Obesity, metabolic disease and the pancreas-Quantitative imaging of pancreatic fat

Abstract

The association between pancreatic fat, obesity and metabolic disease is well-documented, and although a potentially exciting target for novel therapies, remains poorly understood. Non-invasive quantitative imaging-derived biomarkers can provide insights into pathophysiology and potentially provide robust trial endpoints for development of new treatments. In this review, we provide an overview of the pathophysiology of non-alcoholic fatty pancreas disease and associations with metabolic factors, obesity and diabetes. We then explore approaches to pancreatic fat quantification using ultrasound, CT and MRI, reviewing the strengths, limitations and current published evidence in the assessment of pancreatic fat. Finally, we explore the broader challenges of pancreatic fat quantification as we move toward translating these methods into the clinical setting.

Figure 1.

Pathophysiological mechanisms for non-alcoholic fatty pancreas disease. Intracellular accumulation of lipid is thought to precede the rise in pancreatic tissue adipocytes. Subsequent oxidative stress leads to the release of adipokines and cytokines which trigger an inflammatory response that is ultimately thought to drive beta-cell dysfunction leading to Type 2 diabetes mellitus. MPO, myeloperoxidase; TNF-alpha, tumour necrosis factor-alpha.

Figure 2.

Complete fatty replacement of the pancreas. Extreme pancreatic steatosis and atrophy – the residual pancreas with presumed borders outlined (dashed white line). The main pancreatic duct is just visible with only minimal lobular tissue.

Figure 3.

MRS and the differing spectral complexity of pancreatic fat MRS study planning (a) of the liver (i) and pancreas (ii), with corresponding spectra arising from the liver voxel (b) and pancreatic voxel (c). Note differing voxel sizes and the challenge of ensuring inclusion of purely pancreatic tissue within the MRS voxel (a, ii). The water peak is shown in both spectra (b, c) at 4.7 ppm (solid white arrow). Liver fat spectra (b) have been shown to demonstrate six peaks (at 5.3, 4.2, 2.75, 2.1, 1.3 and 0.9 ppm), while pancreatic fat spectra (c) are dominated by single 1.3 ppm methylene peak (clear white arrow). It is not known if this reflects genuine pancreatic fat content or contamination from extra lobular fat. MRS, magnetic resonance spectroscopy.

Figure 4.

Chemical-shift imaging and pancreatic fat quantification in the presence of fatty infiltration in-phase (a) and out-of-phase (b) CSI in the presence of significant pancreatic fatty infiltration. Note the "india-ink" artefact at fat-water interfaces on out-of-phase images (b), which when combined with in-phase imaging makes quantification in small islands of tissue prone to partial voluming errors. CSI, chemical-shift imaging.

Figure 5.

Proton density fat fraction imaging images derived from water-only and fat-only are used to generate a parametric fat fraction map across an axial slice. Note low pancreatic intralobular fat fractions (despite fatty infiltration, white-dashed outline). Errors arising from intraluminal (solid black arrows) and peripheral gas (solid white arrows) result in incoherent pixelation.

Figure 6.

Pancreatic fat segmentation as the pancreas undergoes both fatty infiltration and atrophy, the significance of fat changes (shaded areas) in the intralobular (a), interlobular (b) and extralobular/peripancreatic (c) fat compartments are unknown. Inconsistencies in the way in which these have been reported/evaluated in the literature are a barrier to pooled meta-analyses and robust segmentation methods for quantifying each of these are an important challenge.