摘要
目的:探讨合理简便的乳腺影像报告和数据系统(BI-RADS)4类病变MRI亚分类方法。方法:收集424例行MRI检查且经病理证实的乳腺病变患者。其中肿块样强化病变302例,采用Fischer’s评分联合ADC值进行亚分类;非肿块样强化病变122例,分析其MRI征象,采用ROC曲线确定ADC值的最佳诊断界值,采用Logistic回归分析建立多参数评分模型进行亚分类。以病理为金标准,计算BI-RADS 4a、4b、4c类病变的阳性预测值,并以BI-RADS 4b为诊断界值计算诊断的敏感度、特异度,分析其与病理结果的一致性。结果:BI-RADS 4类病变共165例,其中4a类病变52例,4b类病变41例,4c类病变72例。ROC曲线分析发现,以BI-RADS 4b类为恶性的诊断界值,敏感度和特异度分别为96.77%、81.44%。Kappa一致性分析发现,以BI-RADS 4b类病变为良恶性的诊断界值,与病理结果的一致性显著(K=0.787)。结论:采用多参数鉴别诊断模型可提高乳腺病变诊断的准确率,是一种可靠的、能快速掌握的BI-RADS 4类疾病的MRI亚分类方法。
Objective:To investigate a reasonable and simple method for the subclassification of breast lesions with BI-RADS Category 4.Methods:A total of 424 patients with breast lesions after MRI examination were collected.302 lesions with mass-like enhancement(MLE)was subclassified by Fischer’s score combined with the ADC value.The MRI features of the other 122 lesions with non-MLE(NMLE)were analyzed,and the ROC curve was used to determine the cutoff for ADC values,and logistic regression analysis was used to establish a multiparameter scoring model for subclassification.Taking the pathological results as the gold standard,the positive predictive values of Category 4a,4b and 4c lesions were calculated,the sensitivity and specificity were calculated with Category 4b as the diagnostic cutoff value,and Kappa test was used to analyze their consistency with pathological results.Results:There were 165 lesions of BI-RADS Category 4,including 52 lesions of Category 4a,41 lesions of Category 4b and 72 lesions of Category 4c.ROC curve shwed that,taking BI-RADS Category 4b as the cutoff for the diagnosis of malignant lesions,the sensitivity and specificity were 96.77%and 81.44%,respectively.Kappa analysis showed that Category 4b lesions were used as the malignant cutoff value,which was significantly consistent with the pathological results(K=0.787).Conclusion:The multi-parameter differential diagnosis model enhances the diagnostic precision of breast lesions,offering a dependable subclassification method.
作者
刘丹丹
吕花营
刘俊业
王巧慧
巴照贵
LIU Dandan;LYV Huaying;LIU Junye;WANG Qiaohui;BA Zhaogui(Department of Medical Imaging,Eighth People’s Hospital of Jinan,Jinan 271126,China)
出处
《中国中西医结合影像学杂志》
2023年第6期652-655,662,共5页
Chinese Imaging Journal of Integrated Traditional and Western Medicine
基金
济南市科技计划项目(202019011)。
关键词
乳腺疾病
乳腺影像报告和数据系统
磁共振成像
亚分类
Breast diseases
Breast imaging-reporting and data system
Magnetic resonance imaging
Subcategorization
