^18F-FDG PET／CT最大标准摄取值联合HRCT在肺癌诊断中的价值和影响因素分析  

The maximum standardized uptake value of SF-FDG PET/CT combined with the image features on high resolution CT for the diagnosis of lung cancer

作　　者：王少雁[1] 张建[1] 孙高峰[1] 程超[1] 刘庆华[1] 彭婕[1] 韩一平[2] 郑建明[3] 孔令山[1] 左长京[1] 

机构地区：[1]第二军医大学长海医院核医学科,上海200433 [2]第二军医大学长海医院呼吸科,上海200433 [3]第二军医大学长海医院病理科,上海200433

出　　处：《中华核医学与分子影像杂志》2013年第1期29-33,共5页Chinese Journal of Nuclear Medicine and Molecular Imaging

基　　金：上海市卫生局科研课题计划（XBR2011040）;上海人才发展资金（2010020）;2011年度上海市博士后科研资助计划（11R21410600）;上海市科委科研基金（08411967800）

摘　　要：目的探讨^18F—FDGPET／CT联合高分辨率CT（HRCT）诊断肺癌的价值和主要影响因素。方法回顾分析2010年8月至2011年8月因肺实性病灶行^18F—FDGPET／CT检查及肺部HRCT扫描的122例患者资料，所有病例均经病理证实或影像学检查随访8个月以上确诊。利用∥检验对HRCT上不同影像特征在良恶性病变中的构成比差异进行比较，单因素方差分析比较不同病理类型病灶的SUV差异，多因素logistic回归分析SUV及HRCT影像特征等影响因素，探讨最佳SUV诊断界值和^18F—FDG PET／CT对肺癌的诊断价值。结果122例肺实性病灶患者中，恶性82例，良性40例。肺癌HRCT影像特征中前3位依次为毛刺征64．6％（53／82）、分叶征63．4％（52／82）和胸膜牵拉征39．0％（32／82），高于在良性病变中的比例（X2=19．08、30．89、10．88，均P〈0．01）。肺部鳞状细胞癌（简称鳞癌）、小细胞癌和腺癌的SUV依次为12．57±4．34、10．66±2．90和8．19±6．01，与肺部良性病变SUV（3．01±3．62）相比差异有统计学意义（F=20．83，P〈0．01）。不同病理类型肺癌suV从大到小依次为鳞癌、小细胞癌和腺癌，其中鳞癌与腺癌suV一差异有统计学意义（P〈0．01）；SUV…的ROCAUC为0．863，SUV界值2．99和2．50对肺癌诊断的灵敏度、特异性分别为89．0％（73／82）、75．0％（30／40）和91．5％（75／82）、65．0％（26／40）；诊断一致性SUV2．99优于SUV2．50，Kappa值分别为0．644和0．597。多因素logistic回归分析显示SUV2．99的诊断比值比（OR）优于SUV2．50的OR，分别为5．42和3．93；SUV（OR=5．42，P=0．01）、肿瘤最大径（OR=7．27．P=0．02）、毛刺征（鲫=7．70，P〈0．01））和分叶征（DR=12．38，P〈0．01）均为肺癌与良性病灶鉴别诊断有统计学意义的影响因素。结论SUV对肺癌的诊断和鉴别诊断有较高价值。“F-FDGPET／CT联合HRCT诊断肺实性病�Objective To investigate the diagnostic value of the SUV of lS F-FDG PET/CT combined with the image features on high resolution CT（HRCT） for lung cancer. Methods lS F-FDG PET/CT and HRCT data of 122 patients with pulmonary tumor-like lesions from August 2010 to August 2011 were an alyzed retrospectively. Diagnoses of the subjects were confirmed by pathology or follow-up （ 〉 8 months）. X2 test was performed to compare the differences in the constituent ratio of different HRCT features of the lesions. One-way analysis of variance was used to evaluate the differences of SUVm in different pathological types of lung cancer. Multivariate logistic stepwise regression analysis was performed to evaluate the signifcance of SUVmax and HRCT features for diagnosis of lung cancer. Results Out of the 122 patients, there were 40 with benign lesions and 82 with malignant lesions. Spiculation sign （64.6%, 53/82）, lobulation sign （63.4%, 52/82） and pleural indentation sign （39.0%, 32/82） were the top three HRCT features of the pulmonary malignant tumors,higher than those of benign lesions （X2 = 19.08, 30.89, 10.88, all P 〈 0. 01 ）. There was a statistical difference of SUVmI between benign （3.01±3.62） and respective type of malignant lesions （ F = 20.83, P 〈 0.01 ）. SUVm of squamous cell carcinoma, small cell carcinoma and adenocarcinoma was 12.57 ± 4.34, 10.66 ± 2.90 and 8.19 ±6.01 respectively. There was a statistical difference of SUVmax between adenocarcinoma and squamous cell carcinoma （P 〈 0.01 ）. The AUC on ROC curves of SUVm was 0. 863. The sensitivity, specificity and Kappa value were 89.0% （ 73/82）, 75.0% （30/40）, 0.644 with a SUVm cutoff of 2.99 and 91.5% （75/82）, 65.0% （26/40）, 0. 597 with a SUVm cutoff of 2. 50 for malignant lesion, respectively. Multivariate logistic stepwise regression analysis showed that OR for SUVm= 2.99 was higher than OR for SUVm 2.5 （5.42 vs 3.93 ）. SUVmax （ OR = 5.42, P = 0. 01 ）, tumor maximal diameter （ OR = 7.27, P =

关 键 词：肺肿瘤 发射型计算机 体层摄影术 x线计算机 脱氧葡萄糖 

分 类 号：R730.44[医药卫生—肿瘤]