Compressed Quantitative MRI
Lead Research Organisation:
University of Edinburgh
Department Name: Sch of Engineering
Abstract
The proposed research will provide the first proof-of-principle for a new family of Compressed Quantitative Magnetic Resonance Imaging (CQ-MRI), able to rapidly acquire a multitude of physical parameter maps for the imaged tissue from a single scan.
MRI is the pre-eminent imaging modality in clinical medicine and neuroscience, providing valuable anatomical and diagnostic information. However, the vast majority of MR imaging is essentially qualitative in nature providing a `picture' of the tissue while not directly measuring its physical parameters. In contrast, quantitative MRI aims to measure properties that are intrinsic to the tissue type and independent of the scanner and scanning protocol. Unfortunately, due to excessively long scan times, Quantitative MRI is not usually included in standard protocols.
The proposed research is based on a combination of a new acquisition philosophy for Quantitative MRI, called Magnetic Resonance Fingerprinting, and recent advances in model-based compressed sensing theory to enable rapid simultaneous acquisition of the multiple parameter maps.
The ultimate goal of the research will be to produce a full CQ-MRI scan capability with a scan time not substantially longer than is currently needed for a standard MRI scan.
MRI is the pre-eminent imaging modality in clinical medicine and neuroscience, providing valuable anatomical and diagnostic information. However, the vast majority of MR imaging is essentially qualitative in nature providing a `picture' of the tissue while not directly measuring its physical parameters. In contrast, quantitative MRI aims to measure properties that are intrinsic to the tissue type and independent of the scanner and scanning protocol. Unfortunately, due to excessively long scan times, Quantitative MRI is not usually included in standard protocols.
The proposed research is based on a combination of a new acquisition philosophy for Quantitative MRI, called Magnetic Resonance Fingerprinting, and recent advances in model-based compressed sensing theory to enable rapid simultaneous acquisition of the multiple parameter maps.
The ultimate goal of the research will be to produce a full CQ-MRI scan capability with a scan time not substantially longer than is currently needed for a standard MRI scan.
Planned Impact
The proposed research will develop a radically new MRI technique that promises rapid full quantitative MR imaging. Such a technology will help to address key national and global health challenges such as the diagnosis and treatment of dementia, strokes and other neurodegenerative diseases.
Quantitative imaging acquires tissue specific MR properties that provide invaluable information for tissue or pathology identification. Full quantitative imaging enables the simultaneous acquisition of all these parameters together and has the potential to provide clinicians and neuroscientists with a completely new tool for research and diagnosis giving better sensitivity and specificity. It further offers the opportunity for a single universal scan from which standard 'virtual scans' can be generated in software.
Fast scan times will allow such a technology to be routinely available and would help facilitate the quantitative measurement of diseases. Fast scan times will further reduce motion artefacts, help increase throughput, decrease waiting times and hence reduce total imaging costs.
More broadly the research philosophy forms the basis of a new generation of compressed sensing imaging techniques that fundamentally exploit the underlying physics of the imaging system to enhance sensing and imaging. This is expected to have impact in a very wide range of applications, from seismology to chemical sensing.
Quantitative imaging acquires tissue specific MR properties that provide invaluable information for tissue or pathology identification. Full quantitative imaging enables the simultaneous acquisition of all these parameters together and has the potential to provide clinicians and neuroscientists with a completely new tool for research and diagnosis giving better sensitivity and specificity. It further offers the opportunity for a single universal scan from which standard 'virtual scans' can be generated in software.
Fast scan times will allow such a technology to be routinely available and would help facilitate the quantitative measurement of diseases. Fast scan times will further reduce motion artefacts, help increase throughput, decrease waiting times and hence reduce total imaging costs.
More broadly the research philosophy forms the basis of a new generation of compressed sensing imaging techniques that fundamentally exploit the underlying physics of the imaging system to enhance sensing and imaging. This is expected to have impact in a very wide range of applications, from seismology to chemical sensing.
