【作者】 龚唯；

【导师】 周菊英；

【作者基本信息】 苏州大学 ， 肿瘤学， 2012， 硕士

【摘要】 目的通过每周KV-CBCT监测食管癌三维适形放疗过程中的摆位误差和剩余摆位误差，为制定我科食管癌患者合理的CTV-PTV外放边界提供依据。使用KV-CBCT图像更新治疗计划，评价离线自适应放疗的可行性与优势，并建立一个食管癌在线自适应放疗的模型。资料和方法第一部分我们选取19例食管癌患者。每例患者每周采集CBCT图像，在相应的三维匹配框内与定位CT进行灰度值自动配准，若任意方向的误差超过3mm阈值，则接受重新摆位。在线重新摆位后，再次行KV-CBCT扫描，并按照同样的匹配方式与计划CT进行配准，采集校正后三维方向上的摆位误差。离线分析数据并根据van Herk公式计算出CTV-PTV理论外放边界。第二部分预处理：将19例患者的计划CT图像连同勾画好的靶区分别复制一份，采用容积密度分配的方法，给定肺组织的CT值为：-800Hu，其余组织均定义为水的密度，即Body-Lung的CT值为0Hu，将定位CT的计划（plan1）移植到密度分配好的CT2图像中，生成plan2。根据靶区和正常组织的受量及DVH图比较plan1与plan2的剂量分布差异。主要评价指标有：靶区（GTV、CTV、PTV）平均剂量及95%体积剂量覆盖率（D95）；脊髓平均剂量及1%体积受量（D1）；肺平均剂量（MLD）及V20。自适应：在第一部分结论和预处理结论的基础上，在19例患者第1周的CBCT图像上勾画靶区，生成新的CTV-PTV外放，我们称之为自适应计划（adaptive plan）：planA，与本科室临床治疗计划（clinical plan）planC进行剂量学对比，评估参数有：PTV体积、95%体积PTV覆盖率（PTVD95）、99％体积GTV覆盖率(GTVD99)、脊髓D1cc、双肺V20、双肺V30。分析离线KV-CBCT图像的同时，计算瓦里安OBI系统实现在线更新治疗计划的所需时间。结果第一部分校正前，107次CBCT图像在左右、垂直、前后方向的摆位误差分别为0.39±0.31cm、0.24±0.23cm、0.28±0.22cm，共计有79次（74%）的摆位分次需要接受在线校位。根据van Herk公式计算不校正的CTV-PTV外放边界分别为：1.00cm、0.91cm、0.96cm。校正后，107次CBCT图像在左右、垂直、前后方向的摆位误差分别为:0.17±0.13cm,0.16±0.12cm，0.14±0.14cm。根据van Herk公式计算校正后的CTV-PTV理论外放边界（MPTV）分别为：0.50cm、0.46cm、0.49cm。第二部分容积密度分配后CT计划与常规CT计划的一致性非常好，plan1与plan2的靶区剂量差异在3%以内，正常组织剂量差异均小于2%。自适应计划并没有影响靶区的覆盖率，95%PTV覆盖率、99%GTV覆盖率与临床计划相比，差异无统计学意义。双肺V20和双肺V30较临床计划均有明显的缩小，分别是V20：18.26%±2.98%vs.12.14%±2.89%，V30：9.19%±3.45%vs.5.71%±2.63%，差异有显著统计学意义（P=0.0001，P=0.001）。在1cm3脊髓受量方面，自适应计划较临床计划减少了228cGy±127.75cGy，差异有统计学意义（P=0.04）。瓦里安OBI系统实现在线更新治疗计划从摆位到传输更新的计划，共需耗时约15分钟。结论在线KV-CBCT引导放疗可以降低摆位误差，缩小CTV-PTV外放边界。纠正摆位误差后，采用容积密度分配的方法，可以使用KV-CBCT图像更新治疗计划。第1周的离线自适应计划在不影响靶区剂量覆盖的前提下，能充分保护邻近危及器官。瓦里安的OBI系统在线更新计划约需15分钟，理论上达到了在线ART的标准，实际工作中，实现在线ART远远未达到理想状态。更多还原

