Dr. Daniela Schmitt is a medical physicist in the Department of Radiation Oncology at the Heidelberg University Hospital in Germany. Since the commencement of the CyberKnife Program at Heidelberg in 2016 Dr. Schmitt’s focus has been on CyberKnife planning. Prior to joining Heidelberg University Hospital in 2014 Dr. Schmitt worked at the CyberKnife Center in Frankfurt. She obtained her PhD in physics from the University of Heidelberg working at the German Cancer Research Center under the supervision of Prof. Uwe Oelfke in 2014.
Earlier this year Dr. Schmitt was awarded first place in the 2018 TROG Plan Study for SRS Brain, run by ProKnow, for a multiple brain metastases treatment plan. With a treatment plan for the CyberKnife System created using the Precision Treatment Planning System Dr Schmitt scored 146.23 out of a maximum possible score of 150 beating 159 participants using a range of commercially available products.
The challenge presented to participants was a brain SRS case with 5 GTVs, each treated to 20 Gy in one fraction. The minimum required volume to be covered by the prescription isodose line was 95% for all GTVs. Constraints were defined for the normal brain, brainstem, optic chiasm, optic nerves, hippocampus, lenses and eyes.
Dr. Schmitt chose CyberKnife’s fixed collimators for her plan, using the 5mm cone for all GTVs and also the 7.5 mm cone for GTV2 and the 7.5 and 12.5 cones for GTV4. She chose the “Full_Path” set and limited the total MU to 72,000, the Max MU per beam to 210 MU and the Max MU per Node to 730 MU. Both the left and right eyes were blocked for all beams.
For construction of the shells GTVs 1, 2, and 3 were combined to give the structure “h_GTV_klein” and this along with GTVs 4 and 5 were provided with shells with 2 and 8 mm expansions. Shells for all GTVs combined (“GTV-Total”) were created to control the low dose region. The auto-shell screen is shown in figure 1. Additionally, a shell with 4 mm expansion was created around GTV-Total in the contour tab of Precision to control the 10Gy volume.
Figure 1. Auto-shell box
Dr Schmitt explains that her usual approach to brain mets planning is to set standard maximum constraints for the optical system and the brainstem and then to first optimise the total target volume, followed by the shells from small to large. After that, she sets the achieved max doses for the shells as max constraints and optimises the shells further while attempting to preserving target coverage. If the targets loose coverage, as they did in this case, single targets are optimised separately. For this case the constraints and objectives are changed iteratively, depending on the ProKnow result. Figure 2 shows the final constraints list and the objectives table.
Figure 2. (a) Dose constraints and (b) objectives used in optimizing the plan. h_PTV is the GTV-Total plus a 1 mm margin, h_GTV4_min_3mm is the inner part of GTV4 ( GTV4 minus a margin of 3mm); h_GTV3_GTV5 is the sum of GTV3 and GTV5; h_GTV_klein is the sum of GTVs 1, 2, and 3.
The final dose distributions for each of the 5 GTVs can be seen in figures 3,4,5,6,7.
The dose statistic table for the plan the can be seen in figure 8. The DxVx values assessed by ProKnow for the organs at risk are shown in figure 9. Dr. Schmitt’s report from ProKnow can be accessed here.
Figure 8: The dose statistics Table
Figure 9. DxVx table
As the time reduction in Precision does not work very well for multiple targets, several beam reduction optimzations were performed. The final estimated treatment time was 146 minutes with a total MU of 68,045. The plan contained 159 nodes and 402 beams.