This project explores LLNM-Net, an explainable multimodal AI system that predicts thyroid cancer metastasis by fusing ultrasound images with clinical data. It demonstrates how integrating eight data sources—including nodule location, morphology, and radiology reports—achieves 84.7% accuracy, outperforming human experts by 20.4%.
