Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4221917 | Clinical Imaging | 2012 | 5 Pages |
Abstract
Doppler ultrasound imaging provides vascular information that could characterize benign and malignant breast masses in many previous publications. In this study, we applied vascular quantification and morphology features derived from three-dimensional power Doppler ultrasound as classifiers based on support vector machine. An Az value under the receiver operating characteristic (ROC) curve was used to measure the significance of each vascularization feature. Sixty solid breast tumors were assessed. According to the Az value for the ROC curve of the selected features, the classification performance of the proposed method was 0.8423, indicating that vascular morphologic information is valuable in the classification of breast lesions.
Keywords
Related Topics
Health Sciences
Medicine and Dentistry
Radiology and Imaging
Authors
Ping-Lang Yen, Hwa-Koon Wu, Hsin-Shun Tseng, Shou-Jen Kuo, Yu-Len Huang, Hong-Ting Chen, Dar-Ren Chen,