Rosewall Tara, Bayley Andrew, Catton Charles, Chung Peter, Currie Geoffrey, Heaton Robert, Wheat Janelle, Milosevic Michael
1 Radiation Medicine Program, Princess Margaret Cancer Centre, Toronto, ON, Canada.
2 Department of Radiation Oncology, University of Toronto, Toronto, ON, Canada.
Br J Radiol. 2015 Sep;88(1053):20140818. doi: 10.1259/bjr.20140818. Epub 2015 Jul 24.
To evaluate the methods to delineate the inner bladder (IB) surface using a uniform contraction from the outer bladder (OB) surface, assuming the bladder wall (BW) is either of constant thickness, constant volume or variable volume.
14 prostate intensity-modulated radiotherapy patients with 2 planning CTs were identified. For both CTs, OB was delineated using model-based segmentation. IB was delineated manually. Then, using uniform contractions from OB, the position of IB was approximated using a: 2.5-mm contraction, patient-specific contraction, patient-specific constant wall volume method and variable wall volume method. The structures created using those strategies were compared against the manual IB contours using geometric and dosimetric indices.
In the presence of variable bladder filling, use of a generic or patient-specific constant contraction resulted in a significant overestimation of IB volume (+12 and +13 cm(3), respectively; p < 0.001) that was inversely correlated with the difference in urine volume between the scans (R(2) > 0.86). Mean differences across 95% of IB surfaces were ≤2 mm for methods using either constant or variable wall volume. Mean dose-volume histogram (DVH) differences were <1 cm(3) across the whole BW DVH when using the method that assumed a variable wall volume.
The variable volume BW model provided the best approximation of the IB surface position under varying filling conditions, based on geometric and dosimetric indices.
Use of the equation derived in this research provides a quick and accurate method to delineate the hollow BW on serial imaging for the purposes of dose reconstruction.
假设膀胱壁(BW)厚度恒定、体积恒定或体积可变,评估从膀胱外表面(OB)进行均匀收缩来描绘膀胱内表面(IB)的方法。
确定了14例接受前列腺调强放疗且有2次计划CT扫描的患者。对于这两次CT扫描,均使用基于模型的分割方法描绘OB。手动描绘IB。然后,从OB进行均匀收缩,使用以下方法近似IB的位置:2.5毫米收缩、个体化收缩、个体化恒定壁体积法和可变壁体积法。使用几何和剂量学指标,将使用这些策略创建的结构与手动描绘的IB轮廓进行比较。
在膀胱充盈情况不同时,使用通用或个体化恒定收缩会导致IB体积显著高估(分别高估+12和+13立方厘米;p<0.001),且与两次扫描间尿量差异呈负相关(R²>0.86)。对于使用恒定或可变壁体积的方法,95%的IB表面平均差异≤2毫米。使用假设壁体积可变的方法时,整个BW剂量体积直方图(DVH)的平均差异<1立方厘米。
基于几何和剂量学指标,可变体积BW模型在不同充盈条件下能最好地近似IB表面位置。
使用本研究中推导的公式可提供一种快速准确的方法,用于在系列成像中描绘中空的BW以进行剂量重建。