Automated Dynamic Lattice Construction for Sfrt: Decoupling Biological Effectiveness from Sphere Size In Brain Metastases
Abstract
Purpose
To evaluate the efficacy of dynamic lattice structures in Gamma Kinfe (GK) Spatially Fractionated Radiation Therapy (SFRT) and compare their performance against traditional static lattice designs. Plans were assessed using Equivalent Uniform Dose (EUD), Valley-Peak-Ratio (VPR), Target Mean Dose (TMD) and Therapeutic Ratio (TR).
Methods
CT images and target tumor contours were obtained from 20 patients with large brain metastases under an IRB-approved study protocol. For the GK planning system, SFRT vertices were generated using an in-house script and subsequently imported. Radiation shots were localized within a tumor volume retracted by 3mm margin, delivering a single fraction dose of 18Gy to the 70% isodose line. The treatment plan utilized a face-centered-cubic (FCC) lattice structure with 4 mm or 8 mm spheres. To maintain consistent valley dose regardless of sphere sizes or number of spheres, an iterative optimization algorithm was employed to dynamically adjust lattice constants based on 3D dose profiles of the spheres.
Results
While static lattice plans exhibited significant metric variability between 4 mm and 8 mm sphere sizes, dynamic lattice placement demonstrated superior consistency. Specifically, EUD averages for static lattices were 4.6 (4mm) and 2.6 (8mm), whereas dynamic lattices yielded more uniform averages of 3.9 and 4.1. This trend toward stability was mirrored in other metrics; TR averages were 1.5 and 1.0 vs 1.2 and 1.3; VPR averaged 6.6 and 3.8 vs 6.2 and 5.1; and TMD averaged 0.32 and 0.21 vs 0.26 and 0.33. Shot size (4 mm vs. 8 mm) significantly impacted all metrics in static planning (p0.05).
Conclusion
Dynamic lattice construction facilitates more robust planning by ensuring consistent EUD, VPR, TMD, and TR values independent of sphere size or quantity. This optimization approach represents a clinically viable and reliable method for SFRT planning.