This study presents a few-shot learning approach for the classification of ultrasound breast cancer images using meta-learning methods. The breast ultrasound images (BUSI) dataset, which has three classes and is difficult to use in meta-learning, was used for meta-testing in a cross-domain approach along with other datasets for meta-training. The proposed approach achieved an accuracy range of 0.882–0.889 using the ResNet50 backbone with ProtoNet in a 10-shot setting, which represents a significant improvement over the baseline accuracy of 0.831. ProtoNet outperformed the MAML method for all k-shot settings. The use of ResNet models as the backbone network for feature extraction was found to be more successful than the use of a four-layer convolutional model. The methodology used in this study can be adapted to other datasets with similar problems. The study demonstrates that meta-learning can be effectively applied to few-shot classification in the BUSI dataset, providing higher accuracy compared to deep learning methods for medical images with small-scale datasets and few classes. The results show that ProtoNet outperforms MAML in terms of accuracy across all shot values and backbones. The study also highlights the importance of using deeper and more complex backbones, such as ResNet models, for better performance in medical image classification. The results indicate that the proposed approach is effective in classifying breast ultrasound images with limited data. The study also compares the performance of different meta-learning algorithms and shows that ProtoNet outperforms FOMAML in terms of accuracy, sensitivity, specificity, and F1-score. The study concludes that the proposed approach is a promising solution for few-shot classification in medical imaging, particularly for breast cancer diagnosis.This study presents a few-shot learning approach for the classification of ultrasound breast cancer images using meta-learning methods. The breast ultrasound images (BUSI) dataset, which has three classes and is difficult to use in meta-learning, was used for meta-testing in a cross-domain approach along with other datasets for meta-training. The proposed approach achieved an accuracy range of 0.882–0.889 using the ResNet50 backbone with ProtoNet in a 10-shot setting, which represents a significant improvement over the baseline accuracy of 0.831. ProtoNet outperformed the MAML method for all k-shot settings. The use of ResNet models as the backbone network for feature extraction was found to be more successful than the use of a four-layer convolutional model. The methodology used in this study can be adapted to other datasets with similar problems. The study demonstrates that meta-learning can be effectively applied to few-shot classification in the BUSI dataset, providing higher accuracy compared to deep learning methods for medical images with small-scale datasets and few classes. The results show that ProtoNet outperforms MAML in terms of accuracy across all shot values and backbones. The study also highlights the importance of using deeper and more complex backbones, such as ResNet models, for better performance in medical image classification. The results indicate that the proposed approach is effective in classifying breast ultrasound images with limited data. The study also compares the performance of different meta-learning algorithms and shows that ProtoNet outperforms FOMAML in terms of accuracy, sensitivity, specificity, and F1-score. The study concludes that the proposed approach is a promising solution for few-shot classification in medical imaging, particularly for breast cancer diagnosis.