研究目的
To describe important CAD systems for diagnostically challenging breast lesions in breast MRI, focusing on small lesion detection and non-mass-like-enhancing lesions based on spatio-temporal features.
研究成果
The study concludes that motion compensation improves the diagnostic value for small lesions and non-mass-enhancing lesions. Dynamical features yield the best results for non-mass-like enhancing lesions, followed by morphological features. The study suggests that future studies are necessary to evaluate the effectiveness of novel descriptors for non-mass-enhancing lesion diagnosis.
研究不足
The study acknowledges the challenges in diagnosing non-mass-enhancing lesions due to their heterogeneous appearance and the limited accuracy of kinetic characteristics in discriminating between malignant and benign behavior. The effectiveness of novel descriptors for non-mass-enhancing lesion diagnosis requires evaluation in large trials.
1:Experimental Design and Method Selection:
The study employs a multistep system including motion artifact reduction, morphologic and dynamic enhancement pattern extraction, and feature selection for lesion evaluation.
2:Sample Selection and Data Sources:
A total of 40 patients with indeterminate small mammographic breast lesions and 84 patients with non-mass-enhancing tumors were examined.
3:List of Experimental Equipment and Materials:
MRI was performed with a
4:5 T system (Magnetom Vision, Siemens, Erlangen, Germany) equipped with a dedicated surface coil. Experimental Procedures and Operational Workflow:
The dynamic study consisted of six measurements with an interval of 83 s for small lesions and five measurements with an interval of
5:4 min for non-mass-enhancing lesions. Data Analysis Methods:
The study uses Fisher’s linear discriminant analysis and SVM with different kernels for classification, with the area under the ROC curve serving as a quantitative evaluation measure.
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容