Ansbacher W
BC Cancer Agency, Vancouver Island Centre, 2410 Lee Avenue, Victoria, B.C., Canada V8R 6V5.
Med Phys. 2006 Sep;33(9):3369-82. doi: 10.1118/1.2241997.
A new method for rapid evaluation of intensity modulated radiation therapy (IMRT) plans has been developed, using portal images for reconstruction of the dose delivered to a virtual three-dimensional (3D) phantom. This technique can replace an array of less complete but more time-consuming measurements. A reference dose calculation is first created by transferring an IMRT plan to a cylindrical phantom, retaining the treatment gantry angles. The isocenter of the fields is placed on or near the phantom axis. This geometry preserves the relative locations of high and low dose regions and has the required symmetry for the dose reconstruction. An electronic portal image (EPI) is acquired for each field, representing the dose in the midplane of a virtual phantom. The image is convolved with a kernel to correct for the lack of scatter, replicating the effect of the cylindrical phantom surrounding the dose plane. This avoids the need to calculate fluence. Images are calibrated to a reference field that delivers a known dose to the isocenter of this phantom. The 3D dose matrix is reconstructed by attenuation and divergence corrections and summed to create a dose matrix (PI-dose) on the same grid spacing as the reference calculation. Comparison of the two distributions is performed with a gradient-weighted 3D dose difference based on dose and position tolerances. Because of its inherent simplicity, the technique is optimally suited for detecting clinically significant variances from a planned dose distribution, rather than for use in the validation of IMRT algorithms. An analysis of differences between PI-dose and calculation, delta PI, compared to differences between conventional quality assurance (QA) and calculation, delta CQ, was performed retrospectively for 20 clinical IMRT cases. PI-dose differences at the isocenter were in good agreement with ionization chamber differences (mean delta PI = -0.8%, standard deviation sigma = 1.5%, against delta CQ = 0.3%, sigma = 1.0%, respectively). PI-dose plane differences had significantly less variance than film plane differences (sigma = 1.1 and 2.1%, respectively). Twenty-two further cases were evaluated using 3D EPI-dosimetry alone. The mean difference delta over volumes with doses above 80% of the isocenter value was delta = -0.3%, sigma(delta) = 0.7%, and standard deviations of the distributions ranged from 1.0 to 2.0%. Verification time per plan, from initial calculation, delivery, dose reconstruction to evaluation, takes less than 1.5 h and is more than four times faster than conventional QA.
一种用于快速评估调强放射治疗（IMRT）计划的新方法已经开发出来，该方法利用射野图像重建传递到虚拟三维（3D）体模的剂量。这项技术可以取代一系列不太完整但更耗时的测量方法。首先通过将IMRT计划传输到圆柱形体模来创建参考剂量计算，同时保留治疗机架角度。射野的等中心放置在体模轴上或其附近。这种几何结构保留了高剂量区和低剂量区的相对位置，并且具有剂量重建所需的对称性。为每个射野获取一张电子射野图像（EPI），该图像代表虚拟体模中平面的剂量。该图像与一个核进行卷积以校正散射不足，从而复制剂量平面周围圆柱形体模的效果。这避免了计算注量的需要。将图像校准到一个给该体模等中心提供已知剂量的参考射野。通过衰减和发散校正重建3D剂量矩阵，并求和以在与参考计算相同的网格间距上创建一个剂量矩阵（PI剂量）。基于剂量和位置容差，使用梯度加权3D剂量差异对两种分布进行比较。由于其固有的简单性，该技术最适合检测与计划剂量分布的临床显著差异，而不是用于IMRT算法的验证。对20例临床IMRT病例进行了回顾性分析，比较了PI剂量与计算之间的差异（δPI）和传统质量保证（QA）与计算之间的差异（δCQ）。等中心处的PI剂量差异与电离室差异吻合良好（平均δPI = -0.8%，标准差σ = 1.5%，而δCQ分别为0.3%，σ = 1.0%）。PI剂量平面差异的方差明显小于胶片平面差异（分别为σ = 1.1%和2.1%）。另外22例病例仅使用3D EPI剂量测定法进行评估。在剂量高于等中心值80%的体积上，平均差异δ为δ = -0.3%，σ(δ) = 0.7%，分布的标准差范围为1.0%至2.0%。从初始计算、照射、剂量重建到评估，每个计划的验证时间不到1.5小时，比传统QA快四倍多。
