Wallace D, Ng J A, Keall P J, O'Brien R T, Poulsen P R, Juneja P, Booth J T
School of Physics, University of Sydney, NSW 2006, Australia.
Phys Med Biol. 2015 Jun 21;60(12):4835-47. doi: 10.1088/0031-9155/60/12/4835. Epub 2015 Jun 9.
Kilovoltage intrafraction monitoring (KIM) utilises the kV imager during treatment for real-time tracking of prostate fiducial markers. However, its effectiveness relies on sufficient image quality for the fiducial tracking task. To guide the performance characterisation of KIM under different clinically relevant conditions, the effect of different kV parameters and patient size on image quality, and quantification of MV scatter from the patient to the kV detector panel were investigated in this study. Image quality was determined for a range of kV acquisition frame rates, kV exposure, MV dose rates and patient sizes. Two methods were used to determine image quality; the ratio of kV signal through the patient to the MV scatter from the patient incident on the kilovoltage detector, and the signal-to-noise ratio (SNR). The effect of patient size and frame rate on MV scatter was evaluated in a homogeneous CIRS pelvis phantom and marker segmentation was determined utilising the Rando phantom with embedded markers. MV scatter incident on the detector was shown to be dependent on patient thickness and frame rate. The segmentation code was shown to be successful for all frame rates above 3 Hz for the Rando phantom corresponding to a kV to MV ratio of 0.16 and an SNR of 1.67. For a maximum patient dimension less than 36.4 cm the conservative kV parameters of 5 Hz at 1 mAs can be used to reduce dose while retaining image quality, where the current baseline kV parameters of 10 Hz at 1 mAs is shown to be adequate for marker segmentation up to a patient dimension of 40 cm. In conclusion, the MV scatter component of image quality noise for KIM has been quantified. For most prostate patients, use of KIM with 10 Hz imaging at 1 mAs is adequate however image quality can be maintained and imaging dose reduced by altering existing acquisition parameters.
千伏级分次内监测（KIM）在治疗过程中利用千伏级成像仪对前列腺基准标记物进行实时跟踪。然而，其有效性依赖于用于基准跟踪任务的足够图像质量。为了指导KIM在不同临床相关条件下的性能表征，本研究调查了不同千伏参数和患者体型对图像质量的影响，以及从患者到千伏探测器面板的兆伏级散射的量化。针对一系列千伏采集帧率、千伏曝光、兆伏剂量率和患者体型确定了图像质量。使用两种方法确定图像质量；通过患者的千伏信号与入射到千伏探测器上的来自患者的兆伏散射的比率，以及信噪比（SNR）。在均匀的CIRS骨盆模型中评估了患者体型和帧率对兆伏散射的影响，并利用带有嵌入式标记物的兰多模型确定了标记物分割。结果表明，入射到探测器上的兆伏散射取决于患者厚度和帧率。对于对应于千伏与兆伏比率为0.16且信噪比为1.67的兰多模型，分割代码在所有高于3 Hz的帧率下均成功。对于最大患者尺寸小于36.4 cm的情况，可使用1 mAs下5 Hz的保守千伏参数来降低剂量同时保持图像质量，而当前1 mAs下10 Hz的基线千伏参数在患者尺寸达40 cm时足以进行标记物分割。总之，已对KIM图像质量噪声的兆伏散射分量进行了量化。对于大多数前列腺患者，使用1 mAs下10 Hz成像时KIM是足够的，但是通过改变现有采集参数可以保持图像质量并降低成像剂量。
