2967 - Dosimetric Evolution In Nasopharyngeal Carcinoma: Comparing Conventional VMAT with VMATp Using Dynamic Collimator Rotation and Variable STAMP Configurations

Purpose/Objective(s):

Balancing organ-at-risk (OAR) sparing with potent tumor doses is challenging in nasopharyngeal carcinoma (NPC). This study evaluated the efficacy of VMATp, a novel optimization platform featuring dynamic collimator rotation and IMRT-like static angle modulated ports (STAMPs). We assessed whether VMATp, configured with one (VMATp 1) or five (VMATp 5) STAMPs, provides superior sparing and target homogeneity compared to conventional volumetric modulated arc therapy (VMAT) without compromising clinical coverage.

Materials/Methods:

Twenty Stage III-IVA NPC patients were replanned under a pre-clinical prototype planning system. Prescription doses were 70 Gy and 46 Gy to high-/low-risk target volumes. Three techniques were compared: (1) 2-arc VMAT; (2) 2-arc VMAT with 1 STAMP (VMATp 1); and (3) 2-arc VMAT with 5 STAMPs (VMATp 5). VMATp optimization utilized 2000 iterations, identified as the convergence threshold in our preliminary analysis. Dosimetric coverage, homogeneity index (HI), conformity index (CI), and 41 OAR constraints were evaluated. Optimization time and monitor units (MU) were recorded. Statistical comparisons were performed using repeated-measures ANOVA or Friedman tests.

Results:

All techniques achieved clinically acceptable coverage. Compared to VMAT, VMATp demonstrated superior target coverage (D_99%: 69.66 ± 1.38 Gy [VMATp 1] and 69.68 ± 1.42 Gy [VMATp 5] vs. 69.11 ± 1.12 Gy [VMAT]; p < 0.001) and homogeneity (HI: 0.08 ± 0.02 [VMATp 1] and 0.07 ± 0.02 [VMATp 5] vs. 0.09 ± 0.02 [VMAT]; p < 0.001). Although VMAT maintained better conformity (CI: 1.06 ± 0.03 vs. 1.09 ± 0.03 [VMATp 1] and 1.08 ± 0.04 [VMATp 5]; p < 0.001), VMATp achieved significantly higher OAR pass rates (29.85 ± 4.12 % [VMATp 1] and 29.90 ± 4.04 % [VMATp 5] vs. 28.10 ± 4.19 % [VMAT]; p < 0.001).

Notably, VMATp significantly reduced critical neurological and swallowing structure doses (p < 0.001). Compared to VMAT (59.61 ± 5.81 Gy), brainstem D_max was lowered to 56.75 ± 6.50 Gy (VMATp 1) and 56.77 ± 6.41 Gy (VMATp 5). Spinal cord D_max was reduced from 45.68 ± 6.34 Gy (VMAT) to 38.74 ± 5.58 Gy (VMATp 1) and 38.51 ± 5.53 Gy (VMATp 5). Significant sparing observed in superior-middle pharyngeal constrictor muscles (PCM) (57.21 ± 6.02 Gy [VMATp 1] and 57.48 ± 5.84 Gy [VMATp 5] vs. 58.96 ± 5.74 Gy [VMAT]; p < 0.001) and inferior PCM (24.86 ± 6.46 Gy [VMATp 1] and 24.67 ± 6.46 Gy [VMATp 5] vs. 26.87 ± 6.65 Gy [VMAT]; p < 0.001).

VMATp 1 offered the most optimal balance, with superior optimization speed (91.10 ± 16.56 s [VMATp 1] vs. 107.3 ± 19.6 s [VMATp 5] and 187.90 ± 31.98 s [VMAT]) and intermediate MU (748.05 ± 55.89 [VMATp 1] vs. 861.05 ± 56.63 [VMATp 5] and 586.40 ± 75.24 [VMAT]).

Conclusion:

The VMATp solution represents a significant evolution over standard VMAT for locally advanced NPC, providing superior protection of critical structures through dynamic collimator rotation and STAMP optimization. Our findings suggest that VMATp 1 is recommended for clinical implementation due to its superior therapeutic ratio and efficiency.