Breast mpMRI beyond conventional analysis: a narrative review of current methods and emerging research gaps in segmentation, functional imaging, and AI

Purpose Multiparametric breast magnetic resonance imaging (mpMRI) has become an important tool for the characterization of breast lesions, particularly in high-risk patients and in cases with inconclusive findings from conventional imaging techniques. This review summarizes current methodological approaches to breast mpMRI, through a narrative review of 32 selected studies, with particular attention to lesion segmentation, functional parameter analysis, artificial intelligence (AI)-based models, and non-mass enhancement (NME) lesions. Results mpMRI enables the assessment of functional patterns associated with malignancy by integrating structural and functional information derived from diffusion-weighted imaging (DWI), apparent diffusion coefficient (ADC) maps, and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). AI approaches, including radiomics, deep learning, and multiparametric fusion, have demonstrated improved diagnostic performance. However, clinical evaluation, especially in NME lesions, remains complex, time-consuming, and highly dependent on manual analysis. Additional challenges concern interpretability, standardization, and generalization across datasets and clinical settings. Conclusion The reviewed evidence suggests that progress in breast mpMRI depends less on the isolated development of individual techniques and more on their articulation within coherent analytical frameworks. Approaches combining robust lesion segmentation, quantitative functional information, multiparametric analysis and explainable AI tools may support more consistent, interpretable, and clinically applicable breast MRI workflows, particularly in complex scenarios such as non-mass enhancement lesions.