肺癌患者GTVPET-CT勾画的系统分析

Systematic analysis of GTVpET.CT delineation in lung cancer patients

目的分析肺癌患者GTVPET-CT的勾画、阈值的选择,以及同机和非同机融合靶区的差异。方法在PUBMED和EMBASE中检索用PET-CT融合图像为肺癌患者勾画靶区的现有研究。回顾和分析靶区勾画方法,并对同机和非同机融合GTVPET-CT进行分析。结果文献检索中,共有10篇文献满足研究条件。10项研究总共有198例,其中非同机融合研究153例（77．3%）,同机融合45例（23．7%）。其中,4篇有详细GTV数据的文献共有79例,同机和非同机融合研究的GTVPET-CT分别为（88．0±93．0）cm^3和（59．1±77．9）cm^3。同机融合和非同机融合间GTV绝对差/绝对和差异有统计学意义（t=-6．55,P＜0．001）。即使把Brianzoni（同机融合）的绝对差/绝对和,与其一致性较好的Steenbakkers（非同机融合）单独比较时,差异仍有统计学意义（t=-3．27,P=0．003）。结论与非同机融合相比,同机融合PET-CT图像有更小的器官位移,更适用于靶区的勾画。放疗的靶区勾画方法应首选40%～50%范围的可变阈值,并且选取与肺窗CT靶区相应的阈值。

Objective to analyze the GTVPET-CT delineation, threshold selection and the difference between software and hardware GTV fusion among lung cancer patients. Methods In March 2006, after reviewing the literature, this research was conducted in PUBMED and EMBASE to identify the articles that delineated GTV on PET-CT images among lung caner patients. With methods of target volume delineation reviewed and the differences between software and hardware GTVPETCT fusion analyzed, we wish to present the result here. Results totally 10 articles were found in the literature search. These studies included 198 patients, including 153 patients using fusion software and 45 patients using hardware fusion. In 4 articles which had individual GTV data of 79 patients, GTVPET-CT under software and hardware fusion were （88.0 ±93.0） cm^3 and （59.1 ± 77.9） cm^3. The difference of absolute different value/absolute plus value between the two fusion levels were statistically significant （ t = - 6.55,P 〈 0. 001 ）. Conclusions Compared with software fusion, hardware PET-CT images fusion possesses less organ displacement, which is required in the delineation of target volume. A variable threshold ranging from 40% to 50% according to the CT value under the lung window is suggested.