Dynamic field shaping for stereotactic radiosurgery: a modeling study

Abstract

Purpose: This work assesses the relative field shaping advantages of dynamic field shaping devices for stereotactic radiosurgery using a linear accelerator.

Methods and materials: We selected 43 intracranial tumors (2.0-4.2 cm maximum dimension, 1.5-25.5 cc tumor volume) out of the first 64 intracranial tumors treated with radiosurgery at the Joint Center for Radiation Therapy. We modeled five field shaping devices, each including a fixed auxiliary circular collimator: (a) fixed circular collimator alone; (b) two independent parallel jaws; (c) four independent rectangular jaws; (d) four independent rotatable jaws; and (e) "ideal" multileaf collimator. We adjusted the model parameters until the minimum target isodose was 80% of the dose delivered to isocenter. We defined the treatment volume ratio as the target volume divided by the treatment volume (volume receiving at least the minimum target dose). We used the treatment volume ratio to compare the five models and the actual patient treatments.

Results: For 34 tumors originally treated with one isocenter, the median Treatment Volume Ratio was higher for all of the device models except the fixed circular collimator compared to the actual patient treatments. For the nine tumors originally treated with multiple isocenters, the median Treatment Volume Ratio for the actual multiple isocenter treatments was similar to that for two parallel jaws, four rectangular jaws and four rotatable jaws. Only the median "ideal" collimator treatment volume ratio was higher for these nine tumors.

Conclusion: Simple field shaping devices have approximately 50% of the conformal advantage of an "ideal" multileaf collimator. Approximately 50% of typical radiosurgical tumors between 2 and 4 cm have field shaping advantages which exceed the geometrical uncertainties inherent in linear accelerator radiosurgery treatments. The three models, two parallel, four rectangular, or four rotatable independent jaws would improve current linear accelerator technology by providing homogeneous doses with equivalent field shaping for most tumors originally treated with inhomogeneous multiple isocenter plans (6/9 tumors in the current series).