Vedam S, Archambault L, Starkschall G, Mohan R, Beddar S
Department of Radiation Physics, University of Texas M.D. Anderson Cancer Center, Houston, Texas 77030, USA.
Med Phys. 2007 Nov;34(11):4247-55. doi: 10.1118/1.2794169.
Four-dimensional (4D) computed tomography (CT) imaging has found increasing importance in the localization of tumor and surrounding normal structures throughout the respiratory cycle. Based on such tumor motion information, it is possible to identify the appropriate phase interval for respiratory gated treatment planning and delivery. Such a gating phase interval is determined retrospectively based on tumor motion from internal tumor displacement. However, respiratory-gated treatment is delivered prospectively based on motion determined predominantly from an external monitor. Therefore, the simulation gate threshold determined from the retrospective phase interval selected for gating at 4D CT simulation may not correspond to the delivery gate threshold that is determined from the prospective external monitor displacement at treatment delivery. The purpose of the present work is to establish a relationship between the thresholds for respiratory gating determined at CT simulation and treatment delivery, respectively. One hundred fifty external respiratory motion traces, from 90 patients, with and without audio-visual biofeedback, are analyzed. Two respiratory phase intervals, 40%-60% and 30%-70%, are chosen for respiratory gating from the 4D CT-derived tumor motion trajectory. From residual tumor displacements within each such gating phase interval, a simulation gate threshold is defined based on (a) the average and (b) the maximum respiratory displacement within the phase interval. The duty cycle for prospective gated delivery is estimated from the proportion of external monitor displacement data points within both the selected phase interval and the simulation gate threshold. The delivery gate threshold is then determined iteratively to match the above determined duty cycle. The magnitude of the difference between such gate thresholds determined at simulation and treatment delivery is quantified in each case. Phantom motion tests yielded coincidence of simulation and delivery gate thresholds to within 0.3%. For patient data analysis, differences between simulation and delivery gate thresholds are reported as a fraction of the total respiratory motion range. For the smaller phase interval, the differences between simulation and delivery gate thresholds are 8 +/- 11% and 14 +/- 21% with and without audio-visual biofeedback, respectively, when the simulation gate threshold is determined based on the mean respiratory displacement within the 40%-60% gating phase interval. For the longer phase interval, corresponding differences are 4 +/- 7% and 8 +/- 15% with and without audiovisual biofeedback, respectively. Alternatively, when the simulation gate threshold is determined based on the maximum average respiratory displacement within the gating phase interval, greater differences between simulation and delivery gate thresholds are observed. A relationship between retrospective simulation gate threshold and prospective delivery gate threshold for respiratory gating is established and validated for regular and nonregular respiratory motion. Using this relationship, the delivery gate threshold can be reliably estimated at the time of 4D CT simulation, thereby improving the accuracy and efficiency of respiratory-gated radiation delivery.
四维（4D）计算机断层扫描（CT）成像在整个呼吸周期中对肿瘤及周围正常结构的定位方面愈发重要。基于此类肿瘤运动信息，能够确定呼吸门控治疗计划与实施的合适相位区间。这样的门控相位区间是根据肿瘤内部位移导致的肿瘤运动进行回顾性确定的。然而，呼吸门控治疗是基于主要从外部监测器获取的运动进行前瞻性实施的。因此，在4D CT模拟时为门控选择的回顾性相位区间所确定的模拟门控阈值，可能与治疗实施时根据前瞻性外部监测器位移确定的实施门控阈值不相符。本研究的目的是分别建立CT模拟和治疗实施时确定的呼吸门控阈值之间的关系。分析了来自90例患者的150条外部呼吸运动轨迹，其中部分患者有视听生物反馈，部分没有。从4D CT得出的肿瘤运动轨迹中选择两个呼吸相位区间，即40% - 60%和30% - 70%用于呼吸门控。对于每个这样的门控相位区间内的残余肿瘤位移，基于（a）该相位区间内的平均呼吸位移和（b）最大呼吸位移来定义模拟门控阈值。前瞻性门控实施的占空比是根据所选相位区间和模拟门控阈值内的外部监测器位移数据点的比例来估计的。然后迭代确定实施门控阈值以匹配上述确定的占空比。在每种情况下，对模拟和治疗实施时确定的此类门控阈值之间的差异大小进行量化。模体运动测试表明模拟和实施门控阈值的一致性在0.3%以内。对于患者数据分析，模拟和实施门控阈值之间的差异以总呼吸运动范围的分数形式报告。对于较小的相位区间，当基于40% - 60%门控相位区间内的平均呼吸位移确定模拟门控阈值时，有和没有视听生物反馈的情况下，模拟和实施门控阈值之间的差异分别为8±11%和14±21%。对于较长的相位区间，相应的差异分别为4±7%和8±15%，有和没有视听生物反馈时情况相同。或者，当基于门控相位区间内的最大平均呼吸位移确定模拟门控阈值时，观察到模拟和实施门控阈值之间的差异更大。建立并验证了用于常规和非常规呼吸运动的呼吸门控回顾性模拟门控阈值与前瞻性实施门控阈值之间的关系。利用这种关系，在4D CT模拟时可以可靠地估计实施门控阈值，从而提高呼吸门控放射治疗的准确性和效率。
