293 - Enhancing Proton ARC Treatment Efficiency with a Multi-Arc Approach

J. Ma¹, Y. Zhu², Y. Lin², M. Tang¹, and H. Gao²; ¹Institute of Natural Sciences and School of Mathematics, Shanghai Jiao Tong University, Shanghai, China, ²Department of Radiation Oncology, University of Kansas Medical Center, Kansas City, KS

Purpose/Objective(s): Proton ARC therapy has garnered significant attention due to its ability to improve dose and linear energy transfer (LET) conformality to treatment targets. However, frequent energy layer changes, especially energy jumps, can hinder treatment delivery efficiency. In this study, we propose a novel proton ARC treatment planning method that employs a multi-arc strategy without energy switching within each arc (rotation). The objective of employing the multi-arc and constant-energy-per-arc method is to efficiently deliver proton ARC while simultaneously optimizing plan quality. Materials/

Methods: The proposed method initiates with optimization degrees of freedom equivalent to the number of energy layers, employing one arc per energy layer. Subsequently, a subset of arcs (typically 3 to 5) is selected using a novel block orthogonal matching pursuit (BOMP) optimization algorithm. The spot weights on these selected arcs are then optimized while adhering to the minimum-monitor-unit deliverability constraint. To validate the method, the proposed BOMP algorithm is compared with (1) the heuristic-search (HS) method, where each arc is individually optimized and then the arcs with the best plan quality are jointly selected for final plan optimization; and (2) the conventional (CONV) method, employing a variable-energy arc with multiple energy layers per beam angle.
Results: Compared to HS, BOMP has comparable or even better plan quality, which validates the effectiveness of BOMP algorithm. Compared to CONV, BOMP provides superior plan quality, and a substantial reduction of the energy layer switching time to a few seconds in total: for example, in a head-and-neck (HN) case, when the same target normalization was applied, BOMP had a reduction of optimization objective value from 0.33 to 0.26 (using 3 arcs) or 0.23 (using 5 arcs), max target dose from 117.37% to 113.10% or 110.84%, and energy layer switching time (T_E) from 759.4s to 1.4s or 2.8s which is negligible compared to CONV. Similarly, the results of lung case also show a reduction of optimization objective value from 0.25 to 0.19, max target dose from 118.97% to 115.72% and a substantial reduction of T_E from 1090.5s to 1.4s.
Conclusion: For the first time, a multi-arc and constant-energy-per-arc method is proposed for proton ARC RT. Since there is no need to switch energies in each arc, the novel proton ARC treatment planning method can substantially reduce energy layer switching time to a negligible level, without compromising the plan quality (in fact with an improvement of plan quality), compared to CONV.

Abstract 293 – Table 1

	Quantity	F	CI	DMAX(%)	DOAR(Gy)	Dbody(Gy)	TE(s)
HN	CONV	0.33	0.84	117.37	0.16	0.064	759.4
	HS3	0.29	0.79	113.83	0.17	0.081	1.4
	HS5	0.28	0.83	113.47	0.17	0.080	2.8
	BOMP1	0.53	0.67	120.61	0.17	0.084	0
	BOMP3	0.26	0.80	113.10	0.15	0.085	1.4
	BOMP5	0.23	0.84	110.84	0.15	0.083	2.8
Lung	CONV	0.25	0.91	118.97	0.12	0.059	1090.5
	HS3	0.19	0.85	119.94	0.14	0.070	1.4
	BOMP1	0.71	0.73	148.59	0.15	0.074	0
	BOMP3	0.19	0.87	115.72	0.14	0.070	1.4