Chaikh Abdulhamid, Balosso Jacques
Department of Radiation Oncology and Medical physics, University Hospital of Grenoble Alpes (CHU-GA), France.
France HADRON national research infrastructure, IPNL, Lyon, France.
Quant Imaging Med Surg. 2017 Jun;7(3):292-298. doi: 10.21037/qims.2017.06.03.
During the past decades, in radiotherapy, the dose distributions were calculated using density correction methods with pencil beam as type 'a' algorithm. The objectives of this study are to assess and evaluate the impact of dose distribution shift on the predicted secondary cancer risk (SCR), using modern advanced dose calculation algorithms, point kernel, as type 'b', which consider change in lateral electrons transport.
Clinical examples of pediatric cranio-spinal irradiation patients were evaluated. For each case, two radiotherapy treatment plans with were generated using the same prescribed dose to the target resulting in different number of monitor units (MUs) per field. The dose distributions were calculated, respectively, using both algorithms types. A gamma index (γ) analysis was used to compare dose distribution in the lung. The organ equivalent dose (OED) has been calculated with three different models, the linear, the linear-exponential and the plateau dose response curves. The excess absolute risk ratio (EAR) was also evaluated as (EAR = OED / OED ).
The γ analysis results indicated an acceptable dose distribution agreement of 95% with 3%/3 mm. Although, the γ-maps displayed dose displacement >1 mm around the healthy lungs. Compared to type 'a', the OED values from type 'b' dose distributions' were about 8% to 16% higher, leading to an EAR ratio >1, ranged from 1.08 to 1.13 depending on SCR models.
The shift of dose calculation in radiotherapy, according to the algorithm, can significantly influence the SCR prediction and the plan optimization, since OEDs are calculated from DVH for a specific treatment. The agreement between dose distribution and SCR prediction depends on dose response models and epidemiological data. In addition, the γ passing rates of 3%/3 mm does not translate the difference, up to 15%, in the predictions of SCR resulting from alternative algorithms. Considering that modern algorithms are more accurate, showing more precisely the dose distributions, but that the prediction of absolute SCR is still very imprecise, only the EAR ratio could be used to rank radiotherapy plans.
在过去几十年的放射治疗中，剂量分布是使用以笔形束为“a”型算法的密度校正方法来计算的。本研究的目的是评估和评价剂量分布偏移对预测的二次癌症风险（SCR）的影响，采用现代先进的剂量计算算法——点核算法，即“b”型算法，该算法考虑了侧向电子传输的变化。
对儿科颅脊髓照射患者的临床实例进行评估。对于每个病例，使用相同的靶区处方剂量生成两个放射治疗计划，每个射野的监测单位（MU）数量不同。分别使用两种算法计算剂量分布。采用伽马指数（γ）分析来比较肺部的剂量分布。使用线性、线性 - 指数和平台剂量响应曲线这三种不同模型计算器官等效剂量（OED）。还评估了超额绝对风险比（EAR），计算公式为（EAR = OED / OED ）。
γ分析结果表明，在3%/3 mm条件下，剂量分布一致性可接受，为95%。不过，γ图显示健康肺部周围剂量位移>1 mm。与“a”型算法相比，“b”型算法剂量分布的OED值高出约8%至16%，导致EAR比值>1，根据SCR模型，EAR比值范围为1.08至1.13。
在放射治疗中，根据算法的剂量计算偏移会显著影响SCR预测和计划优化，因为OED是从特定治疗的剂量体积直方图（DVH）计算得出的。剂量分布与SCR预测之间的一致性取决于剂量响应模型和流行病学数据。此外，3%/3 mm的γ通过率并未体现替代算法导致的SCR预测中高达15%的差异。考虑到现代算法更准确，能更精确地显示剂量分布，但绝对SCR的预测仍然非常不精确，因此只能使用EAR比值对放射治疗计划进行排序。
