鼻咽癌18F-FDG PET/CT 代谢参数与EGFR表达水平的相关性研究

目的 探讨18F-氟脱氧葡萄糖（FDG） PET/CT相关代谢参数预测鼻咽癌治疗前表皮生长因子受体（EGFR）表达水平的价值。 方法 回顾性分析2017年1月至2019年11月广西省柳州市工人医院收治的61例男性44例、女性17例，年龄21~84（57.8±6.2）岁]经组织病理学检查确诊的鼻咽癌患者的临床资料及PET/CT显像资料。分析所有患者的18F-FDG PET/CT图像，计算最大标准化摄取值（SUV_max）、肿瘤代谢体积（MTV）、肿瘤糖酵解总量（TLG）。所有患者的肿瘤标本经链霉菌抗生物素蛋白-过氧化物酶连结（sp）法免疫组织化学染色法染色后，计算标本的EGFR阳性表达情况。采用χ2检验或Fisher确切概率法分析性别、肿瘤T分期、最大径、病理分型等计数资料；SUV_max、MTV、TLG两样本的组间比较采用Mann-Whitney U秩和检验；年龄的组间比较采用独立样本t检验。采用单因素Logistic回归方程分析临床病理因素、代谢参数等因素对EGFR表达水平的影响，把单因素分析中差异有统计学意义的因素纳入多因素Logistic回归分析模型中，评估EGFR表达水平的独立影响因素。绘制鼻咽癌患者代谢参数的受试者工作特征（ROC）曲线，评估其诊断效能。 结果 61例鼻咽癌患者的EGFR阳性表达率为93%（57/61），EGFR高表达34例（高表达组），EGFR低表达27例（低表达组）。EGFR高表达组的SUV_max、MTV、TLG明显高于低表达组（U=?4.197、?2.273、?2.425，均P<0.05）。EGFR高表达组与EGFR低表达组在肿瘤T分期、肿瘤最大径之间的差异均有统计学意义（χ2=11.128、5.165，均P<0.05），而在性别（χ2=0.720）、年龄（t=?0.087）、病理分型（χ2=1.914）之间的差异均无统计学意义（均P>0.05）。单因素回归分析结果显示，肿瘤T分期（OR=0.103，95%CI:0.018~0.582，P=0.025）、肿瘤最大径（OR=1.612，95%CI:1.090~2.385，P=0.017）、SUV_max（OR=1.270，95%CI:1.115~1.446，P<0.01）、MTV（OR=1.008，95%CI:1.002~1.014，P=0.015）、TLG（OR=1.085，95%CI:1.015~1.160，P=0.016）是EGFR表达水平的影响因素。多因素回归分析结果显示，SUV_max（OR=1.340，95%CI:1.019~1.764，P=0.036）是预测EGFR高表达的独立影响因素。ROC曲线结果表明，SUV_max曲线下面积（AUC）=0.815，95%CI:0.709~0.921，P<0.01]对于预测EGFR高表达的诊断效能优于MTV（AUC=0.682，95%CI:0.548~0.816，P=0.015）、TLG（AUC=0.670，95%CI:0.535~0.805，P=0.023）。SUV_max、MTV、TLG临界值分别为16.39、136.29 cm3、19.28时诊断鼻咽癌的灵敏度分别为58.8%、50.0%、41.2%，特异度分别为96.3%、85.2%、96.3%。 结论 18F-FDG PET/CT代谢参数SUV_max可作为预测鼻咽癌患者EGFR表达水平的独立影响因素。

Correlation study of 18F-FDG PET/CT metabolic parameters and EGFR expression level in nasopharyngeal carcinoma

Objective To investigate the value of 18F-fluorodeoxyglucose (FDG) PET/CT related metabolic parameters in predicting epidermal growth factor receptor (EGFR) expression levels before the treatment of nasopharyngeal carcinoma (NPC). Methods The clinical and PET/CT imaging data of 61 patients (44 males and 17 females, aged 21–84 (57.8±6.2) years) with NPC that was diagnosed via hiopathological examination and who were admitted to the Nuclear Medicine Department of Guangxi Liuzhou Worker's Hospital from January 2017 to November 2019 were retrospectively analyzed. The 18F-FDG PET/CT images of all patients were analyzed to calculate the maximum standard uptake value (SUV_max), tumor metabolic volume (MTV), and total lesion glycolysis (TLG). The samples of all patients were subjected to streptavidin-perosidase immunohistochemical staining, and the positive expression of EGFR was calculated. Gender, T stage, pathological classification, maximum tumor diameter, and other enumeration data were analyzed through χ2 test or Fisher's exact probability. Mann–Whitney U test was adopted for the comparison of SUV_max, MTV, and TLG between two samples. Independent samples t test was applied for the comparison of ages between groups. Single factorial Logistic regression equation was used for the analysis of clinicopathological factors and metabolic parameters to evaluate the predictors of EGFR expression levels, and factors (P<0.05) with statistical significance in single factorial analysis were included in the multiple factorial Logistic regression analysis model to evaluate the independent predictors of EGFR expression levels. The receiver operator characteristic (ROC) curves of the three groups of the metabolic parameters of patients with NPC were drawn to evaluate diagnostic efficacy. Results The 61 NPC patients had positive EGFR expression rates of 93% (57/61). Among the patients, 34 had high EGFR expression levels and 27 had low EGFR expression levels. The SUV_max, MTV, and TLG of the high EGFR expression group were evidently higher than those of the low EGFR expression group (U=?4.197, ?2.273, ?2.425; all P<0.05). T stage and maximum tumor diameter were related to EGFR expression levels (χ2=11.128, 5.165; both P<0.05), whereas the difference in gender (χ2=0.720), age (t=?0.087), and pathological classification (χ2=1.914) between the high and low EGFR expression groups lacked statistical significance (all P>0.05). Single factorial regression analysis showed that patients with high T stage (OR=0.103, 95%CI: 0.018–0.582, P=0.025), maximum tumor diameter (OR=1.612, 95%CI: 1.090–2.385, P=0.017), SUV_max (OR=1.270, 95%CI: 1.115–1.446, P<0.01), MTV (OR=1.008, 95%CI: 1.002–1.014, P=0.015), and TLG (OR=1.085, 95%CI: 1.015–1.160, P=0.016) were likely to have high EGFR expression, and multiple factorial regression analysis indicated that SUV_max (OR=1.340, 95%CI: 1.019–1.764, P=0.036) was the independent predictor of high EGFR expression levels. The ROC curves indicated that the diagnostic efficacy of SUV_max (AUC=0.815, 95%CI: 0.709–0.921, P<0.01) in predicting high EGFR expression levels was obviously superior to that of MTV (AUC=0.682, 95%CI: 0.548–0.816, P=0.015) and TLG (AUC=0.670, 95%CI: 0.535–0.805, P=0.023). When the critital values of SUV_max, MTV and TLG were 16.39, 136.29 cm3, 19.28 respectively, the sensitivities were 58.8%, 50.0%, and 41.2%, respectively, and the specificities were 96.3%, 85.2%, and 96.3%, respectively. Conclusions The metabolic parameter SUV_max of 18F-FDG PET/CT can serve as the independent predictor of EGFR expression levels in patients with NPC. Moreover, 18F-FDG PET/CT examination can enable the noninvasive assessment of EGFR expression levels.