SRS using linac and cones offers steep dose fall-off but a tradeoff exists between conformality and treatment time, which depends on the number of isocentres. Purpose of this study is to quantify planning metrics between cones- and MLC-based SRS for arteriovenous malformation(AVM).
Seven AVM cases treated with cones were re-planned with MLC on Pinnacle treatment planning system. Planning target volume(PTV) was created with 1mm uniform margin to the AVM to account for MLC positional variation. Clinically-planned prescription dose(15–25Gy) was used. Four plans were generated per case:non-coplanar VMAT(ncV), single-arc VMAT(saV), non-coplanar IMRT(ncI), non-coplanar conformal(ncC). Plans were compared for conformity(CI), heterogeneity(HI) and gradient(GI) indices and brain doses. Estimated treatment times and monitor units(MU) were compared.
Cone-based plans required 2–6 isocentres. Though CI-RTOG was similar for plans(median=0.98), CI-Paddick was most favourable for ncV(median=0.86) and worst for cones(0.54). HI for MLC plans(median=1.19–1.27) were lower than cone-based plans(1.43). GI was similar for all plans. For 2/7 ncC had brainstem maximum dose>16.7Gy and therefore were clinically unacceptable. Brain V12Gy,V10Gy,V2Gy were lowest in the cones plan. ncV brain V12Gy,V10Gy,V2Gy were lowest of all MLC-based plans studied. Treatment MUs were similar for MLC-based plans and up to 70% lower than clinically delivered plans.
ncV showed best conformality in this study. Of the MLC-based plans, ncV also showed lowest normal tissue dose with reasonable treatment time.