This paper presents a framework for breast MRI data processing in oncology applications with a focus on image registration and deep learning based segmentation. In this study, we present a proposed approach for flexible breast dynamics registration using B-spline and segmentation using U-Net based architecture. For this, we used a combination of the Duke-Breast-Cancer-MRI dataset and dynamic MRI scans from a local hospital, which provide more robust results, achieving a Dice coefficient of 0.96 ± 0.03 for the optimized model. The main objective of this work is to develop a standardized and user-friendly tool for comprehensive breast MRI analysis and to highlight that data preprocessing is essential for clinical studies.