Invited Lecture 3.2

Abstract

Current MRI clinical practice is almost always qualitative, with the interpretation relying on contrast changes in weighted images. This has associated limitations, in particular in the context of standardisation and reproducibility. Qualitative conventional MRI also suffers from poor integration into multimodal modern medicine, in its trajectory moving towards precision and personalised diagnosis and treatment planning.In quantitative MRI (qMRI) techniques, the interpretation of the images can rely on the intensities having physical units and providing estimations of a physical or physiological parameter. The widespread adoption of qMRI is often referred to as a desirable paradigm change, since qMRI could drastically improve the reproducibility of MRI measurements and harness the power of longitudinal scans, clinical trials and multicentre studies.  Nowadays there are only a few qMRI applications, with limited clinical availability, since realising its full benefits comes with technical and translational challenges. On the technical side, recent technical developments in MRI are very promising for qMRI development, but conscious efforts from all the stakeholders involved (manufacturers, researchers and clinical professionals) will be required to bring qMRI to the clinic. Advances in AI acceleration techniques have driven an interest in standardisation and optimisation of MRI scanners outputs, but these are time consuming problems that might not even be solvable under the current framework of qualitative MRI. In this talk, I will discuss the current state of affairs of conventional MRI and its limitations, as well as outlining the potential benefits of qMRI in the clinic, reviewing some of the most promising applications (with an emphasis in relaxometry, diffusion and spectroscopy) and the barriers standing on their way