Demol Benjamin, Viard Romain, Reynaert Nick
Centre Oscar Lambret; AQUILAB SAS; UMR CNRS 8520.
J Appl Clin Med Phys. 2015 Sep 8;16(5):117–130. doi: 10.1120/jacmp.v16i5.5586.
The purpose of this study was to demonstrate that Monte Carlo treatment planning systems require tissue characterization (density and composition) as a function of CT number. A discrete set of tissue classes with a specific composition is introduced. In the current work we demonstrate that, for megavoltage photon radiotherapy, only the hydrogen content of the different tissues is of interest. This conclusion might have an impact on MRI-based dose calculations and on MVCT calibration using tissue substitutes. A stoichiometric calibration was performed, grouping tissues with similar atomic composition into 15 dosimetrically equivalent subsets. To demonstrate the importance of hydrogen, a new scheme was derived, with correct hydrogen content, complemented by oxygen (all elements differing from hydrogen are replaced by oxygen). Mass attenuation coefficients and mass stopping powers for this scheme were calculated and compared to the original scheme. Twenty-five CyberKnife treatment plans were recalculated by an in-house developed Monte Carlo system using tissue density and hydrogen content derived from the CT images. The results were compared to Monte Carlo simulations using the original stoichiometric calibration. Between 300 keV and 3 MeV, the relative difference of mass attenuation coefficients is under 1% within all subsets. Between 10 keV and 20 MeV, the relative difference of mass stopping powers goes up to 5% in hard bone and remains below 2% for all other tissue subsets. Dose-volume histograms (DVHs) of the treatment plans present no visual difference between the two schemes. Relative differences of dose indexes D98, D95, D50, D05, D02, and Dmean were analyzed and a distribution centered around zero and of standard deviation below 2% (3 σ) was established. On the other hand, once the hydrogen content is slightly modified, important dose differences are obtained. Monte Carlo dose planning in the field of megavoltage photon radiotherapy is fully achievable using only hydrogen content of tissues, a conclusion that might impact MRI dose calculation, but can also help selecting the optimal tissue substitutes when calibrating MVCT devices.
本研究的目的是证明蒙特卡罗治疗计划系统需要将组织特征（密度和成分）作为CT值的函数。引入了一组具有特定成分的离散组织类别。在当前工作中，我们证明，对于兆伏级光子放疗，仅不同组织的氢含量是重要的。这一结论可能会对基于MRI的剂量计算以及使用组织等效物的MVCT校准产生影响。进行了化学计量校准，将具有相似原子组成的组织分组为15个剂量学等效子集。为了证明氢的重要性，推导了一种新方案，该方案具有正确的氢含量，并以氧作为补充（所有不同于氢的元素都被氧取代）。计算了该方案的质量衰减系数和质量阻止本领，并与原始方案进行了比较。使用从CT图像得出的组织密度和氢含量，通过内部开发的蒙特卡罗系统重新计算了25个射波刀治疗计划。将结果与使用原始化学计量校准的蒙特卡罗模拟进行了比较。在300 keV至3 MeV之间，所有子集中质量衰减系数的相对差异在1%以内。在10 keV至20 MeV之间，硬骨中质量阻止本领的相对差异高达5%，而所有其他组织子集的相对差异均保持在2%以下。治疗计划的剂量体积直方图（DVH）在两种方案之间没有视觉差异。分析了剂量指标D98、D95、D50、D05、D02和Dmean的相对差异，并建立了以零为中心、标准差低于2%（3σ）的分布。另一方面，一旦氢含量略有改变，就会得到显著的剂量差异。在兆伏级光子放疗领域，仅使用组织的氢含量就完全可以实现蒙特卡罗剂量计划，这一结论可能会影响MRI剂量计算，但也有助于在校准MVCT设备时选择最佳的组织等效物。