【Abstract】 PurposeThe set-up error and the residual set-up error are monitored on esophageal patientswith three-dimensional conformal radiotherapy through the weekly KV-CBCT, forproviding the basis for reasonable esophageal CTV-PTV margin in our department. Thetreatment plan is updated through the KV-CBCT image to evaluate the feasibility andadvantage of off-line adaptive radiotherapy, and establish an esophageal on-line adaptiveradiotherapy model.Materials and methodsSection oneWe select19cases of esophageal cancer patients. Each patient takes CBCT imageevery week, the images in the corresponding3D match box are got to be automatic greyregistration with the positioning CT images, if error of any direction more than3mmthreshold, then accept a online placement reset. After that, we do again KV-CBCTscanning, and follow the same way for images registration, the set-up error we collected isthe error after correction in3D direction. Then analysis data offline and calculatetheoretical CTV-PTV margin according to van Herk formula.Section twoPretreatment: Copy the planning CT images together with target, use the method of bulkdensity assignment, give the Lung tissue CT value is:-800Hu, the rest are defined as theorganization of the density of water, that is, Body-Lung CT value:0Hu,Transplantate CTplan (plan1) to the density distribution CT2images, named plan2. According to the targetand normal tissue dose and DVH charts, we compare the dose distribution of differences ofplan1and plan2. Main evaluating indexes: the average dose of target (GTV, CTV, PTV) and95%volume dose coverage (D95); average dose of spinal cord and1%volumequantity (D1); Lung average dose (MLD) and V20.Adaptive: On the basis of the first part conclusion and preprocessing conclusion, wedelineate the target and the organs at risk in the first week of CBCT image of the19patients, creat a new CTV-PTV margin, called adaptive plan: planA, and compare with theclinical plan: planC in dose distribution. Evaluation parameters are: PTV volume,95%volume PTV coverage (PTVD95),99%volume GTV coverage (GTVD99), spinal cordD1cc, total lung V20, total lung V30. When analyzing the off-line KV-CBCT images, wecalculate the required time of online updating treatment plan with Varian OBI system.ResultsSection oneBefore correction, the set-up error of107CBCT images in Lat, Vrt, Lng direction is0.39±0.31cm,0.24±0.23cm,0.28±0.22cm respectively, altogether79times (74%) needto accept a online correction. According to van Herk formula, the no correction CTV-PTVmargin is:1.00cm,0.91cm,0.96cm respectively. After correction, the set-up error of107CBCT images is:0.17±0.13cm,0.16±0.12cm,0.14±0.14cm respectively. According tovan Herk formula, the no correction CTV-PTV margin is:0.50cm,0.46cm,0.49cmrespectively.Section twoThe plan with bulk density assignment and the conventional plan show very goodconsistency. The dose difference of target about plan1and plan2is within3%, of normaltissue is less than2%. Compare with clinical treatment plan, adaptive plan have noeffection on target coverage, the difference in95%PTV coverage,99%GTV coverage wasnot statistically significant. Total lung V20and total lung V30are obviously shrinks,respectively V20:18.26%±2.98%vs.12.14%±2.89%, V30:9.19%±3.45%vs.5.71%±2.63%, there is a significant statistical difference(P=0.0001, P=0.001). In1cc dose ofspinal cord, adaptive plan is cutted by228cGy±127.75cGy, the difference is statisticallysignificant (P=0.04). A model of online updating treatment plan is builded up, frompositioning to transporting the updated plan, Varian OBI system totally need to take about 15minutes.ConclusionsThe set-up error can be reduced and the CTV-PTV margin can be narrowed throughon-line KV-CBCT guided radiotherapy. After the correction of set-up error, we can use thebulk density assignment method, making the KV-CBCT images to participate in replanning.In the premise of not affect the target coverage, the first week of offline adaptive plan canavailably protect the organs at risk. Varian OBI system totally need to take about15minutesfor online updating the plan, theoretically up to the online ART standard, however,implementing online ART is far away to be reached.更多还原