Room: AAPM ePoster Library
Purpose: evaluate the relevance between various calculated modulation indices for predicting the delivery accuracy and patient-specific QA measurement results for volumetric modulated arc therapy (VMAT).
Methods: treatment planning data set of 129 patients with head and neck cancer were analyzed. These VMAT plans using 6-MV linear accelerator were performed on Eclipse treatment planning system (Varian, USA), and the prescribed dose was 66-70 Gy in 33-35 fractions at D95 of the planning target volume. The patient-specific QA of each treatment plan was measured using Delta4 system (Scandidos, Sweden), and then the median value of dose difference (DD) and the local gamma passing rate (GPR) with gamma criteria of 3%/2mm were calculated.We evaluated six modulation indices including the modulation complexity score for VMAT (MCSv) from Masi et al.; the plan-averaged beam irregularity (PI), the plan-averaged beam aperture (PA) and the plan-averaged beam modulation (PM) from Du et al.; the ratio of the average area of an aperture over the area defined by the jaws (AAJA) from Lam et al. In addition, we proposed a new modulation index that the plan-averaged velocity of the multi-leaf collimator (PV) was calculated using Python 3.7. Correlations between these modulation indices and DD, GPR were analyzed.
Results: correlation coefficients (r) were -0.300, 0.760, 0.757, 0.461, -0.371, 0.679 between the DD and MCSv, PI, PA, PM, AAJA, PV, respectively. The DD tended to increase linearly with increasing PI, PA, PM, PV. PI was the strongest correlation (r=-0.672) with the GPR among the modulation indices, on the other hand, both MCSv and AAJA were a weak correlation (r=0.366 and 0.332, respectively) with the GPR.
Conclusion: has a moderate correlation with DD and GPR, in our results, suggesting that it is one of the indicators suitable for predicting delivery accuracy for VMAT.