Introduction:
The Gamma Knife (GK) was originally introduced as a radiotherapy unit for solid brain lesions. The original skull-attached G-frame enabled submillimetre accuracy but allowed only single-fraction treatments, thus limiting to small lesion treatments. To enable fractionated Stereotactic Radiosurgery, the GK Perfexion model (with eXtend) was introduced leading to new questions:
• What is the repositioning system accuracy?
• What effect have positional uncertainties for the target; and for normal tissue (NT) and Organs at risk (OAR)?
• How can positional uncertainties be handled?
Material and Method:
eXtend uncertainty was investigated using trigger levels of the vacuum-surveilled patient positioning mouthpiece, measured for ten patients’ dental models mounted on a 0.05mm-accuracy computer-controlled positioner. The dosimetric effect of these displacements in a single fraction was evaluated by calculating minimum dose, coverage and conformity-index changes for displacements of 0 to 4mm.The impact of positional displacements for 3 and 5 fraction treatments was evaluated for targets and organs-at-risk (OaR) by calculating total doses with simulated displacements of 0-4mm in combinations of X, Y, and Z directions for nine patients (in total almost 300 sumPlans). All plans were prescribed to the 50% isodose. Effects were calculated in physical dose and in biological effective dose (BED).
Finally, from this a novel approach to minimize effects of displacement was proposed and feasibility-tested for 5-fraction plans. Instead of applying margins to cover potential uncertainties, a correction is applied in the last two fractions only after any observed systematic uncertainty is quantified in the first three fractions. 100 displacement scenarios were tested, comparing for each the dose volume histogram (DVH) and dose profiles and the total dose with and without the correction procedure.
Results:
Mean positional uncertainty with a clinical vacuum setting was 0.15mm (SD ±0.05mm, range 0.05-0.29mm) and 0.33° (SD ±0.15°, 0.05°-1.0°).
The most critical parameter in the dose distribution is minimum dose D99%. In the evaluated scenarios D99% fell by 2% for a displacement of 0.5mm and 16% with 2mm in X or Y direction and >20% in Z direction, dependent on arrangement.
Moving to hypofractionation increases the BED inside the target and reduces it outside, i.e. the dose gradient is increased at the 50% (prescription) isodose. Displacement in opposing directions is predominantly random, which reduces the dose gradient but does not introduce an underdosage to the target for displacements of up to 2 mm. At the same time the BED to OAR is increased but remains below single fraction dose.
Feasibility testing the proposed correction method showed a significant improvement of the total BED indicating the method’s potential.
Discussion:
Positional uncertainty is generally low, better than 0.5mm for Perfexion whilst the later GK version, Icon. starts treatment with minimal uncertainties following an initial Cone Beam CT (CBCT). However, for both treatment units, target movement during treatment is possible and may be up to 2mm. The most critical aspect is underdosage of the target. Prescribing the dose to 50% reduces random uncertainty effects and ensures BED is at an optimal value (gradient) to avoid underdosage and protect OAR.
A novel strategy is proposed and feasibility-tested to correct positional uncertainties, showing improved dose distribution in the scenarios tested. However, this test was for static displacements only. GK Icon corrects any positional displacement observed at the initial CBCT, so the correction method would only be needed when the system identifies further movement. For practical clinical use, further investigation is needed.
