Multiparametric Magnetic Resonance Imaging of Pancreas Disease

This project falls within the EPSRC Healthcare Technologies (Medical Imaging) research area

Collaboration and Supervision
This project will be carried out in collaboration with Perspectum Diagnostics Ltd. The project will be supervised by Prof. Daniel Bulte at the University of Oxford and Prof. Sir Michael Brady and Dr. Ged Ridgway at Perspectum Diagnostics Ltd.

Project Details
Much like chronic liver disease, chronic pancreas disease is on the rise together with the increasing prevalence of obesity and its associated conditions, such as the metabolic syndrome, cardiovascular disease and cancer. Obesity causes ectopic fat accumulation in organs such as the liver, heart or pancreas, and triggers a set of inflammatory responses that ultimately lead to fibrosis and increased risk of primary cancer. Fatty infiltration of organs such as the liver or the pancreas is initially asymptomatic and is usually reported incidentally as part of an abdominal imaging routine. Quantitative imaging becomes essential to the early detection and diagnosis of these conditions in order to monitor these subjects more closely. MRI provides safe, accurate and precise tools compared to other imaging techniques such as Computed Tomography (CT) or ultrasound.

Perspectum Diagnostics Ltd have developed a set of non-contrast, quantitative MRI protocols for the assessment of liver disease. Multiparametric imaging of the liver allows for the reliable quantification of fat, iron and fibro-inflammation in parenchymal tissue; these parameters predict clinical outcomes in patients with NAFLD, hemochromatosis, steatohepatitis or autoimmune hepatitis, to name a few, as well as alcohol-related liver disease. While these MRI protocols are not readily applicable to pancreas disease, they constitute excellent first candidates. We believe this for two main reasons, (1) the fact that the pancreas shows similar disease progression to the liver when exposed to fatty infiltration -the analogous term Non-Alcoholic Fatty Pancreas Disease (NAFPD) has been coined- and (2) similar conditions have been described in the pancreas, including pancreatic steatosis, iron overload, pancreatic fibrosis, chronic pancreatitis, pancreatic cancer, pancreatic cysts or pancreatic duct obstruction. More adventurous research has attempted to establish a link between certain pancreas conditions such as NAFPD and beta-cell dysfunction and the onset of type II diabetes mellitus (T2DM), with inconclusive results. In pancreatic transplantation, steatosis is currently assessed subjectively through visual inspection upon organ retrieval and is a frequent cause of organ discard.

The first year of this research project will focus on the technical MRI acquisition challenges and the validation of the Proton Density Fat Fraction (PDFF) metric for fat quantification in the pancreas. For this problem, certain modelling assumptions in liver fat assessment do not hold; also, the variable geometry and location of the pancreas make it a challenging organ to image without artefacts, compromising fat quantification. The second year of the DPhil will aim to investigate protocols for the segmentation of the pancreas in order to assess the spatial heterogeneity of pancreatic steatosis. Segmentation will help towards the assessment of pancreatic volume and surface irregularity, which have been linked to certain conditions such as T1DM and may themselves constitute useful imaging biomarkers tied with pancreas health state. The third year of the project is planned for the development of relaxometry protocols for the quantification of fibro-inflammation and its association with the risk of developing pancreatic cancer.

Nearly 20,000 UK Biobank volunteer acquisitions have become available to us, each with multiple pancreas MRI protocols. We are also aiming to acquire patient data within the course of the project through research collaborations.