Aims of the Project:
- To develop, implement and evaluate cardiac magnetic resonance fingerprinting (MRF) for simultaneous T1, T2 and T1rho tissue characterization at 0.55T Free.Max Siemens scanner
- To develop physics-informed deep-learning undersampled reconstruction and transfer. learning approaches from 1.5T to 0.55T to enable accelerated acquisitions and compensate for the low signal to noise at lower field.
- To extend the proposed approach to further characterize T2* and Fat Fraction (FF), incorporating deep learning based cardiac motion correction in the reconstruction framework.
- To evaluate the proposed approach in standardised phantoms, healthy subjects, and patients in comparison to standard references at 0.55T and current cardiac MRF approaches at 1.5T.
Cardiovascular disease (CVD) is the leading single cause of morbidity and mortality in the Western world. Quantitative cardiac magnetic resonance imaging (MRI) has emerged in recent years as an important non-invasive tool to evaluate a range of CVD conditions, showing great potential for objective characterisation of myocardial tissue properties. Several clinical studies have shown the potential of tissue specific MRI parameters such as T1, T1, T2 and T2* relaxation times and fat fraction (FF) to improve the assessment of CVD by characterising fibrosis, inflammation, iron overload and steatosis. However, multiparametric mapping requires sequential acquisitions under several breath-holds before and after contrast injection, often leading to non-registered maps, bias due to inter-parameter dependencies, as well as potential patient discomfort. Cardiac Magnetic Resonance Fingerprinting (MRF) has been proposed to rapidly and simultaneously quantify multiple tissue parameters including T1 and T2. More recently our group has extended this approach to simultaneous T1, T2 and T1ρ cardiac MRF and T1, T2, T2* and FF cardiac MRF. However, these approaches have been developed at conventional MRI field strengths of 1.5 Tesla (T) and 3T.
MRI relies on a powerful and heavy magnet, radio-frequency waves and magnetic gradient fields; and on different acquisition and reconstructions strategies to create the images. The first clinical MRI systems in the early 80’s had magnets of ~0.05–0.35T. Over the last 40 years, hardware advances have moved towards high (1.5-3T) and ultra-high field (up to 7T) MRI. While resulting in significant improvements, these developments meant that whole-body MRI remains expensive (~USD 1-1.3million/Tesla). However, advances in current state of the art MRI are not only explained by the higher magnetic field strength but also by a plethora of other remarkable developments in hardware and software. Recently, leading MRI vendors have taken full advantage of these cutting-edge developments to provide diagnostic image quality at lower field strengths (< 1T) while reducing not only the equipment costs but also the siting, maintenance and running requirements. In addition, lower-field MRI is safer and more comfortable for patients (especially paediatric, bariatric, and claustrophobic ones and those with implants) and clinical staff due to the weaker projectile forces, the reduced acoustic noise, the increased bore size (80cm) and the much smaller tissue heating effects. At KCL we are lucky to be the first centre in the UK to have access to a latest generation 0.55T scanner.
Thus, taking advantage of these novel methodologies and technologies, the objective of this project is to move the state of the art forward by developing novel methods to enable fully quantitative, affordable, accessible, and accurate myocardial tissue characterisation via the development of a single, fast, comprehensive exam operating on a lower-cost (therefore more affordable and accessible) scanner.