Ultrasound Imaging Shows Promise in Diagnosing Non-Massive Breast Cancer

Research conducted in Guangdong, People's Republic of China, has found that ultrasound imaging, particularly in the intratumoral and peritumoral 2 mm regions, has significant potential for diagnosing non-massive breast cancer. A team of researchers from the Department of Ultrasonography at the Tenth Affiliated Hospital of Southern Medical University developed a nomogram that demonstrated high predictive performance in distinguishing between benign and malignant breast lesions. The study used a combination of machine learning algorithms and radiomics models to analyze ultrasound images and identify key features that could indicate the presence of cancer.

Key Takeaways:

• The research used a total of 851 radiomics features extracted from intratumoral and peritumoral regions of interest (ROIs) to develop machine learning models for breast cancer diagnosis.
• The best model combined with clinical ultrasound predictive factors was evaluated using ROC curves, calibration curves, and decision curve analysis (DCA) and demonstrated high predictive performance with C-index values of 0.982 and 0.978.
• The nomogram developed from this model showed high predictive performance in distinguishing between benign and malignant breast lesions, outperforming other models that did not include the intratumoral and peritumoral 2 mm regions.
• The study concluded that ultrasound imaging has significant potential for diagnosing non-massive breast cancer, and the nomogram can assist clinical decision-making.
• A total of 13 machine learning algorithms were employed to create radiomics models for the intratumoral and peritumoral areas.
• Five peritumoral ROIs were generated by extending the contours of the intratumoral ROI by 1 to 5 mm.
• Manual segmentation of ultrasound images was performed to define the intratumoral region of interest (ROI).

Statistics:

• A total of 851 radiomics features were extracted from the intratumoral and peritumoral ROIs.
• C-index values of 0.982 and 0.978 were achieved using the best model combined with clinical ultrasound predictive factors.
• The model incorporating the intratumoral and peritumoral 2 mm regions outperformed other models, indicating its effectiveness in distinguishing between benign and malignant breast lesions.

Sources:

• The value of intratumoral and peritumoral ultrasound radiomics model constructed using multiple machine learning algorithms for non-mass breast cancer. Scientific Reports, 2025;15(1):19953.
• Nature Publishing Group - www.nature.com/
• Scientific Reports - www.nature.com/srep/