The dosimetric impact of trade-off optimization in high-quality Craniospinal Irradiation plans based on volumetric modulated arc therapy technique

This study assesses the dosimetric effectiveness of the commercial trade-off optimization (TO) module in comparison to iterative optimization for volumetric modulated arc therapy (VMAT) in craniospinal irradiation technique.Fifteen patients who had previously undergone VMAT-based craniospinal irradiation (CSI) using manual optimization (TP) underwent re-optimization with trade-off optimization (MCO). All patients were treated using the Halcyon-E O-ring linear accelerator, with maximum field size of 28×28 cm², a 6MV unflattened beam, and adjacent isocenter field overlap of 10 cm. Plans were compared based on PTV dose coverage (D95%), maximum dose (Dmax), conformity index (CI), heterogeneity index (HI), maximum and mean dose to serial and parallel organs, respectively. Statistical evaluation was conducted using paired sample t-tests. The PTVD95% for TO and MCO plans were 98.0% ± 1.0% and 97.4% ± 0.7%, respectively. In the same sequence, HIs were 1.06 ± 0.01 and 1.07 ± 0.01. CIs for both arms were 0.9 ± 0.0 and its variation was statistically significant (p = 0.027). The differences in dose for bilateral cochlea and left optic nerves were statistically significant (0.022≤ p ≤0.049). The ΔDmax for serial organs and mean dose for parallel organs did not exceed 1%, except for the bilateral optic nerve, mandible, oral cavity, right parotid, and stomach. No parallel organ showed a statistically significant dose variation. Clinically significant reductions in dose were noted for three organs; the average dose reduction in MCO plans for bilateral optic nerves was 3.9%, and for the larynx, it was 8.5%. In this study, trade-off optimization did not demonstrate any significant improvement over the iteratively optimized plans, primarily because the planners were highly skilled and could already generate high-quality plans using iterative optimization alone. However, this finding may not necessarily apply universally to all treatment planners or clinical settings.