The University of Manchester, Manchester Academic Health Science Centre, The Christie NHS Foundation Trust, Manchester, M20 4BX, United Kingdom. Christie Medical Physics and Engineering, The Christie NHS Foundation Trust, Manchester, M20 4BX, United Kingdom.
Phys Med Biol. 2018 Mar 12;63(6):065003. doi: 10.1088/1361-6560/aab0f0.
Radiotherapy dose calculations based on cone-beam CT (CBCT) images can be inaccurate due to unreliable Hounsfield units (HU) in the CBCT. Deformable image registration of planning CT images to CBCT, and direct correction of CBCT image values are two methods proposed to allow heterogeneity corrected dose calculations based on CBCT. In this paper we compare the accuracy and robustness of these two approaches. CBCT images for 44 patients were used including pelvis, lung and head & neck sites. CBCT HU were corrected using a 'shading correction' algorithm and via deformable registration of planning CT to CBCT using either Elastix or Niftyreg. Radiotherapy dose distributions were re-calculated with heterogeneity correction based on the corrected CBCT and several relevant dose metrics for target and OAR volumes were calculated. Accuracy of CBCT based dose metrics was determined using an 'override ratio' method where the ratio of the dose metric to that calculated on a bulk-density assigned version of the same image is assumed to be constant for each patient, allowing comparison to the patient's planning CT as a gold standard. Similar performance is achieved by shading corrected CBCT and both deformable registration algorithms, with mean and standard deviation of dose metric error less than 1% for all sites studied. For lung images, use of deformed CT leads to slightly larger standard deviation of dose metric error than shading corrected CBCT with more dose metric errors greater than 2% observed (7% versus 1%).
基于锥形束 CT(CBCT)图像的放射治疗剂量计算可能不准确,因为 CBCT 中的 Hounsfield 单位(HU)不可靠。将计划 CT 图像与 CBCT 进行变形图像配准,并直接校正 CBCT 图像值是两种允许基于 CBCT 进行不均匀性校正剂量计算的方法。本文比较了这两种方法的准确性和稳健性。使用包括骨盆、肺和头颈部部位的 44 名患者的 CBCT 图像。使用“阴影校正”算法校正 CBCT HU,并使用 Elastix 或 Niftyreg 通过对计划 CT 到 CBCT 的变形配准进行校正。基于校正后的 CBCT 重新计算具有不均匀性校正的放射治疗剂量分布,并计算目标和 OAR 体积的几个相关剂量指标。使用“覆盖比”方法确定基于 CBCT 的剂量指标的准确性,其中剂量指标与相同图像的体密度分配版本计算的剂量指标的比值对于每个患者假定为常数,允许与患者的计划 CT 进行比较作为金标准。阴影校正的 CBCT 和两种变形配准算法都具有相似的性能,所有研究部位的剂量指标误差的平均值和标准差均小于 1%。对于肺部图像,使用变形 CT 导致剂量指标误差的标准差略大于阴影校正的 CBCT,观察到更多的剂量指标误差大于 2%(7%对 1%)。