Early generation dynamic and static proton arc treatment planning algorithms assessment in oropharyngeal cancer patients.

BACKGROUND

Compared with intensity modulated proton therapy (IMPT), proton arc treatment (PAT) employs an increased number of gantry angles, potentially reducing healthy tissues doses, especially for complex target geometries found in oropharyngeal cancer (OPC) treatment. PAT plans can be optimized with algorithms, based either on "static" gantry position or "dynamic" gantry movement during dose delivery. Recent results have shown target coverage may suffer more from inter-fraction patient anatomical- and setup changes In PAT than IMPT.

PURPOSE

We assessed if PAT plans generated with one static and two dynamic PAT planning algorithms can improve expected plan toxicity, delivery time, and inter-fraction robustness compared to clinical volumetric modulated arc therapy (VMAT) and IMPT plans for OPC patients.

METHODS

Six OPC patients were included that qualified for proton therapy based on model-based selection, with IMPT plans superior to the VMAT plans in terms of toxicities. Static PAT plans were produced using an energy layer filtration (ELF) algorithm, and dynamic PAT plans were produced with two different methods: (1) Spot scanning Proton Arc (SPArc), and (2) Early Layer and Spot Assignment (ELSA). Two sets of PAT plans with about 360 or 240 energy layers and an additional ELF plan employing anterior oblique range shifted fields were produced. All proton plans were robustly optimized. Expected plan toxicity was determined using normal tissue complication probability (NTCP) models for dysphagia and xerostomia. A delivery time model was calibrated using experimental machine log-files and gantry dynamics from an IBA Proteus PLUS system (IBA Ltd, Belgium). Inter-fraction robustness was evaluated on a fraction-wise (on weekly repeated CT) and course-wise (accumulated over all weekly repeated CTs) basis.

RESULTS

All PAT strategies showed significantly (p-values < 0.05) reduced NTCPs for dysphagia and xerostomia grade ≥ 2 compared to IMPT and VMAT. Relative to VMAT, NTCP for xerostomia reduced by on average 4.0% in IMPT plans, 9.8% in ELF-, 9.6% in SPArc-, and 8.7% in ELSA PAT plans with 360 energy layers, and NTCP for dysphagia reduced by on average 9.6% in IMPT plans, 13.1% in ELF, 12.9% in SPArc, and 12.3% in ELSA PAT plans. Average clinical IMPT delivery time was 11.4 ± 2.1 min, while dynamic PAT delivery time was on average 7.6 ± 0.5 min for ELSA and 8.2 ± 0.5 min for SPArc plans with 360 energy layers and auto beam sequencing delivery for 360 energy layer 30 beam angle ELF plans was 9.9 ± 1.0 min. Reduction of energy layers to 240 resulted in limited NTCP increase, globally < 1%, and reduced delivery time by up to 3 min, where changes in delivery time and NTCP were largest in ELF plans. Fraction-wise target coverage was worse in the PAT plans compared to IMPT; however the course-wise coverage showed to be less impacted by inter-fraction changes. The addition of anterior oblique fields with range shifters markedly improved fraction- and course-wise target coverage for the ELF plans.

CONCLUSIONS

The tested static and dynamic PAT planning strategies showed similar significant reductions in NTCP compared to IMPT and VMAT. Estimated PAT delivery times were shorter compared to times for current delivery procedures of clinical IMPT plans. Dynamic PAT delivery is faster than static PAT. Fraction-wise robustness suffered more than course-wise robustness due to anatomical changes found in repeated CTs. Range shifters can be employed to improve PAT plan robustness.