An AI algorithm may assist radiologists in improving their performance on breast ultrasound and identifying early breast cancer, according to research published May 25 in eClinicalMedicine.
Researchers led by Jianwei Liao, PhD, from Southwest Hospital of Third Military Medical University in Chongqing, China, found that their deep-learning algorithm could help by identifying subtle elements on breast ultrasound. The model itself also achieved high diagnostic performance.
"The proposed model can extract the morphological features from early breast lesions, conduct effective and objective image analysis, and provide precise outcome for early breast cancer," Liao and colleagues wrote.
Although breast ultrasound is a go-to supplemental imaging method for confirming suspicious findings on mammography, previous reviews suggest that cancer indicators are visible on about 11% of early examinations interpreted as normal. The researchers pointed out that this may be due to factors such as noise, contrast, illumination, and resolution.
Previous studies also suggest that AI can effectively assist radiologists in breast image interpretation. In their project, Liao and colleagues developed a deep-learning model called EDL-BC to identify high-risk lesions in ultrasound images. It has two separate feature extraction modules for B-mode and Doppler images. The team wrote that it directed the model to uncover more discriminative representation from ultrasound image samples.
The researchers found that the model had high performance in both internal and external validation cohorts, marked by high area under the curve (AUC) measures. They tested the model on one internal validation cohort and two external validation cohorts. The internal cohort consisted of 7,955 lesions from 6,795 women at the First Affiliated Hospital of Army Medical University in Chongqing. The external validation cohorts included 448 lesions from 391 patients in the Tangshan People's Hospital, and 245 lesions from 235 patients in the Dazu People's Hospital, both in Chongqing.
| Performance of deep learning model in training, validation sets | |||
| Internal validation set | External validation set No. 1 | External validation set No. 2 | |
| AUC | 0.95 | 0.956 | 0.907 |
| Sensitivity | 94.4% | 100% | 80% |
The team also found that the model itself (AUC = 0.945) and the radiologists who were assisted by the model (AUC = 0.899) had higher diagnostic performance than those who had no such assistance (AUC = 0.716, p < 0.0001).
The researchers also found no significant differences between the model and radiologists with AI assistance (p = 0.099).
They made two recommendations based on the results of this study. These include highlighting that the model could improve the accuracy of diagnosis and that the model can discriminate against the high risk of early breast cancer, as well as reduce the misdiagnosis rate and avoid unnecessary invasive biopsies.
"Notably, the most crucial role of the proposed model is improving diagnostic accuracy by assisting clinicians," the study authors wrote. "The performance of this model in clinical cases shows its effectiveness in assisting the doctors to improve the diagnostic accuracy."
![Examples of ultrasound findings and techniques. (A) Images in a 39-year-old male patient with a mass in the left thigh. The mass is heterogeneous on the B-mode US image (compared with the patient in D) and showed increased microvascularity (superb microvascular imaging [SMI]) and shear-wave elastography (SWE) values. Undifferentiated pleomorphic sarcoma was diagnosed at biopsy (with pleomorphic rhabdomyosarcoma in surgical specimen). (B) Images in an 18-year-old male patient with a mass in the left leg. The mass is hypoechoic on the B-mode image, with no other findings suggestive of malignancy. The lesion is in contact with the cortex of the tibia, which is slightly irregular. CT revealed a doubtful anteromedial tibial erosion. The microvascular study demonstrated high vascularization, suggestive of malignancy. Periosteal Ewing sarcoma was diagnosed with both histologic and immunohistochemical confirmation. (C) Images in a 69-year-old female patient with a lump growing on the outside of the left leg. Multiple SWE examinations were performed (please note the high values obtained in the measurements, whereas the color map highlights the stiffness relative to adjacent tissues). SMI showed areas of increased vascularization to target for sampling. Undifferentiated spindle cell sarcoma was diagnosed at biopsy, with residual leiomyosarcoma in the surgical specimen after neoadjuvant therapy. (D) Images in a 56-year-old female patient with a mass in the right thigh. The mass is heterogeneous at both B-mode ultrasound (similar to patient A) and MRI (coronal T2-weighted spectral attenuated inversion recovery [SPAIR]; T1-weighted pre-contrast and postcontrast imaging), which even shows uptake after the administration of paramagnetic contrast material, which is traditionally suggestive of malignancy. Low values at SMI and elastography are suggestive of benignity. Spindle cell lipoma was diagnosed at biopsy, with atypical spindle cell lipomatous tumor in the surgical specimen.](https://img.auntminnie.com/mindful/smg/workspaces/default/uploads/2026/08/images-radiol250278fig2.APCFLSvX6p.jpg?auto=format%2Ccompress&dpr=2&fit=crop&h=167&q=70&w=250)