Publications
Bano W
(2020)
Model-based super-resolution reconstruction of T2 maps.
in Magnetic resonance in medicine
Benjamin AJV
(2020)
Diagnostic quality assessment of IR-prepared 3D magnetic resonance neuroimaging accelerated using compressed sensing and k-space sampling order optimization.
in Magnetic resonance imaging
Benjamin AJV
(2019)
Multi-shot Echo Planar Imaging for accelerated Cartesian MR Fingerprinting: An alternative to conventional spiral MR Fingerprinting.
in Magnetic resonance imaging
Duarte R
(2020)
Greedy approximate projection for magnetic resonance fingerprinting with partial volumes
in Inverse Problems
Golbabaee M
(2019)
CoverBLIP: accelerated and scalable iterative matched-filtering for magnetic resonance fingerprint reconstruction*
in Inverse Problems
Golbabaee M
(2021)
Compressive MRI quantification using convex spatiotemporal priors and deep encoder-decoder networks.
in Medical image analysis
Golbabaee M
(2019)
Geometry of Deep Learning for Magnetic Resonance Fingerprinting
Golbabaee M
(2018)
Inexact Gradient Projection and Fast Data Driven Compressed Sensing
in IEEE Transactions on Information Theory
Golbabaee M.
(2017)
Cover Tree Compressed Sensing
Golbabaee M.
(2017)
Cover tree compressed sensing for fast MR fingerprint recovery
Golbabaee Mohammad
(2017)
Inexact Gradient Projection and Fast Data Driven Compressed Sensing
in arXiv e-prints
Iwaya LH
(2023)
On the privacy of mental health apps: An empirical investigation and its implications for app development.
in Empirical software engineering
Tang J,
(2017)
Exploiting the Structure via Sketched Gradient Algorithms.
Tang Junqi
(2017)
Structure-Adaptive, Variance-Reduced, and Accelerated Stochastic Optimization
in arXiv e-prints
| Description | We have shown that compressed sensing techniques can be applied successfully to perform rapid multi-parameter quantitative MR imaging. |
| Exploitation Route | We engaged with GE who are looking at various aspects of quantitative imaging technologies |
| Sectors | Digital/Communication/Information Technologies (including Software),Healthcare |
| Description | Deep compressive quantitative MRI imaging |
| Amount | £268,932 (GBP) |
| Funding ID | EP/X001091/1 |
| Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
| Sector | Public |
| Country | United Kingdom |
| Start | 02/2023 |
| End | 01/2025 |
| Description | GE Global Research Munich |
| Organisation | General Electric |
| Department | GE Global Research HQ |
| Country | United States |
| Sector | Private |
| PI Contribution | We have collaborated with GE global research in Munich on our development of compressed sensing MRF algorithms. The RA, Mohammad Golbabaee visited GE during the project and was able to carry out tests on their MRI scanners |
| Collaborator Contribution | GE contributed to the research and hosted the RA visit |
| Impact | joint publications |
| Start Year | 2018 |
| Description | Superresolution Collaboration with Seimens |
| Organisation | Siemens Healthcare |
| Country | Germany |
| Sector | Private |
| PI Contribution | Working with the Advanced Clinical Imaging Technology, group in Siemens Healthcare, based at EPFL in Lausanne we developed a new imaging protocol to super resolve quantitative T2 maps. |
| Collaborator Contribution | The work began while RA Wajiha Bano was on secondment to Siemens in 2018 and resulted in a publication in Magnetic Resonance in Medicine this year. |
| Impact | The work was presented in the MRM paper: Model-Based Super-Resolution Reconstruction of T2 Maps, Bano, W., Piredda, G. F., Davies, M., Marshall, I., Golbabaee, M., Kober, T., Thiran, J-P., Meuli, R. & Hilbert, T. In Magnetic Resonance in Medicine, 2019. |
| Start Year | 2018 |
| Description | Keynote talk in iTWIST workshop 2018 |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | keynote talk on compressed sensing MRI |
| Year(s) Of Engagement Activity | 2018 |
| URL | https://sites.google.com/view/itwist18 |