Adaptive particle swarm optimizer for beam angle selection in radiotherapy planning

As an emerging stochastic optimization paradigm, particle swarm optimization (PSO) algorithm has received a lot of attention in recent years. In this paper, a method named adaptive PSO is introduced to automatically select the beam angles for intensity-modulated radiotherapy (IMRT) planning. To date, the improvements in automatic beam angle selection are still not quite satisfying, especially on the optimization efficiency because of the handicap of huge hyperspace of solutions. In the proposed algorithm, the beam angles are selected using PSO, followed by a beam intensity map optimization using conjugate gradient (CG) algorithm for each updated beam configuration. This PSO-based algorithm is verified by a relatively complex clinical prostate tumor case. In addition, the efficiency is compared with a genetic algorithm (GA)-based approach. The preliminary results show that the proposed algorithm is feasible for the beam angle selection problem in IMRT planning. Furthermore, PSO appears more efficient than GA, according to the limited test case in this paper. Further study is needed to test the proposed method with more clinical cases.