Breast cancer is among the most diagnosed cancers and a leading cause of cancer-related deaths in women. Early detection and accurate classification into benign or malignant types are crucial for improving survival rates. This study presents a deep learning model for binary breast tumor classification based on the Wisconsin Breast Cancer Dataset (WBCD) diagnostic version. Unlike traditional approaches, the proposed model introduces 3D tensorization of the input, treating feature vectors as pseudo-sequences. This approach leverages the memory retention and dynamic learning capabilities of LSTMs, enabling not only more efficient and faster model training but also improving performance. Distributed data processing enables the parallelization of computational workloads across multiple nodes, without sharing sensitive patient data, which aligns well with the nature of LSTM networks. By exchanging only model updates, this federated approach preserves data privacy and scalability, making it highly suitable for collaborative healthcare applications and distributed computing environments.