超声征象logistic回归分析鉴别诊断腮腺良恶性肿瘤  

Differential Diagnosis of Benign and Malignant Parotid Gland Tumors by Logistic Regression Analysis of Ultrasonic Signs

作　　者：张丽 肖颖[1] 张纾[1] 厉亚林 Zhang Li;Xiao Ying;Zhang Shu;Li Yalin(Department of Ultrasound Medicine,Taizhou Municipal Hospital,Taizhou,Zhejiang 318000,China)

机构地区：[1]台州市立医院超声医学科,浙江省台州市318000

出　　处：《中国超声医学杂志》2023年第7期724-727,共4页Chinese Journal of Ultrasound in Medicine

摘　　要：目的应用logistic回归研究腮腺肿瘤的超声征象在鉴别其良恶性中的价值。方法回顾分析193例腮腺良恶性肿瘤患者的声像图特征(形状、边界、回声、液化、钙化、后方回声情况、血流分级及血流模式),以病理结果为金标准,应用单因素分析,将有统计学意义的参数建立logistic回归模型,绘制相应的受试者工作特征(ROC)曲线并评估诊断效能。结果浅表超声检查,腮腺恶性肿瘤相较良性肿瘤而言,形状更不规则、边界不清、回声不均匀、多伴有钙化及囊性变、血流更丰富,两组间差异具有统计学意义(P<0.05)。二分类logistic回归提示边界、钙化、血流分级为相对独立危险因素。ROC曲线显示建立的回归模型对腮腺恶性肿瘤预测的准确度为91.4%,灵敏度为80.7%,特异度为95.5%,阳性预测值为87.5%,阴性预测值为92.8%。结论本研究建立的logistic回归模型,能较好地预测腮腺肿瘤的良恶性程度。Objective To study the value of ultrasonic signs of parotid gland tumor in distinguishing benign and malignant lesions by logistic regression.Methods The ultrasonographic features(shape,boundary,echogenicity,liquification,calcification,posterior echo,blood flow grade and blood flow pattern)of 193 cases,including benign and malignant parotid gland tumors were retrospectively analyzed.Using the pathological results as the gold standard,univariate analysis was used to establish a logistic regression model with statistically significant parameters.Corresponding ROC curve was plotted and diagnostic efficiency was evaluated.Results In superficial ultrasound examinations,parotid gland malignant tumors showed more irregular shape,unclear boundary,uneven echo,calcification and cystic changes,and more abundant blood flow,in contrast with the benign tumors,and the difference between the two groups was statistically significant(P<0.05).Binary logistic regression indicated that boundary,calcification and blood flow grade were relatively independent risk factors.ROC curve showed that the accuracy,sensitivity,specificity,positive predictive value,and negative predictive value of parotid gland malignancy were 91.4%,80.7%,95.5%,87.5%and 92.8%respectively.Conclusions The logistic regression model established in the present study may help tobetter predict the benign and malignant degree of the parotid gland tumors.

关 键 词：腮腺 LOGISTIC回归 超声 鉴别诊断 

分 类 号：R445.1[医药卫生—影像医学与核医学] R739.87[医药卫生—诊断学]