Assessing the potential and pitfalls of spot sequence optimization for OAR-specific dose rate control in proton PBS Bragg peak FLASH radiotherapy.

[PURPOSE] To evaluate the impact of key treatment planning parameters-including beam arrangement, minimum monitor unit (MMU) settings, and anatomical site variability-on the ability to achieve ultra-high dose rate (UHDR) delivery in Bragg peak FLASH radiotherapy.

[METHODS] FLASH dose rate coverage can be assessed by dose-rate volume histogram (DRVH), thus objective functions based on DRVH can be constructed to optimize the dose rate distribution for individual regions of interest (ROIs). The optimization of each ROI, as part of the final objective function, is integrated into a multi-objective optimization problem that can be solved using a heuristic algorithm. A phantom-based study was conducted to investigate the effect of beam number on optimization performance in proton therapy planning. Treatment plans of 8 consecutive node-negative non-small cell lung cancer and 5 consecutive liver cancer patients were initially optimized, followed by optimizing the spot delivery sequence to enhance dose rate ratios without compromising dose performance. A thorough evaluation was conducted to assess the optimization of the scanning pattern in improving the FLASH ratio of critical OARs in Bragg peak FLASH-RT, considering beam currents, beam arrangement, MMU constraints, and anatomical sites in lung and liver cases.

[RESULTS] The phantom study demonstrated that the effectiveness of the spot pattern in dose rate depends on the number of beams and beam arrangement, and the 3-field arrangement can achieve better optimization effects. In lung cases, using a MMU of 600 (nozzle current of 252nA), scanning pattern optimization increased the average dose rate (V40Gy/s) for the esophagus, heart, spinal cord, and lung-GTV from 38.3 %, 62.8 %, 59.6 % and 61.9 % to 74.4 %, 85.5 %, 83.3 % and 78.6 %, respectively (all p-values < 0.001). When a higher MMU of 1200 (nozzle current of 504nA) was used, the benefits brought by optimization are not as obvious as the previous situation. For all liver cases with an MMU of 600, the average FLASH dose rate (V40Gy/s) for the esophagus, heart, spinal cord, and liver-GTV increased from 60.5 %, 52.7 %, 60.3 %, and 59.1 % to 75.1 %, 69.4 %, 80.2 %, and 75.9 %, respectively, after optimization (all p-values < 0.001). However, when a higher MMU of 1200 was used, the V40Gy/s for all four OARs increased from approximately 93.3 % to 97.0 %, showing only limited additional improvement.

[CONCLUSION] This approach successfully optimized FLASH dose rate coverage for specific OARs, enhancing BP-FLASH effectiveness by improving OAR protection while maintaining dosimetric quality. However, the impact of spot pattern optimization is influenced by factors such as the number of beams, MMU constraints, and spot distribution, with limited effectiveness in significantly increasing the FLASH ratio.