Room: AAPM ePoster Library
Purpose: investigate, under a robust beam model building process, the impact of optimizing the effective target spot size on dose prediction accuracy of AcurosXB model for SRS and SBRT treatment.
Methods: AcurosXB models, A and B, of 6XFFF were built in Eclipse (ver 13.6), using vendor representative data and in-house measured output factors down to 1x1cm2. The spot size of Model-A was optimized, 1.0 and 0.5mm in X and Y, such that the calculated PDDs and profiles of MLC defined small fields match the best to the measured. The spot size of model-B was vendor recommended, 1.5 and 0mm in X and Y. MLC dosimetric leaf gap of each model was subsequently optimized such that the calculated and measured inline profiles of MLC defined small fields match in shoulder area. Both models were each used to calculate 14 SRS and 8 SBRT plans. The discrepancies between the predicted dose and measured were examined. Measurement details include PTW Microdiamond in water tank for small field PDDs, profiles and output; and A16 micro-ion-chamber and GafChromic-XD films in anthropomorphic phantoms for treatment dose.
Results: predicted point dose by both models was within 3% of the measured in all SBRT and single-target SRS cases, and 4.5% for multi-target SRS cases. The average dose difference between the predicted and measured was -0.5±1.1 and 0±1.0% for Model-A and Model-B in single target cases, and 0.6±0.6 and 2.4±0.8% in multi-target SRS cases. The gamma pass rate at 2mm2% of all cases were within ~98% between the measured and predicted by both models.
Conclusion: spot size in building AcurosXB models could improve dose prediction accuracy for multiple target SRS cases, but the effect is minimal for single target SRS or SBRT cases as long as the robust model building process is followed.