Breast MRI: Where are we currently standing?

Breast cancer is the most frequently occurring malignancy among women, having a great impact on society, economy, and healthcare. It is therefore vital to develop effective imaging methods to perform breast screening, diagnosis, and treatment follow-up. Breast MRI is the most efficient method for screening high-risk patients, for breast lesion differentiation and characterization, and for the assessment of response to treatment. Some novel MRI imaging techniques, such as Diffusion Kurtosis Imaging, perfusion imaging, MR Spectroscopy, hybrid PET/MRI imaging, fMRI and ultra-high field MRI imaging offer the capacity to improve the diagnostic accuracy of breast MRI, while reducing unnecessary biopsies. However, any techniques used in breast MRI should be treated with caution, and after a thoughtful consideration of its main strengths and weaknesses. Fast, unenhanced MRI protocols will benefit our patients, improving their overall MRI experience and avoiding the potential risks of contrast media administration. The implementation of AI-based algorithms, using Deep Learning, Convolutional Neural Networks and Radiomics, will certainly increase the superiority of breast MRI and improve patient outcomes, as they can facilitate lesion differentiation, predict response to treatment, reduce unnecessary biopsies, and also reduce scan times and artefacts.