Dosimetric Impact of Systematic and Random Errors With Multi-Leaf Collimator Leaf Positioning in Intracranial Stereotactic Radiosurgery

Abstract

Modern medical Linac equipped with multileaf collimators (MLC) are widely used for intracranial stereotactic radiosurgery (SRS). This study aimed to evaluate dosimetric impact of systematic and random positioning errors with MLC leaves and to establish practical limits on MLC leaf positioning uncertainties in SRS treatments.A single institution retrospective review identified 10 clinical single-target conformal arc SRS plans and 7 single- or multiple-target VMAT SRS plans treating a total of 15 PTVs. Each conformal arc SRS plan was prescribed 20 Gy in one fraction to a PTV using 2-5 arc fields with PTV volume ranging from 0.10 to 2.48 cc, and each volumetric modulated arc therapy (VMAT) SRS plan was prescribed 20 - 27 Gy in 1 - 5 fractions to 1 - 4 PTVs, with PTV volume ranging from 0.11 to 7.73 cc. 6-MV flattening-filter free photon fields were used in all the plans. A 120-leaf MLC was used for beam modulation with a leaf width of 2.5 mm for the central 32 leaf-pairs and 5.0 mm for the peripheral 28 leaf-pairs. Systematic or random leaf positioning errors ranging from 0.1 to 1.0 mm at 0.1-mm intervals were introduced to each SRS plan. Plan dose was re-calculated with original field MUs with a 1.0 mm calculation resolution using a linear Boltzmann transport algorithm and was verified using another dose calculation engine with a convolution superposition algorithm. Variations in PTV dosimetric parameters including mean dose and D(95%) due to systematic and random leaf positioning errors were analyzed. PTV dose variation was also verified with measurement using an ion chamber in water equivalent phantom.Both the PTV mean dose and D(95%) decreased linearly with increasing systematic and random positioning errors in all the SRS plans. In systematic leaf positioning errors with a 1-mm shift towards the isocenter for all MLC leaves, the PTV mean dose and D(95%) decreased by an average of 8.7% (range: 5.1 - 12.4%) and 15.2% (range: 10.4 - 20.1%) respectively in the conformal arc plans; the PTV mean dose and D(95%) decreased by an average of 18.2% (range: 11.6 - 25.0%) and 21.8% (range: 15.1 - 27.3%) respectively in the VMAT plans. The PTV volume positively correlated with the declining rate of mean PTV dose (correlation coefficient: 0.80, P-value = 0.01) with the conformal arc SRS plans. In random leaf positioning errors, with a standard deviation of up to 1 mm, the decrease of PTV mean dose and D(95%) was up to 2.4% and 3.3% respectively in the conformal arc plans, and decreased up to 2.7% and 2.8% respectively in the VMAT plans.Manufacturers' specifications for MLC leaf positioning accuracy are not adequate for accurate intracranial SRS dose delivery. To ensure that both the PTV mean dose and D(95%) are within 5% from plan dose, the systematic MLC leaf positioning error needs to be < 0.3 mm for conformal arc SRS plans and < 0.2 mm for VMAT SRS plans. On the other hand, random leaf positioning errors have smaller impact on dose delivery accuracy within the limits specified by manufacturers.C. Han: None. J.R. Liu: None. D. Du: None. J. Liang: None. K. Qing: None. W.T. Watkins: None. S. Zhang: None. A. Liu: None.