Braunstein Steve E, Dionisio Sebastian A, Lometti Michael W, Pinnaduwage Dilini S, Chuang Cynthia F, Yom Sue S, Gottschalk Alexander R, Descovich Martina
University of California, San Francisco, Department of Radiation Oncology, 1600 Divisadero St., Suite H1031, San Francisco, CA 94143, USA.
J Radiosurg SBRT. 2014;3(1):67-79.
PURPOSE/OBJECTIVE: Dose calculation in treatment planning must account for tissue heterogeneity, especially for tumors within low-density lung tissues. While Monte Carlo (MC) calculation methods are the most accurate, Ray Tracing (RT) methods are also commonly employed. We evaluated dose calculation differences between the RT and MC algorithms in central and peripheral lung tumors treated with CyberKnife SBRT to determine which planning parameters may predict dose differences. We also examined clinical outcomes of local-regional control (LRC) and long-term treatment-related toxicity as a function of calculation method.
MATERIALS/METHODS: A retrospective series of 70 patient plans (19 central and 51 peripheral lung lesions) treated between 2009 and 2011 were analyzed. Among those, 33 were primary lung cancer and 37 were metastatic lesions. Thirty-three treatment plans were developed with the RT method, and 37 plans used MC. Groups were recalculated with the reciprocal method for dose comparison. Parameters examined to quantify dose differences between the two algorithms included: dose delivered to 95% (D95) of the planning target volume (PTV), dose heterogeneity, and dose to organs at risk (OAR). Dose differences were analyzed as a function of target volume, distance to soft tissue, and fraction of target overlap with soft tissue. For the subset of primary lung tumors, LRC was assessed radiographically at a median follow-up of 19 months (mo) (range, 2 to 41 mo).
Compared to MC, the RT algorithm largely overestimated the dose delivered to the PTV. The dose difference between RT and MC plans correlated to the volume of PTV overlapping with soft tissue; the smaller the overlap volume, the larger the dose differences between RT and MC. Compared to MC, the RT algorithm overestimated the dose delivered to 10% of the ipsilateral lung (D10%). Evidence of local progression was noted in only one of the 31 patients treated for primary lung malignancy. DFS and OS were not significantly different between RT and MC plans.
There is a significant range of discordance between MC and RT dose calculations for SBRT treated peripheral lung tumors. While variation is correlated to target size and proximity to soft tissue, no single parameter can reliably predict dose differences. Ultimately, local control and long-term toxicity appear independent of the dose calculation method.
目的/目标：治疗计划中的剂量计算必须考虑组织异质性，尤其是对于低密度肺组织内的肿瘤。虽然蒙特卡罗（MC）计算方法最为准确，但光线追踪（RT）方法也常用。我们评估了射波刀立体定向体部放疗（CyberKnife SBRT）治疗的中央型和周围型肺肿瘤中RT和MC算法之间的剂量计算差异，以确定哪些计划参数可预测剂量差异。我们还研究了局部区域控制（LRC）的临床结果以及作为计算方法函数的长期治疗相关毒性。
材料/方法：分析了2009年至2011年期间治疗的70例患者计划（19例中央型和51例周围型肺部病变）的回顾性系列。其中，33例为原发性肺癌，37例为转移性病变。33个治疗计划采用RT方法制定，37个计划使用MC。用相互方法重新计算各组以进行剂量比较。为量化两种算法之间的剂量差异而检查的参数包括：计划靶体积（PTV）95%（D95）处的剂量、剂量不均匀性以及危及器官（OAR）的剂量。剂量差异作为靶体积、到软组织的距离以及靶与软组织重叠部分的函数进行分析。对于原发性肺肿瘤子集，在中位随访19个月（范围2至41个月）时通过影像学评估LRC。
与MC相比，RT算法在很大程度上高估了输送到PTV的剂量。RT和MC计划之间的剂量差异与PTV与软组织重叠的体积相关；重叠体积越小，RT和MC之间的剂量差异越大。与MC相比，RT算法高估了输送到同侧肺10%（D10%）的剂量。在31例接受原发性肺恶性肿瘤治疗的患者中仅1例出现局部进展迹象。RT和MC计划之间的无病生存期（DFS）和总生存期（OS）无显著差异。
射波刀立体定向体部放疗治疗的周围型肺肿瘤的MC和RT剂量计算之间存在显著差异范围。虽然差异与靶大小和与软组织的接近程度相关，但没有单一参数能够可靠地预测剂量差异。最终，局部控制和长期毒性似乎与剂量计算方法无关。
