Use of Gadoxetic Acid-enhanced MRI to Predict the Development of Postoperative Pancreatic Fistulas by Estimating the Degree of Pancreatic Fibrosis

Abstract

Background: Post-operative pancreatic fistula (POPF) can be life-threatening, and gadoxetic acid-enhanced MRI is routinely performed in patients undergoing pancreatic surgery. However, previous reports have not investigated if gadoxetic acid-enhanced MRI can be used to predict POPF risk. Objective: This study aims to explore if gadoxetic acid-enhanced MRI can predict pancreatic fibrosis and the need for POPF treatment before surgery. Methods: We retrospectively analyzed gadoxetic acid-enhanced MR images from 142 patients who underwent pancreatic surgery between January 1, 2011, and April 30, 2018. Pre-dynamic signal intensity (SI) and values for the portal, transitional, and hepatobiliary phase standardized based on pre-dynamic study values were analyzed. The diameter of the main pancreatic duct (DMPD) was measured, and the degree of pancreatic fibrosis was classified as F0 – F3. We defined POPF higher than grade B as significant. Results: Odds ratios for combinations that led to any degree of fibrosis higher than grade B were defined as significant risk factors. The highest odds ratio was obtained for F0 vs. F1 – F3 (p = 0.038). DMPD (p<0.001), pre-SI (p = 0.008), portal-SI/pre-SI (p<0.001), transitional-SI/pre-SI (p<0.001), and hepatobiliary-SI/pre-SI (p = 0.012) were significantly correlated with the presence of fibrosis. Moreover, the presence of fibrosis was best detected by DMPD (AUC = 0.777). Individual specificity values of transitional-SI/pre-SI and DMPD were 95.5% and 86.6%, respectively, and their combined specificity was 97.7%. Conclusion: The absence of pancreatic fibrosis is a risk factor for developing POPF higher than grade B. DMPD was the most useful diagnostic indicator for the presence of fibrosis among our analysis, and its specificity increased when combined with transitional-SI/pre-SI.