两种算法在非小细胞肺癌调强放疗中的剂量学比较

目的 比较两种算法在非小细胞肺癌（NSCLC）调强放疗（IMRT）中的剂量学差异，为非小细胞肺癌IMRT计划设计提供数据参考。方法 选择确诊为NSCLC的患者10例，分别制定采用Sliding Wnd（SLD）和Smart Sequence（SMT）算法的IMRT计划，比较两种算法在IMRT计划中的剂量分布、机器跳数、治疗时间以及二维剂量验证结果。结果 SLD算法的计划靶区D_2%、D_98%、D_mean、CI、HI值均优于SMT算法（P<0.05），差异有统计学意义；SLD算法双肺的D_mean、V₂₀、V₃₀、V₄₀、V₅₀均优于SMT算法，差异有统计学意义；心脏、脊髓等其它参数相差不大，差异无统计学意义。SMT算法的机器跳数与SLD算法相比减少了48 MU，治疗时间减短了7.2 s，差异有统计学意义。SLD算法和SMT算法的IMRT计划二维剂量验证3%/3 mm的γ通过率分别为（99.2±0.5）%和（99.3±0.6）%，2%/2 mm的γ通过率分别为（95.1±1.5）%和（96.4±1.8）%（P>0.05），差异无统计学意义。结论 两种算法的靶区剂量分布、危及器官受量以及二维剂量验证通过率均能满足临床要求，可应用于非小细胞肺癌IMRT计划设计。SLD算法的靶区剂量分布以及对双肺的保护均优于SMT算法，在NSCLC调强放疗中可优先考虑SLD算法。

Dosimetric comparison of two algorithms in intensity-modulated radiation therapy for NSCLC

Objective To evaluate the dose distribution of two algorithms in intensity-modulated radiation therapy (IMRT) for a basis for treatment planning of non-small cell lung cancer (NSCLC).Methods Ten NSCLC patients were selected in this analysis. IMRT plans were created with Sliding Wnd (SLD) algorithm and Smart Sequence (SMT) algorithm. The dose distribution, MU, treatment time and result of 2D dose verification were the parameters for comparison between those two plans.Results D_2%, D_98%,D_mean, CI and HI of PTV from SLD algorithm were superior than that with SMT algorithm (P﹤0.05). D_mean, V₂₀, V₃₀, V₄₀, V₅₀ of lungs with SLD algorithm were better than that with SMT algorithm (P﹤0.05), and no significant difference in other parameters was observed between the two algorithms. The MU of SMT algorithm was reduced by 48 and the treatment time was shortened by 0.12 min compared to SLD algorithm. The dosimetry of 3%/3 mm γ passing rate of SLD and SMT were (99.2±0.5)% and (99.3±0.6)%;, and 2%/2 mm γ passing rate of SLD and SMT were (95.8±2.3)% and (95.7±2.5)%.Conclusion The dose distribution of target, radiation dose to OAR and dosimetry passing rate of two algorithms can meet the clinical requirements, all of them can apply in IMRT for NSCLC. The dose distribution of target and protection of lungs with SLD algorithm were better than that with SMT algorithm. Our study reveals that SLD algorithm favors a more precise IMRT planning for NSCLC.