基于CT影像特征构建联合预测因子在腮腺肿瘤良恶性鉴别中的价值  

The value of constructing joint predictors based on CT image features in the differentiation of benign and malignant parotid gland tumors

作　　者：洪悦[1] 田慧[1] 陈杰[1] 王艳[1] HONG Yue;TIAN Hui;CHEN Jie;WANG Yan(Imaging Center,the People’s Hospital of Xinjiang Uygur Autonomous Region,Urumqi 830001,China)

机构地区：[1]新疆维吾尔自治区人民医院放射影像中心,新疆乌鲁木齐830001

出　　处：《实用放射学杂志》2024年第8期1233-1237,共5页Journal of Practical Radiology

摘　　要：目的探讨基于CT影像特征构建的联合预测因子在鉴别腮腺肿瘤(PGT)良恶性中的价值。方法回顾性分析经手术病理证实的105例PGT患者临床特征和CT征象,采用单因素、二元logistic回归分析筛选独立危险因素。根据其回归系数拟合生成联合预测因子,绘制受试者工作特征(ROC)曲线、计算临界点。应用联合预测因子预测2022年至2023年36例患者PGT的良恶性。结果良恶性PGT在吸烟史、形态、边界、大小、门脉期CT值和强化方式有显著性差异(P<0.05)。logistic回归分析确定形态、边界、门脉期CT值和吸烟史为恶性PGT的独立危险因素(P<0.05)。联合预测因子的曲线下面积(AUC)为0.942,高于单一独立危险因素。联合预测因子的预测正确率80.6%,敏感度93.3%,特异度71.4%。结论联合预测因子对鉴别PGT的良恶性具有较高的效能,可为临床优化治疗方案提供帮助。Objective To investigate the value of constructing joint predictors based on CT image features in the differentiation of benign and malignant parotid gland tumors(PGT).Methods The clinical features and CT signs of 105 patients with PGT confirmed by surgical pathology were analyzed retrospectively,and independent risk factors were screened using univariate,binary logistic regression analysis.Joint predictors were generated by fitting according to their regression coefficients,receiver operating characteristic(ROC)curve were plotted and thresholds were calculated.The joint predictors were applied to predict the benign and malignant of PGT in 36 patients from 2022 to 2023.Results There were significant differences between benign and malignant PGT in smoking history,morphology,borders,size,portal phase CT value,and enhancement method(P<0.05).Logistic regression analysis identified morphology,borders,portal phase CT value,and smoking history as independent risk factors for malignant PGT(P<0.05).The joint predictor’s area under the curve(AUC)was 0.942,higher than the single independent risk factor.The predictive correctness,sensitivity and specificity of the joint predictors were 80.6%,93.3%and 71.4%,respectively.Conclusion The joint predictor has high efficacy in distinguishing benign and malignant PGT and can provide clinical optimization of treatment options.

关 键 词：腮腺肿瘤 良恶性 联合预测因子 计算机体层成像 

分 类 号：R739.91[医药卫生—肿瘤] R814.42[医药卫生—临床医学]