Possibility of increasing linear energy transfer in multi-ion radiotherapy for bone and soft tissue sarcomas

Multi-ion radiotherapy using carbon, oxygen, and neon ions aims to improve local control by increasing dose-averaged linear energy transfer (LET_d) in the target. However, there has been limited understanding of how to utilize variables for multi-ion treatment planning such as the selection and arrangement of ion species.

An in silico study was conducted to explore the feasibility of increasing a minimum LET_d, and the optimal selection and arrangement of ion species in multi-ion therapy for increasing LET_d in tumors of varying sizes mimicking bone and soft tissue sarcomas (BSTS). Additionally, the robustness of multi-ion therapy against setup and range errors was evaluated.

Spherical targets of 500, 1000, and 1500-mL volumes were placed at the center or 80 mm horizontally displaced from the center of a numerical phantom to simulate BSTS treatments. Treatment plans were made for these targets with two orthogonal fields of carbon-only, oxygen+carbon, and neon+carbon ions with a total dose of 70.4 Gy (RBE). The treatment parameters were optimized to increase the LET_d in the targets while ensuring adequate target dose coverage and dose homogeneity. The plans were evaluated based on the dose covering 95% of the target (D_95%), skin dose (D_skin), and the minimum LET_d excluding the 1 mL volume with the lowest LET_d (L_1mL). Multi-ion radiotherapy treatment plans were also developed for 12 patients with BSTS who had previously received carbon-ion radiotherapy. D_95% and L_1m of the target, and the dose to organs at risk (OARs) such as the rectum, intestine, and spinal cord were assessed. The robustness of the plans created in the phantom against setup and range errors was evaluated under 2 mm shifts in six directions combined with 2.5% variation of the stopping power ratio, resulting in 12 scenarios. Differences in the target D_95% and L_1mL, and D_skin in each scenario from those in the nominal plan were evaluated.

The target dose coverage was comparable for any ion species combinations regardless of target size and position. The L_1mL in the target increased by 7-9 and 15-20 keV/µm with the oxygen+carbon and neon+carbon plans, respectively, compared to the carbon-only plans, while maintaining homogeneity index values below 0.10. Additionally, the skin dose increased by 2.2-7.0 and 9.2-14.6 Gy (RBE) for the oxygen+carbon and neon+carbon plans. The L_1mL was greater than or equal to 40 keV/µm in all phantom targets and most clinical cases for the oxygen+carbon and neon+carbon plans, while meeting the target and OAR dose requirements. In the robustness evaluation, the variations in D_95% were comparable or smaller in the oxygen+carbon and neon+carbon plans than in the carbon-only plan. The maximum decrease in L_1mL in the target was 1.5 keV/µm. The maximum increase in D_skin was 2.4 Gy (RBE) in the target closest to the skin.

The LET_d was successfully increased with the oxygen+carbon and neon+carbon ions, while meeting the dose requirements. The multi-ion therapy plans created using the method presented in this study were robust to setup and range errors.