双能CT评估口腔咽喉部鳞癌分化程度的研究  

Dual-energy CT evaluation of differentiation of head and neck squamous cell carcinoma

作　　者：李程超 林桂涵 陈炜越 陈春妙 应海峰 纪建松[1] LI Cheng-chao;LIN Gui-han;CHEN Wei-yue;CHEN Chun-miao;YING Hai-feng;JI Jian-song(Department of Radiology,Key Laboratory of Imaging Diagnosis and Minimally Invasive Intervention Research,The Fifth Affiliated Hospital of Wenzhou Medical University,Zhejiang 323000,China)

机构地区：[1]温州医科大学附属第五医院放射科、浙江省影像诊断与介入微创研究重点实验室,浙江丽水323000

出　　处：《影像诊断与介入放射学》2022年第5期351-357,共7页Diagnostic Imaging & Interventional Radiology

摘　　要：目的本研究旨在探讨双能CT在口腔、咽喉部鳞状细胞癌(SCC)分化程度评估中的应用价值。方法回顾性分析经手术病理证实的148例口腔、咽喉部SCC患者,收集临床资料包括年龄、性别、吸烟史及肿瘤位置。所有患者术前均接受双能CT扫描,测量并计算病灶动脉期和静脉期标准化碘浓度(NIC)、能谱曲线斜率(λ_(HU))及标准化有效原子序数(nZ_(eff))。根据病理结果,将所有患者分为低分化组和中高分化组,比较两组患者临床特征和双能CT参数的差异。使用多因素逻辑回归分析筛选影响口腔、咽喉部SCC分化程度的独立预测因素,分别构建临床特征模型、双能CT参数模型和联合模型。采用受试者工作特征(ROC)曲线评估各模型的诊断效能,并计算曲线下面积(AUC)、敏感度和特异度。结果低分化组46例,年龄(49.28±11.67)岁,其中口腔和口咽部10例,下咽部28例,喉部8例;高分化组102例,年龄(55.03±9.07)岁,其中口腔和口咽部12例,下咽部18例,喉部72例。年龄和肿瘤位置在两组患者间差异均有统计学意义(均P<0.05)。低分化组患者的动、静脉期NIC、λ_(HU)及nZ_(eff)[M(Q_(1),Q_(3))]均显著高于中高分化组[动脉期:0.26(0.22,0.33)比0.20(0.15,0.22),2.03(1.86,2.31)比1.56(1.40,1.93),0.80(0.78,0.83)比0.77(0.75,0.80);静脉期:0.48(0.45,0.58)比0.42(0.35,0.49),2.75(2.20,3.06)比2.26(2.02,2.66),0.87(0.83,0.90)比0.85(0.82,0.88),均P<0.05]。临床特征模型的AUC、敏感度及特异度分别为0.781、76.47%和71.74%;双能CT参数模型的AUC、敏感度及特异度分别为0.863、83.33%和76.09%;联合模型诊断效能最高,AUC可达0.919,敏感度及特异度分别为86.27%和86.96%,明显优于临床特征模型(Z=2.596,P<0.001)和双能CT参数模型(Z=3.589,P<0.001)。结论双能CT用于术前无创评估口腔、咽喉部SCC分化程度具有较好的应用价值,联合临床特征后可明显提升诊断效能。Objective To evaluate dual-energy CT(DECT)in determining the degree of head and neck squamous cell carcinoma(SCC)differentiation.Methods Clinical data including age,gender,smoking history,and tumor location were retrospectively collected on 148 patients with pathologically confirmed head and neck SCC.All patients were divided into a poorly(46)and moderately-highly differentiated(102)groups and underwent preoperative DECT.The normalized iodine concentration(NIC),slope of the energy spectrum curve(λ_(HU)),and the normalized effective atomic number(nZ_(eff))in the arterial and venous phases of the lesions were measured.The clinical characteristics and DECT parameters were compared between the two groups.Multivariate logistic regression analysis was used to screen the independent predictors of oral SCC differentiation.The clinical characteristics model,DECT parameter model,and combined model were constructed.The diagnostic performance of each model was assessed using receiver operating characteristic(ROC)curve analysis.Results The 46 poorly differentiated SCC[age:(49.28±11.67)years]was in the oral cavity and oropharynx(10),hypopharynx(28),and larynx(8).The 102 well-differentiated SCC[age:(55.03±9.07)years]was in the oral cavity and oropharynx(12),hypopharynx(18),larynx(72).There were significant differences in age and tumor locations between the two groups(all P<0.05).The NIC,λHU and nZeff of the poorly differentiated group in the arterial(0.26,2.03,0.80)and venous(0.48,2.75,0.87)phases were significantly(all P<0.05)higher than those of the moderately-highly differentiated group(arterial 0.20,1.56,0.77;venous 0.42,2.26,0.85).The combined model had significantly better diagnostic performance(Z=3.589,P<0.001)with AUC of 0.919,sensitivity of 86.27%,and specificity of 86.96 compared to that of clinical model(0.781,76.47%,71.74%)and DECT parameter model(Z=2.596,P<0.001;0.863,83.33%,76.09%).Conclusion DECT combined with clinical features can significantly improve the diagnostic efficiency of head and neck SCC different

关 键 词：口腔部肿瘤 咽喉部肿瘤 鳞状细胞癌 双能成像 体层摄影术 X线计算机 

分 类 号：R739.8[医药卫生—肿瘤] R739.6[医药卫生—临床医学] R814.42