Logistic回归联合ROC曲线评价CT增强定量值在术前预测胸腺瘤危险度中的诊断价值  

Value of quantitative computed tomography values on enhancement by Logistic regression and ROC curve in predicting the pathological grade of thymoma

作　　者：章婷婷[1] 宋文[1] 黄国权 余永强[1] Zhang Tingting;Song Wen;Huang Guoquan;Yu Yongqiang(Dept of Radiology,The First Affiliated Hospital of Anhui Medical University,Hefei 230022;Dept of Radiology,The Second People’s Hospital of Wuhu,Wuhu 241000)

机构地区：[1]安徽医科大学第一附属医院放射科,合肥230022 [2]芜湖市第二人民医院放射科,芜湖241000

出　　处：《安徽医科大学学报》2023年第2期315-321,共7页Acta Universitatis Medicinalis Anhui

基　　金：国家自然科学基金(编号:81801679)。

摘　　要：目的探讨临床资料、CT定性特征及定量值在术前预测胸腺瘤危险度中的价值。方法回顾性分析经手术病理证实的81例胸腺瘤患者的临床和CT资料,依据病理结果将患者分为低危组和高危组。对两组患者临床资料、CT定性特征和定量值进行组间比较;采用Logistic单因素和多因素回归分析,寻找预测胸腺瘤危险度的相关因素;对胸腺瘤危险度有预测意义的因素绘制受试者工作特征曲线(ROC曲线)评估其预测效能。结果低危组62例,高危组19例,两组的最大增强CT值和增强前后CT最大差值差异有统计学意义(P<0.001)。Logistic回归分析显示,最大增强CT值(OR=0.88,95%CI:0.83~0.94)和增强前后CT最大差值(OR=0.92,95%CI:0.88~0.97)为胸腺瘤危险度的独立预测因子(P<0.001)。最大增强CT值的ROC曲线下面积(AUC)为0.884,敏感性为68.4%,特异性为96.8%;增强前后CT最大差值的AUC为0.833,敏感性为89.5%,特异性为72.6%。结论基于增强CT的定量参数有助于术前预测胸腺瘤不同风险分层,最大增强CT值预测效能优于增强前后CT最大差值。Objective To investigate the value of clinical data and qualitative and quantitative computed tomography(CT)in predicting the pathological grade of thymoma.Methods The clinical data and CT images of 81 patients with confirmed pathological diagnoses of thymoma were retrospectively analyzed.The patients were divided into low-risk and high-risk groups according to the pathological results,and the clinical features and qualitative and quantitative CT findings were compared between the groups.The efficacy of different parameters in predicting the pathological thymoma grade was evaluated by logistic univariate and multivariate regression analyses and receiver operating characteristic(ROC)curves.Results Sixty-two tumors were classified as low-risk thymoma and 19 as high-risk thymoma.The efficacies of both the maximum CT values on enhancement and maximum degree of enhancement for risk prediction were statistically significant(P<0.001).Logistic regression analysis showed that both the maximum CT values on enhancement(OR=0.88,95%CI:0.83-0.94)and maximum degree of enhancement(OR=0.92,95%CI:0.88-0.97)were independent predictors of thymoma risk(P<0.001).The area under the ROC curve(AUC)of the maximum CT values on enhancement was 0.884,with a sensitivity of 68.4%and a specificity of 96.8%;the AUC of the maximum degree of enhancement was 0.833,with a sensitivity of 89.5%and a specificity of 72.6%.Conclusion Quantitative parameters based on enhanced CT are helpful for the preoperative prediction of thymoma risk stratification.The predictive efficacy of the maximum CT values on enhancement is superior to that of the maximum degree of enhancement.

关 键 词：电子计算机断层扫描 增强定量值 预测 胸腺瘤 危险度 

分 类 号：R445.4[医药卫生—影像医学与核医学]