Quantification of pancreas fat on dual-energy computed tomography: comparison with six-point Dixon magnetic resonance imaging

Although it is important to quantify the degree of fatty degeneration of the pancreas, it is difficult to make such a quantification using conventional computed tomography (CT). The present study evaluated the feasibility of pancreatic fat quantification by dual-energy CT (DECT) compared with T2*-corrected six-point Dixon magnetic resonance imaging (MRI). Twenty-eight patients who underwent both DECT (100 and 150 kVp) and Dixon MRI without the use of contrast agents were analyzed. The region of interest (ROI) was placed at the head and body/tail of the pancreas on fat volume fraction (FVF) maps generated using the multi-material decomposition (MMD) algorithm on DECT. The FVF (%) of pancreatic parenchyma measured by DECT (CT-FVF) was compared with that measured on FVF maps calculated using Dixon MRI (MR-FVF) using the Spearman rank correlation coefficient. The median CT-FVF (%) values of the head and body/tail of the pancreas on DECT were 14.2% (range 0.1–81.2%) and 9.4% (range 0–40.8%), respectively. The median MR-FVF (%) values of the head and body/tail of the pancreas on Dixon MRI were 12.2% (range 1.2–80.9%) and 8.1% (range 0.3–43.7%), respectively. CT-FVF (%) measured by DECT showed a significant correlation with the MR-FVF (%) measured by Dixon MRI in the head of the pancreas (ρ = 0.631, P