Department of Medical Physics, Greater Poland Cancer Centre, Garbary 15, 61-866, Poznan, Poland.
Department of Technical Physics, Poznan University of Technology, Poznan, Poland.
Sci Rep. 2022 Apr 8;12(1):5974. doi: 10.1038/s41598-022-10072-8.
Metal artefacts degrade clinical image quality which decreases the confidence of using computed tomography (CT) for the delineation of key structures for treatment planning and leads to dose errors in affected areas. In this work, we investigated accuracy of doses computed by the Eclipse treatment planning system near and inside metallic elements for two different computation algorithms. An impact of CT metal artefact reduction methods on the resulting calculated doses has also been assessed. A water phantom including Gafchromic film and metal inserts was irradiated (max dose 5 Gy) using a 6 MV photon beam. Three materials were tested: titanium, alloy 600, and tungsten. The phantom CT images were obtained with the pseudo-monoenergetic reconstruction (PMR) and the iterative metal artefact reduction (iMAR). Image sets were used for dose calculation using an Eclipse treatment planning station (TPS). Monte Carlo (MC) simulations were used to predict the true dose distribution in the phantom allowing for comparison with doses measured by film and calculated by TPS. Measured and simulated percentage depth doses (PDDs) were not statistically different (p > 0.618). Regional differences were observed at edges of metallic objects (max 8% difference). However, PDDs simulated with and without film were statistically different (p < 0.002). PDDs calculated by the Acuros XB algorithm based on the dose-to-medium approach best matched the MC reference regardless of the CT reconstruction methods and inserts used (p > 0.078). PDDs obtained using other algorithms significantly differ from the MC values (p < 0.011). The Acuros XB algorithm with a dose-to-medium approach provides reliable dose calculation in all metal regions when using the Varian system. The inability of the AAA algorithm to model backscatter dose significantly limits its clinical application in the presence of metal. No significant impact on the dose calculation was found for a range of metal artefact reduction strategies.
金属伪影会降低临床图像质量，从而降低使用计算机断层扫描（CT）进行治疗计划中关键结构勾画的信心，并导致受影响区域的剂量误差。在这项工作中，我们研究了两种不同计算算法在金属元素附近和内部计算的 Eclipse 治疗计划系统剂量的准确性。还评估了 CT 金属伪影减少方法对计算出的剂量的影响。使用 6 MV 光子束照射包括 Gafchromic 胶片和金属插件的水模体（最大剂量 5Gy）。测试了三种材料：钛、600 合金和钨。使用伪单能重建（PMR）和迭代金属伪影减少（iMAR）获得幻影 CT 图像。使用 Eclipse 治疗计划站（TPS）计算图像集的剂量。使用蒙特卡罗（MC）模拟来预测真实的剂量分布在允许与胶片测量和 TPS 计算的剂量进行比较的幻影中。测量和模拟的百分深度剂量（PDD）没有统计学差异（p>0.618）。在金属物体边缘观察到区域差异（最大 8%差异）。然而，在没有胶片的情况下模拟和没有胶片的 PDDs 有统计学差异（p<0.002）。基于剂量到介质方法的 Acuros XB 算法计算的 PDDs 无论使用 CT 重建方法和插入物如何，都与 MC 参考值最匹配（p>0.078）。使用其他算法获得的 PDDs 与 MC 值显著不同（p<0.011）。当使用 Varian 系统时，基于剂量到介质方法的 Acuros XB 算法在所有金属区域都能提供可靠的剂量计算。AAA 算法无法对背散射剂量进行建模，这极大地限制了其在存在金属的情况下的临床应用。没有发现金属伪影减少策略的范围对剂量计算有重大影响。
