非小细胞肺癌原发灶代谢参数与区域淋巴结转移的关系研究

The relationship between metabolic parameters of primary focus of non-small cell lung cancers and regional lymph node metastasis

目的探讨非小细胞肺癌(NSCLC)18F-FDG PET/CT显像中原发病灶的各项代谢参数SUVmax、SUVmean、MTV、TLG和病灶T分期与区域淋巴结有无转移及N分期的关系及其临床价值。方法回顾性分析93例18 F-FDG PET/CT检查后行肺癌切除+区域淋巴结清扫术患者的资料,与病理对照,统计分析有无淋巴结转移、N分期与肺癌原发灶SUVmax、SUVmean、MTV、TLG、T分期之间是否有差异及相关性;通过ROC曲线方法评价SUVmax、SUVmean、MTV、TLG、T分期对肺癌患者区域淋巴结有无转移判断的价值。结果93例患者中N0 48例,N1 15例,N2 30例。Mann-Whitney U非参检验分析显示:肺癌原发灶SUVmax、SUVmean、MTV、TLG、T分期在区域淋巴结有无转移两组间均有统计学差异(P均<0.05)。Spearman相关分析显示:肺癌原发灶SUVmax、SUVmean、MTV、TLG、T分期与区域淋巴结转移均有相关性(P均<0.05)。非参数假设检验独立样本Kruskal-Wallis检验分析显示:除MTV外,肺癌原发灶SUVmax、SUVmean、TLG、T分期与N分期均有统计学差异(P均<0.05)。Spearman相关分析,肺癌原发灶SUVmax、SUVmean、TLG、T分期与N分期均有相关性(P均<0.05),MTV与N分期无明显相关性(P>0.05)。利用ROC曲线分析,SUVmax、SUVmean、MTV、TLG、T分期诊断是否有区域淋巴结转移的曲线下面积分别为0.699、0.695、0.640、0.728、0.626(P均<0.05)。TLG在这些参数中曲线下面积最大,具有较高诊断效能,以TLG>13.765(g)作为诊断界值,诊断的敏感性83.7%,特异性45.5%,阳性预测值65.1%,阴性预测值71.4%,准确性67.7%。结论 SUVmax、SUVmean、MTV、TLG和病灶T分期与非小细胞肺癌区域淋巴结转移及N分期有一定的相关性,TLG诊断区域淋巴结是否转移效能较高,可作为判断非小细胞肺癌区域淋巴结转移的辅助手段。

Objective To evaluate the relationship between metabolic parameters of the original non small cell lung canc- er （SUVmax, SUVmean, MTV, TLG and T stage ） and regional lymph node metastasis and N stage. Methods We ret rospectively analyzed the 18F-FDG PET/CT imagines of 93 patients with non-small cell lung cancer （NSCLC） before they had lung cancer resection and regional lymph node cleaning operation. Compared to pathology, cases were classified into groups according to whether regional lymph node metastasis and N stage. We analyzed the differences and correlation of SUVmax, SUVmean, MTV, TLG , T stage of lung cancer primary tumor between the groups. Meanwhile, we used ROC to analyze the value of SUVmax, SUVmean, MTV, TLG, T stage in judgment of regional lymph node metastasis in lung cancer. Results In the all 93 cases, 48 cases were NO, 15 cases were N1, and 30 cases were N2. The Mann-Whitney Utest analysis showed： SUVmax, SUVmean, MTV, TLG and T stage of the original NSCLC had statistically significant differences in the two groups of whether regional lymph node metastasis （ P 〈0. 05）. Spearman Correlation Analysis showed ： there were correlation between SUVmax, SUVmean, MTV, TLG and T stage of the original NSCLC and region- al lymph node metastasis （ P 〈 0. 05）. Dependent samples nonparametric hypothesis test Kruskal-Wallis test analysis showedz In addition to MTV, SUVimix, SUVmean, TLG and T stage of the original NSCLC had statistically significant differences in N stage （ P 〈0.05 ）. Spearman Correlation Analysis showed： SUVmax, SUVmean, TLG and T stage of the original NSCLC and N stage had correlation （ P G0.05）, but MTV had no correlation with N stage （ P 〉0.05）. AUC of SUVmax. Vmean, MTV, TLG, T stage diagnosed the regional lymph node metastasis were 0. 699, 0. 695, 0. 640, 0. 728, 0. 626 respectivel y（ P〈0. 05）. TLG had the largest AUC in those parameters, and had higher diagnosis efficiency. When the optimal cut off value of TLG was 13. 765（g）, the sensitivity, specificity, positive predictive value, negative predictive value and accuracy of TLG to diagnose regional lymph node metastasis were 83.7 %, 45.5 %, 65.1 %.4, 71.4% and 67.7% respectively. Conclusion SUVmax, SUVmean, MTV, TLG and T stage are related to regional lymph node metastasis and N stage in NSCLC, TLG has higher diagnosis efficiency in whether regional lymph node metastasis, and it can be used as a judgment of NSCLC with regional lymph node metastasis.