中文摘要：

目的：研究肿瘤放疗中由体表标记和激光灯摆位的床值偏差（体外误差）和由每周一次验证影像匹配所得靶区位置偏移（体内误差）之间的关联性，为评估每周一次影像引导放射治疗（IGRT）方案的有效性提供参考。方法：鼻咽癌、肺癌和宫颈癌患者共计28 例。技术员摆位后，记录体外误差；每周一次kV或锥形束CT（CBCT）扫描，并与数字重建影像或定位CT图像配准，记录体内误差。分析两类误差的差异及相关性。结果：3 类肿瘤的体外误差和体内误差的系统误差之间相近（P〉0.05）。随机误差：鼻咽癌无显著差异（P〉0.05）；宫颈癌有显著差异（P〈0.05）；肺癌头-脚（SI）方向无显著差异（P〉0.05），腹-背（AP）和左-右（LR）方向有显著差异（P〈0.05）。由体内误差得临床靶区（CTV）-计划靶区（PTV）边界（AP/SI/LR）：鼻咽癌2.80 mm/1.73 mm/2.73 mm；肺癌9.80 mm/5.98 mm/5.09 mm；宫颈癌3.94 mm/5.20 mm/2.73 mm。结论：鼻咽癌体外误差与体内误差之间有较好一致性，前者可以作为衡量后者的一个指标。肺癌和宫颈癌体外误差与体内误差之间有显著差异，为减小体内误差带来的放疗精度影响，需要考虑高频次的影像引导，或采用先进定位技术（如CRadCatalyst）进行患者摆位，以完成精确放疗。

英文摘要：

Objective To provide a reference for evaluating the effectiveness of image-guided radiotherapy （IGRT） once a week by studying on the correlation between couch errors and target displacements in radiotherapy. Couch errors were caused by body surface marking and laser beam positioning; target displacements were obtained from the verification of image matching once a week. Methods Totally, 28 patients respectively with nasopharyngeal cancer, lung cancer and cervical cancer were selected. Couch errors were recorded after the position. Cone beam computed tomography or kV scan were carried out once a week, and the obtained image was registered with the digitally reconstructed radiographs or located CT images to record the target displacement. The differences and correlations of couch errors and target displacements were analyzed. Results No significant differences were found in the systematic error of couch errors and target displacements among three kinds of cancers （P〉0.05）. Significant differences were not found in the random error of nasopharyngeal cancer （P〉0.05）, but found in that of cervical cancer （P〈0.05）. For lung cancer, significant differences were not found in superior-inferior （SI） direction （P〉0.05）, but found in anterior-posterior （AP） and left-right （LR） directions （P〈0.05）. The margins of clinical target volume and planning target volume obtained from target displacements were respectively 2.80 mm, 1.73 mm, 2.73 mm in AP, SI, LR directions for nasopharyngeal cancer, 9.80 mm, 5.98 mm, 5.09 mm for lung cancer, and 3.94 mm, 5.20 mm, 2.73 mm for cervical cancer. Conclusion The couch errors and target displacements has a good consistency, and the couch errors could be an indicator to measure target displacements. Significant differences are existed between the couch errors and target displacements of lung cancer and cervical cancer. The IGRT of higher frequencies need to be considered, or the advanced location technique such as C- Rad Catalyst should be adopted