Towards 10 Minute Magnetic Resonance Imaging scans in children with machine learning

点击翻译按钮获取中文摘要

1. 1. BackgroundMagnetic Resonance Imaging (MRI) has enabled significant advances in the diagnosis and management of many diseases. However, MRI is challenging in the pediatric population as it is time consuming (~1 hour to perform) and requires patient cooperation. Hence it is often necessary to use general anesthesia (GA) in children below 8 years of age, which is both costly and carries some risk. One way of overcoming these problems would be to speed up the MRI scans so children do not have to keep still or hold their breath. The simplest way of doing this is to collect less data for each image, however, this results in artefacts that make the images unusable. Current reconstruction methods for removing these artefacts, allow limited acceleration, or use time consuming algorithms which hamper their clinical uptake. A new approach is Machine Learning (ML) that aims to 'learn' how to remove undersampling and motion artefacts. The purpose of this project is to develop fast MRI acquisitions, with rapid Machine Learning reconstruction technologies for use in childhood diseases of the abdomen.Building on work in Super Resolution and Deep Artefact Suppression, this project aims to reframe the reconstruction of MRI data, to remove aliases caused by data undersampling and motion corruption, as an image de-noising problem that can be initially performed by a CNN. This strategy requires specific sampling patterns that produce noise-like aliasing and large amounts of high quality, application-specific training data. Thus, this work will leverage the extremely large amounts image data that is available at Great Ormond Street Hospital (GOSH). 2. 2. Research Aims and ObjectivesThis study aims to develop novel accelerated MRI technologies which will allow scan times to be reduced from ~1 hour to ~10 minutes in children with diseases within the abdomen. This will be achieved this through development of optimised MR acquisition strategies combined with ML reconstruction techniques. Specific work packages include development of techniques to correct for artefacts caused by respiratory motion, development of fast 3D imaging for assessment of small bowel motility in Chron's disease, as well as development of real-time imaging to assess the filling and emptying function of the stomach. The resulting networks will be integrated into standard clinical workflow to enable clinical validation studies, as well as simple translation into routine clinical practise.3. 3. Novelty of Research MethodologyA few recent papers have shown the benefit of using ML methods for reconstruction of MR images. These have mostly been in knee or cardiac applications, and very few have been in abdominal MR imaging or in paediatrics. 4. 4. Alignment to EPSRC's strategies and research areasThis research aligns with the EPSRC Healthcare Technologies theme. The work falls within the "Optimising Treatment" grand challenge and is thoroughly aligned with the "Novel imaging technologies" cross-cutting research capabilities area. The proposed technologies satisfy the key fields; techniques for image reconstruction, lower cost image acquisition technologies, high throughput, and automated image interpretation. The work is also aligned with the EPSRC Research area, "Medical imaging". In particular, it addresses the high priority areas of this delivery plan; enabling earlier and more effective diagnosis, and novel imaging technologies that offer a significant benefit over current technologies. 5. 5. Any companies or collaborators involvedThe National Institute for Health Research Biomedical Research Centre (BRC) is partly funding this PhD studentship. Siemens Healthcare have pledged their support, as Project Partner, for Jennifer Steeden's UKRI grant which is closely related to this work.Great Ormond Street Hospital (GOSH) will be close collaborators in this project, providing access to MRI data and MRI scanning facilities.