调强放疗中强度矩阵分解算法的研究

Highly-efficient greedy algorithm for field splitting in intensity-modulated radiation therapy

强度矩阵分解作为调强放疗(IMRT)中的一个核心技术,是一个NP-hard问题。调强一般分为静态模式和动态模式两种。针对静态调强放疗,对总治疗时间主要依赖的四个参数:总机器跳数(TNMU)、子野个数(NS)、多叶准直器叶片移动速率和验证记录(V&R)进行了研究,基于Xia&Verhey算法提出了一种高效的算法,能达到总治疗时间尽可能少的效果。经过随机与临床强度矩阵数据的大量测试,提出的算法与同类算法如Galvin算法、Bortfeld算法、Xia&Verhey算法、Siochi算法以及Luan算法相比,都能取得最小的总治疗时间,从而更具有实际可行性。

Field splitting is one of the core techniques in the intensity-modulated radiation therapy（ IMRT）,which is an NPhard problem. Intensity modulation includes two cases： static mode（ i. e,step and shoot） and dynamic mode. Regarding the static mode,this paper considered how to reduce the total treatment time of IMRT,which depended on four parameters： the total number of monitor units（ TNMU）,the number of segments（ NS）,the MLC＇s speed,and the verification and recording time（ VR）. Based on Xia Verhey＇s algorithm,this paper presented an optimal algorithm by taking all these four factors into account. Experiments with random data and clinical data show that,compared with the classic algorithms proposed by Galvin,Bortfeld,Xia Verhey,Siochi and Luan,the new algorithm always produces optimal results on total treatment time and speeds up significantly.