Automatic selection of representative slice from cine-loops of real-time sonoelastography for classifying solid...

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Radiology

Background:

• Accurate classification of breast masses as benign or malignant is crucial for effective patient management.
• Real-time sonoelastography provides valuable information on tissue stiffness, aiding in breast mass characterization.
• Manual selection of representative slices from ultrasound elastography cine-loops can be subjective and time-consuming.

Purpose of the Study:

• To evaluate the performance of an automated system for selecting representative slices from real-time sonoelastography cine-loops.
• To compare the diagnostic accuracy of automated slice selection methods against radiologist selection for breast mass classification.

Main Methods:

• A novel computer-assisted system was developed for automatic lesion segmentation from cine-loops.
• Methods included edge-detection, region growing, and motion registration for segmentation.
• Signal-to-noise ratio (SNR(e)) and contrast-to-noise ratio (CNR(e)) were computed to identify optimal slices for differentiation.

Main Results:

• The automated selection using maximum SNR(e) demonstrated the highest diagnostic performance (accuracy 84.4%, sensitivity 83.3%, specificity 85.0%, A(z) 0.90).
• This outperformed other automated methods (maximum CNR(e), maximum compression, maximum strain) and manual radiologist selection (accuracy 77.3%).
• The study included 141 ultrasound elastographic studies (93 benign, 48 malignant masses).

Conclusions:

• Automatic selection of representative slices from real-time sonoelastography cine-loops is a practical and objective approach.
• This automated method offers an accurate strategy for classifying solid breast masses.
• The developed system has the potential to improve the efficiency and reliability of breast mass diagnosis.