论文部分内容阅读
目的探讨计算机辅助诊断系统(CADx)在微钙化检测与特征提取基础上的分类对于导管原位癌(DCIS)的诊断价值。方法回顾性分析经南方医科大学附属南海医院及中山大学肿瘤防治中心行乳腺X线摄影检查发现微钙化并经病理学证实的623例患者影像资料,其中,良性病变378例,DCIS 245例。用受试者操作特征曲线(ROC)分别分析采用计算机方法提取的每个微钙化特征对于这两类病变判别的诊断效能,和应用所有微钙化特征集合并基于支持向量机(SVM)分类器的CADx的分类诊断效能。结果 CADx对于良性病变和DCIS这两类病变微钙化分类的ROC曲线下面积(Az)为0.853;特异度、准确率、敏感度分别为70.1%、82.1%、90.7%,高于单个微钙化特征的诊断效能。结论采用CADx对于DCIS微钙化能较好的检测与定位,对乳腺癌早期病变的识别能提供有益的参考。
Objective To investigate the diagnostic value of computer-aided diagnosis system (CADx) on the diagnosis of ductal carcinoma in situ (DCIS) based on the detection of micro-calcifications and feature extraction. Methods The data of 623 patients with microcalcifications and pathologically confirmed by mammography examination at Nanhai Hospital affiliated to Southern Medical University and Cancer Center of Sun Yat-sen University were analyzed retrospectively. Among them, 378 were benign lesions and 245 were DCIS. The receiver operating characteristic curve (ROC) was used to separately analyze the diagnostic efficacy of each of the microcalcifications extracted using computer-based methods in discriminating between the two groups of lesions and using the set of all the microcalcification features and based on Support Vector Machine (SVM) classifiers CADx classification diagnostic efficacy. Results The area under the ROC curve of CADx was 0.853 for the classification of benign lesions and DCIS. The specificity, accuracy and sensitivity of CADx were 70.1%, 82.1% and 90.7% respectively, which were higher than those of single micro-calcifications The diagnostic efficacy. Conclusion The detection and localization of DCIS microcalcification using CADx can provide a good reference for the identification of early stage breast cancer.