Development of an Automated CAD System for Lesion Detection in DCE-MRI.

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Radiology

Background:

• Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is crucial for early breast lesion detection and characterization.
• Automated systems can enhance the efficiency and accuracy of interpreting DCE-MRI scans.

Purpose of the Study:

• To develop and evaluate an automated computer-aided diagnosis (CAD) system for detecting breast lesions in DCE-MRI.
• To assess the performance of different classifiers in identifying suspicious regions.

Main Methods:

• The system preprocesses DCE-MRI images by cropping breast tissue and segmenting regions of interest (ROIs) using Otsu's multilevel thresholding.
• A two-stage false positive reduction process is employed, followed by feature extraction and classification using feed-forward backpropagation neural networks (FFBPN) and support vector machines (SVM).

Main Results:

• The CAD system demonstrated high performance on a dataset of 52 DCE-MRI exams.
• Sensitivity reached 83% with FFBPN and 92% with SVM.
• The SVM classifier achieved an area under the curve (AUC) of 0.95, indicating excellent diagnostic capability.

Conclusions:

• The developed CAD system effectively detects enhancing breast lesions in DCE-MRI images.
• The SVM classifier shows superior performance in lesion identification and differentiation from healthy tissue.
• Further research with larger datasets is recommended to validate these findings.