AAA和PDIP算法在非均整模式容积调强放射治疗剂量预测方面的差异

Differences between anisotropic analytical algorithm and portal dose image prediction algorithm in dose prediction in volumetric modulated arc therapy using flattening filter free beams

目的:探究各项异性算法(AAA)和射野剂量图像预测(PDIP)算法在非均整模式(FFF)容积调强放射治疗计划治疗前验证γ分析中的差异以及计划复杂程度对这种差异的影响,为临床上基于电子射野影像系统(EPID)的剂量预测算法的选择提供依据。方法:选取能量为6 MV FFF的两种测试野和16例头颈部肿瘤治疗计划,利用PDIP和AAA两种算法分别生成预测数据并与EPID实测数据进行γ分析,统计两种算法在不同γ评判标准下的通过率并计算通过率差异(Deltaγ)。计算上述病例每个射野的复杂系数,分析不同标准下两种算法的Deltaγ与复杂系数的相关性;利用γmean、γsd、γ1和γ通过率共同描述γ分布,并分析其与复杂系数间的相关性。结果:当评判标准为3%/3 mm或2%/2 mm时,不同算法下测试射野的Deltaγ较小。当评判标准为1%/1 mm,不同开野的Deltaγ变化明显:射野较小时,PDIP算法的通过率低于AAA;当射野增大到(10×10)cm2时,通过率基本一致;当射野继续增大时,PDIP算法的通过率逐渐高于AAA。全部射野的通过率与评判标准的关系类似:在3%/3 mm标准下,两种算法的结果基本一致;随着标准的提高,两种算法的通过率逐渐下降,二者之间的差异也逐渐明显。复杂系数与Deltaγ、γmean、γsd和γ1为正相关,与γ通过率为负相关。结论:PDIP算法对于有机械臂支撑的EPID的剂量预测更准确;AAA则适用于无机械臂支撑的EPID或机械臂反散射影响较小的射野。当计划复杂程度或评判标准提高时,两种算法的差异也增大。计划复杂程度对FFF计划验证结果的影响是负面的。上述结果提示临床应针对性地选择计划验证工具来确保治疗的安全有效。

Objective To explore the difference between anisotropic analytical algorithm(AAA) and portal dose image prediction(PDIP) algorithm in γ analysis for the pre-treatment verification of volumetric modulated arc therapy using flattening filter free(FFF) beams and to investigate the effects of plan complexity on the difference for providing a basis for the selection of dose prediction algorithm based on electronic portal imaging device(EPID). Methods Two test fields and 16 head and neck treatment plans using 6 MV FFF were selected. AAA and PDIP algorithm were used to generate predicted data, and a γ analysis between predicted data and the data measured by EPID was conducted. The Delta γ of the two algorithms under different γ criteria were analyzed. The complexity metric(CM) of each field in the above cases was calculated, and the correlation between Delta γ and CM under different γ criteria was investigated. Finally, the γ distribution was described by γmean、γsd、γ1 and γ passing rates, and their correlations with CM were discussed. Results For the test field under the criterion of 3%/3 mm or 2%/2 mm, the Delta γ of two algorithms was small. When the criterion was 1%/1 mm, there was significant difference in Delta γ for different open fields. The passing rate of PDIP algorithm was lower than that of AAA when the field was smaller than(10×10) cm2;and the passing rates of two algorithms was basically the same when the field was(10×10) cm2;with the further enlargement of field, the passing rate of PDIP algorithm was significantly higher than that of AAA. The relationship between the passing rates of all fields and evaluation criteria was similar. Under the criterion of 3%/3 mm, there was trivial difference in the passing rate of two algorithms. As the criterion increases, the passing rates of two algorithms were gradually decreased and the difference between them become obvious.The CM was positively correlated with Delta γ, γmean、γsd、γ1, and negatively related with γ passing rate. Conclusion PDIP algorithm has a higher accuracy for dose prediction based on EPID with arm support, while AAA is suitable for dose prediction based on EPID without arm support or field with less backscattering effect. When plan complexity or evaluation criterion increases, the difference between two algorithms increases. The plan complexity has a negative effect on FFF plan verification. An appropriate plan verification tool should be adopted to ensure the safe and effective treatment in clinic.