摘要
目的探讨双层探测器光谱CT(DLCT)虚拟平扫(VNC)和碘图在结核性胸膜炎(TP)、肿瘤性胸膜炎(NP)鉴别诊断中的价值。方法选取行胸部DLCT增强检查的22例TP(TP组)和20例NP患者(NR组)的临床和影像资料。通过DLCT静脉期增强图像重建VNC图像和碘图,并分别在VNC图像测量渗出液的CT值,在碘图测量胸膜病变的碘浓度(IC)值。比较TP、NP两组患者临床资料、渗出液CT值和胸膜病变IC值的差异。对组间存在统计学差异的指标采用多因素Logistic回归,寻找鉴别TP和NP的独立指标。采用ROC曲线分析渗出液CT值、胸膜病变IC值及两者联合应用鉴别TP和NP的效能,通过Delong检验对各自鉴别效能进行比较。结果TP组患者年龄和胸膜病变的IC值明显小于NP组,渗出液CT值明显高于NP组,差异有统计学意义(均P<0.001)。多因素二项Logistic回归发现年龄并非鉴别TP和NP的独立指标,而渗出液CT值和胸膜病变IC值均为鉴别两者的独立指标,差异有统计学意义(均P<0.05)。渗出液CT值、胸膜病变IC值及两者联合鉴别TP和NP的曲线下面积(Area under the curve,AUC)分别为0.830、0.869和0.966,其中两者联合的AUC明显大于CT值和IC值的AUC,差异有统计学意义(均P<0.05)。渗出液CT值和胸膜病变IC值的最佳阈值分别为14.5 HU和1.85 mg/ml。渗出液CT值、胸膜病变IC值和两者联合鉴别TP和NP的敏感性分别为68.2%、70.5%和90.4%,特异性分别为95.0%、94.5%和95.5%,约登指数分别为0.632、0.650和0.859。结论基于DLCT的VNC图像和碘图均是鉴别TP和NP的独立指标,联合应用在鉴别TP和NP中具有较高的应用价值。
Objective To investigate the value of dual-layer spectral detector CT(DLCT)Virtual noncontrast(VNC)image and iodine map in the differential diagnosis of tuberculous pleurisy(TP)and neoplastic pleurisy(NP).Methods The clinical and imaging data of 22 TP and 20 NP patients who underwent chest DLCT enhancement examination in the Second Hospital of Nanjing were retrospectively analyzed.Clinical data of patients were recorded through electronic medical record system.The VNC image and iodine map were reconstructed by DLCT venous phase enhanced image.The CT value of effusion was measured on VNC image,and the iodine concentration(IC)value of pleural lesion was measured on iodine map.The clinical data,CT value of effusion and IC value of pleural lesions were compared between TP and NP groups.Multivariate Logistic regression was used to identify the independent factors of TP and NP for the indicators with statistical differences between groups.The receiver operating characteristic(ROC)curve was used to analyze the differential efficacy of CT value of effusion,IC value of pleural lesion and their combined application in differentiating TP and NP,and the differential efficacy was compared by Delong test.Results The age and IC values of pleural lesions of TP group were significantly lower than those of NP group,and the CT values of effusion were significantly higher than those of NP group(all P<0.001).Multivariate binomial logistic regression showed that age was not an independent indicator in the differentiation of TP and NP,while CT value of effusion and IC value of pleural lesion were independent indicators in the differentiation of TP and NP(both P<0.05).The area under the curve(AUC)of the effusion CT value,pleural lesion IC value and their combination to distinguish TP and NP were 0.830,0.869 and 0.966,respectively.The AUC of the combination of the two independent factors was significantly greater than that of the CT value and IC value(both P<0.05).The optimal thresholds for CT value of effusion and IC value of pleural lesion were 14.5 HU and 1.85 mg/ml,respectively.The sensitivity of CT value of effusion,IC value of pleural lesion and their combination in the identification of TP and NP were 68.2%,70.5%and 90.4%,the specificity were 95.0%,94.5%and 95.5%,and the Youden index were 0.632,0.650 and 0.859,respectively.Conclusion VNC image and iodine map based on DLCT are both independent indicators for the identification of TP and NP,and their combined application has high application value in the identification of TP and NP.
作者
田玉亭
李代欣
付志浩
邢富臣
许传军
TIAN Yuting;LI Daixin;FU Zhihao;XING Fuchen;XU Chuanjun(Department of Radiology,Nanjing Hospital Affiliated to Nanjing University of Traditional Chinese Medicine,The Second Hospital of Nanjing,Nanjing 210003,China;Department of Thoracic Surgery,Nanjing Hospital Affiliated to Nanjing University of Traditional Chinese Medicine,The Second Hospital of Nanjing,Nanjing 210003,China)
出处
《医学影像学杂志》
2023年第9期1575-1580,共6页
Journal of Medical Imaging
基金
江苏省南京市卫生科技发展专项基金项目(编号:YKK21126)。