Model Weights and Environment Configuration for BCBM-Net: A Multimodal Transformer Framework for Predicting Postoperative Bone Metastasis in ER-Positive/HER2-Negative Breast Cancer

This dataset provides the pretrained model weights and computational environment configuration associated with the study entitled “Transformer-Based Multimodal Fusion Network for Predicting Postoperative Bone Metastasis in ER-Positive/HER2-Negative Breast Cancer: A Multi-Institutional Retrospective Study.” The deposited files are intended to support reproducibility and independent validation of BCBM-Net, a multimodal deep learning framework integrating clinical variables, ultrasound imaging features, and digital pathology-derived features for individualized prediction of postoperative bone metastasis.The resource includes pretrained model weights and environment configuration files specifying the software dependencies and computational settings required for model implementation and evaluation.

These materials can be used together with the publicly available source code to reproduce the model workflow and facilitate further methodological research and external validation.