Simultaneous radiosurgery for multiple brain metastases: technical overview of the UCLA experience

Abstract

Purpose/objective(s): To communicate our institutional experience with single isocenter radiosurgery treatments for multiple brain metastases, including challenges with determining planning target volume (PTV) margins and resulting consequences, image-guidance translational and rotational tolerances, intra-fraction patient motion, and prescription considerations with larger PTV margins.

Materials/methods: Eight patient treatments with 51 targets were planned with various margins using Elements Multiple Brain Mets SRS treatment planning software (Brainlab, Munich, Germany). Forty-eight plans with 0 mm, 1 mm and 2 mm margins were created, including plans with variable margins, where targets more than 6 cm away from the isocenter were planned with larger margins. The dosimetric impact of the margins were analyzed with V5Gy, V8Gy, V10Gy, V12Gy values. Additionally, 12 patient motion data were analyzed to determine both the impact of the repositioning threshold and the distributions of the patient translational and rotational movements.

Results: The V5Gy, V8Gy, V10Gy, V12Gy volumes approximately doubled when margins change from 0 to 1 mm and tripled when change from 0 to 2 mm. With variable margins, the aggregated results are similar to results from plans using the lower of two margins, since only 12.2% of the targets were more than 6 cm away from the isocenter. With 0.5 mm re-positioning threshold, 57.4% of the time the patients are repositioned. Reducing the threshold to 0.25 mm results in 91.7% repositioning rate, due to limitations of the fusion algorithm and actual patient motion. The 90th percentile of translational movements in all directions is 0.7 mm, while the 90th percentile of rotational movements in all directions is 0.6 degrees. Median translations and rotations are 0.2 mm and 0.2 degrees, respectively.

Conclusions: Based on the data presented, we have switched our modus operandi from 2 to 1 mm PTV margins, with an eventual goal of using 0.5 and 1.0 mm variable margins when an automated margin assignment method becomes available. The 0.5 mm and 0.5 degrees repositioning thresholds are clinically appropriate with small residual patient movements.

Fig. 1

Simulation of 3-degree rotation and the impact of the rotation on each of the targets. The first row is the Multiple Isocenter plan and the second row is the Single Isocenter plan. With a single isocenter treatment plan, the effects of these rotations become much more pronounced

Fig. 2

ExacTrac displacement data from a cohort of 12 patients were analyzed. With the 0.5 mm threshold, 57.4% of the time the patients are repositioned after the ExacTrac images

Fig. 3

Patient data from a cohort of 12 patients were analyzed for the translational movements. 90th percentile of translational movements in all directions is 0.7 mm with median translations of 0.2 mm. 90th percentile of translational movements in all directions is 0.6 degrees with median rotations of 0.2 degrees

Fig. 4

The dosimetric impact of the margins are shown for a cohort of 8 patients with 51 targets. Overall, 48 plans were created using various margins and V5Gy, V8Gy, V10Gy, V12Gy values were analyzed

Fig. 5

Histogram of targets binned by the distance from the isocenter. The data shown may further support the use of the variable margins, since a small fraction of targets are more than 6 cm away from the isocenter

Fig. 6

Depending on the clinical prescription strategy, if the target size is strictly followed for determining the prescription levels, then larger margins may also have an impact on prescriptions