) metrics (size, volume, and distance to isocenter) with volume expansion and prescription selection; 3.Technique selection. A second script was developed to provide summarized dosimetric information for each target including individual conformity index (CI). Five plans were created using the script simulating targets in an anonymized brain CT scan. Five plans were created with a different number of targets (3, 4, 5, 7 and 9). For GTVs smaller than 2 cm a 20 Gy prescription was selected. For GTVs larger than 2 cm an 18 Gy prescription was selected. After one optimization and calculation was completed, each plan was normalized so all PTVs received V100%≥99%. All plans were created for Varian TrueBeam using 10MV-FFF energy and four non-coplanar arcs.
Results: PTV sizes and volumes ranged between 0.92 and 2.16 cm, and 0.33 and 1.04 cc, respectively (n=28). Coverage for each individual PTV was V100%≥99% for all plans. The mean dose to the Brain-PTV was 138.2, 168.1, 213.0, 310.4, and 411.1 cGy for the 3, 4, 5, 7 and 9 PTVs plans, respectively. The CI for all individual PTVs ranged between 1.09 (three targets plan) and 1.89 (nine targets plan). Dosimetric goals for brainstem (Dmax<1500cGy, D0.5cc<1000cGy), optics (Dmax<1000cGy, V800cGy<0.2cc) and cochleas (Dmax<900cGy, Dmean<530cGy) were met for all plans. Planning time was under two hours for all cases.
Conclusion: We have demonstrated that it is feasible to establish a scripted framework for SIMT-SRS planning that incorporates plan checks, user input for decision-making and different levels of automation.