Cross-sectional comparison of magnetic resonance imaging and ultrasound on the quantitative assessment of pancreatic fat

Purpose: Fat accumulation in the pancreas is associated with obesity and nonalcoholic fatty liver disease (NAFLD). It is also associated with pancreatic dysfunction or even cancer. Therefore, a reliable and noninvasive method for pancreatic fat quantification is indispensable. Materials and Methods: We retrospectively evaluated pancreatic steatosis by multiple magnetic resonance imaging (MRI) sequences and ultrasound (US) images from December 2015 to March 2018 of patients diagnosed with NAFLD. The multiple MRI sequences mentioned above include in/out of phase (IP/OP), apparent diffusion coefficient, and fat fraction (FF) images. The determination of a fatty pancreas based on ultrasound imaging depends on the pancreas’s higher echogenicity than the kidney. The correlation between US and MRI sequences was analyzed. Results: We reviewed the MR studies of 235 (192 males and 43 females) consecutive patients; 147 of them with abdominal ultrasound were recruited. The FF ranged from 2.91% to 39.69% on MR-FF images. This value of the ultrasound-negative group was significantly lower than that of the ultrasound-positive group (P = 0.004). The IP/OP ratio was 0.92–1.84. The values in the ultrasound-negative group were significantly lower than those in the ultrasound-positive group (P = 0.001). The MR-FF value was significantly correlated with the IP/OP ratio (P < 0.0001, R = 0.73). Conclusion: The ultrasound findings to determine fatty pancreas significantly corresponded to the FF on MR-FF and the IP/OP images. The ratio obtained from IP/OP images, a routine sequence in abdominal MR study, was also reliable for assessing fatty pancreatic disease.