摘要
目的探讨PET/CT图像分割技术对肺癌放疗计划制定的影响。方法对12例无转移的肺癌患者行PET/CT扫描。采用自主编写的基于PCNN模型的自动分割程序对PET靶区进行分割处理,再分别以CT图像、PET/CT图像为依据采用目测法手动勾画肿瘤靶区,以相同参数制定调强放疗计划,对比分析靶区体积和剂量分布。结果 PET自动分割靶区与PET手动勾画靶区之间未见统计学差异(P>0.05),分割方法准确可靠;与CT手动勾画靶区之间差异有统计学意义(P<0.05),前者<后者,PET/CT图像较CT能更准确的区分肿瘤与肺不张。与基于CT的放疗计划相比,PET计划正常肺组织V20、V30均有显著降低,差异均有统计学意义(P<0.05)。脊髓、心脏和食管的受量差异无统计学意义(P>0.05)。结论 PET/CT图像分割技术提高了肿瘤靶区勾画的准确性,依据分割靶区制订的放疗计划能降低正常组织受照范围,减少并发症的发生率。
Objective To investigate the inlfuence on radiotherapy planning of lung cancer with the PET/CT segmentation technology.Methods 12 patients with non-metastatic lung cancer were scanned by PET/CT. The PET targets were segmented by automatic segmenting program written independently based on PCNN model.According to CT images and PET/CT images, the tumor targets were delineated manually by ocular estimating, radiotherapy planning was formulated with the same parameters, the targets volume and dose distribution were compared and analyzed.Results The differences between automatic segmentation target area and manual delineation target area of PET weren’t of statistical signiifcance （P〉0.05）, which proved the PET automatic segmentation method was more accurate and reliable. Statistically signiifcant differences （P〈0.05） existed between the PET automatic segmentation target area and CT manual delineation target area, and the former was less than the latter, which proved that PET/CT images were more accurate in distinguishing between the tumor and the atelectasis. Compared with the radiotherapy planning based on CT, PET planning signiifcantly reduced the V20 and V30 normal lung tissues, statistically signiifcant differences （P〈0.05） were discovered between them. However, there were no statistical significances （P〉0.05） in the quantity difference of spinal cords, hearts and esophaguses.Conclusion PET/CT image segmentation improves the accuracy of tumor target delineation. The radiotherapy planning based on target segmentation can reduce the illuminated scope of normal tissues and the incidence rate of complications.
出处
《中国医疗设备》
2014年第6期160-163,共4页
China Medical Devices
基金
国家自然科学基金项目(81170078)
广东省教育厅科技创新项目(2013KJCX152)
广州医科大学青年科研项目(2013A42)
关键词
PET
CT
图像分割
肺癌
调强放疗治疗
脉冲耦合神经网络
靶区勾画
PET/CT
image segmentation
lung cancer
radiotherapy planning
pulse coupled neural network
target volume delineation
