Static proton arc therapy: Comprehensive plan quality evaluation and first clinical treatments in patients with complex head and neck targets

Abstract

Background

Proton Arc Treatment (PAT) has shown potential over Multi-Field Optimization (MFO) for out-of-target dose reduction in particular for head and neck (H&N) patients. A feasibility test, including delivery in a clinical environment is still missing in the literature and a necessary requirement before clinical application of PAT.

Purpose

To perform a comprehensive comparison between clinically delivered MFO plans and static PAT plans for H&N treatments, followed by end-to-end commissioning of the system to prepare for clinical treatments.

Methods

Anonymized datasets of 10 patients treated for H&N cancer (median prescription dose 70 GyRBE) were selected for this study. Both MFO and PAT plans were created in RayStation and robustly optimized for setup and range uncertainties as in our clinical routine. PAT plans were created with 30 angle directions.

1. Comparisons were performed regarding:

2. nominal dose distributions in terms of target coverage, dose to primary and secondary OARs

3. robustness evaluation (D₉₅ of the target and D₁ of primary OARs)

4. Normal tissue complication probability (NTCP) values for xerostomia, swallowing dysfunction, tube feeding, and sticky saliva

5. D·LET_d distributions

6. the probability of replanning at least once due to anatomical changes

7. delivery time: MFO and PAT plans, for one patient, were delivered in a clinical gantry room. For PAT, two plans with 30 and with 20 discrete beam directions were optimized and delivered.

Results

In PAT plans, a significant reduction was observed in the near maximum dose to the brainstem, while no statistically significant differences were found for other primary OARs or target coverage metrics (D₉₅ and D₉₈) in both nominal plans and robustness evaluation scenarios. For secondary OARs, PAT plans achieved an impressive reduction in mean dose. Max D·LETd distributions in brainstem, brain, and temporal lobes showed no statistical differences between MFO and PAT plans while mean D·LETd values were lower with PAT. Median NTCP was significantly reduced for xerostomia as endpoint (ΔNTCP = 8.5%), while reductions in other endpoints were not statistically significant. The number of patients that would need at least one replanning during the treatment for PAT was similar to MFO, showing that the established clinical workflow for monitoring of anatomy changes will remain the same for both delivery methods. Comparison in terms of delivery time from the start of the first beam until the end of the last (comprising all the technically motivated delays due to operation of OIS/Therapy Control System operation, gantry rotations, couch rotations, beam line preparation etc.) resulted in delivery times that were similar for both techniques.

Conclusion

Static PAT plans demonstrate the capability to increase plan quality with respect to state-of-the-art MFO planning, since dose reduction outside of the target is significant with no reduction of the quality of the target dose distribution. NTCP evaluations, as well as linear energy transfer (LET) distributions, do not indicate risks for unexpected toxicity. Delivery time tests with different beam direction configurations have shown that PAT plans can already be delivered within similar time slots as highly conformal MFO plans. The successful end-to-end commissioning led to the world's first patient treatments using PAT, with eight patients treated to date.