
External validation testing revealed a deep learning combination of breast MRI, ultrasound and clinical factors had a 10 percent higher AUC for predicting axillary lymph node metastasis than sole use of MRI- or ultrasound-based deep learning models in patients with breast cancer.



















































