Related Experiment Video
Updated: Mar 27, 2026
15:48
Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging
Published on: December 15, 2014
23.5K
A computer-aided diagnosis system for breast DCE-MRI at high spatiotemporal resolution
Mehmet Ufuk Dalmış1, Albert Gubern-Mérida1, Suzan Vreemann1
1Radiology and Nuclear Medicine, Radboud University Medical Center, Geert Grooteplein 10, Route 766 Nijmegen, Gelderland 6500 HB, The Netherlands.
Medical Physics
|January 10, 2016
Summary
A new computer-aided diagnosis (CADx) system for breast dynamic contrast-enhanced MRI (DCE-MRI) accurately differentiates benign from malignant lesions. This advanced system shows improved performance over previous methods, aiding in breast cancer diagnosis.
Area of Science:
- Radiology and Medical Imaging
- Artificial Intelligence in Medicine
- Biomedical Engineering
Background:
- High spatiotemporal resolution in breast dynamic contrast-enhanced MRI (DCE-MRI) presents challenges in distinguishing benign lesions from malignancies due to similar enhancement patterns.
- Accurate characterization of breast lesions is crucial for effective diagnosis and treatment planning.
Purpose of the Study:
- To develop and evaluate a computer-aided diagnosis (CADx) system for characterizing breast lesions using high spatiotemporal resolution DCE-MRI.
- The system aims to improve the accuracy of differentiating benign from malignant breast lesions.
Main Methods:
- A CADx system was developed, incorporating semiautomated lesion segmentation, computation of morphological and dynamic features, aorta detection, and random forest classification.
- Features were derived from lesion segmentation and contrast enhancement curves, with aorta enhancement time information integrated for automated analysis.
- The system was evaluated on a dataset of 325 patients (223 malignant, 172 benign lesions) using leave-one-out cross-validation and ROC analysis.
Main Results:
- The proposed CADx system achieved an area under the ROC curve (AUC) of 0.8543, significantly outperforming a previous system (AUC = 0.8172, p = 0.007).
- Specific AUC values for ductal carcinoma in situ (DCIS), invasive ductal carcinoma (IDC), and invasive lobular carcinoma (ILC) were 0.7924, 0.8688, and 0.8650, respectively.
- The system demonstrated improved classification performance with reduced user interaction requirements.
Conclusions:
- A novel CADx system for high spatiotemporal resolution DCE-MRI of the breast has been successfully developed.
- The system offers superior performance in classifying benign and malignant breast lesions compared to existing approaches.
- The developed CADx system requires less user interaction, making it a more efficient diagnostic tool.
