Breast cancer is the most frequent malignant tumor among women globally and the primary cause of cancer-related deaths in this demographic, excluding non-melanoma skin cancer. For the treatment of individuals with breast cancer, imaging is essential, particularly for the early detection of breast tumors that cannot be felt.
Ultrasound and mammography are the imaging modalities most frequently utilized in this situation. In Brazil, breast ultrasonography is a popular technique because to its affordability and accessibility. While it can be used as a screening tool for young people with thick breasts and a high risk of breast cancer, it is often advised for the supplemental study of regions identified worrisome on mammography or clinical examination.
With the development of imaging biomarkers that directly affect patient care, diagnostic imaging is going through a paradigm shift as a result of the ongoing integration of new technologies, which has increased diagnostic accuracy to a level sufficient to meet the current ideas of personalized medicine. An evaluation of imaging techniques that is more precise, impartial, effective, and repeatable may be possible with the use of artificial intelligence (AI).
AI-based studies have previously been used in a variety of clinical situations and breast imaging modalities, such as lesion identification and classification and breast cancer risk prediction. Thus, this study assessed the outcomes of using AI-based software to forecast the risk of breast mass cancer based on ultrasound pictures.
The patients with MRI-identified breast lesions who had targeted ultrasound and percutaneous ultrasound-guided biopsy were included in this investigation. AI-based software was used to assess the ultrasound data, which were then compared to the abnormal findings.
A total of 334 lesions (183 mass and 151 non-mass) were assessed. Histological investigation revealed 77 (23.1%) malignant lesions and 257 (76.9%) benign ones. Radiologists and AI algorithms both showed great sensitivity in predicting the lesions' likelihood of malignancy. When compared to the radiologist's evaluation alone, the specificity was greater when the radiologist used the AI program (p < 0.001).
On targeted ultrasonography, every lesion that the radiologist or the AI program identified as BI-RADS 2 or 3 (n = 72; 21.6%) displayed benign pathology findings. Overall, and non-mass lesions on targeted ultrasound. This should lead to fewer false positives without missing tumors and assist more precise biopsy decisions.
Reference:
Lima, I. R. M., Cruz, R. M., de Lima Rodrigues, C. L., Lago, B. M., da Cunha, R. F., Damião, S. Q., Wanderley, M. C., & Bitencourt, A. G. V. (2025). Performance of AI-Based software in predicting malignancy risk in breast lesions identified on targeted ultrasound. European Journal of Radiology, 191(112339), 112339. https://doi.org/10.1016/j.ejrad.2025.112339
Disclaimer: This website is primarily for healthcare professionals. The content here does not replace medical advice and should not be used as medical, diagnostic, endorsement, treatment, or prescription advice. Medical science evolves rapidly, and we strive to keep our information current. If you find any discrepancies, please contact us at corrections@medicaldialogues.in. Read our Correction Policy here. Nothing here should be used as a substitute for medical advice, diagnosis, or treatment. We do not endorse any healthcare advice that contradicts a physician's guidance. Use of this site is subject to our Terms of Use, Privacy Policy, and Advertisement Policy. For more details, read our Full Disclaimer here.
NOTE: Join us in combating medical misinformation. If you encounter a questionable health, medical, or medical education claim, email us at factcheck@medicaldialogues.in for evaluation.