LAUSR

AIM To evaluate the potential value of magnetic resonance imaging (MRI) for predicting postoperative pancreatic fistula (POPF) in patients with pancreatic cancer (PC) and non-pancreatic cancer (non-PC). MATERIAL AND METHODS This retrospective study was approved by the institutional review board and written informed consent was waived. Forty patients underwent pancreatoduodenectomy due to PC (n=31) and non-PC (n=9). The pancreas-to-muscle signal intensity ratio (SIR) on three-dimensional (3D)- fast field echo (FFE) T1-, in- and opposed-phase T1-, and T2-weighted images, as well as the apparent diffusion coefficient (ADC) value of the pancreas were measured. The frequency of POPF and MRI measurements were compared between patients with PC and non-PC. The MRI measurements were also compared with the grade of pancreatic fibrosis on pathological findings, fat deposition, and interstitial oedema. RESULTS The frequency of POPF was significantly higher in patients with non-PC than in those with PC (p=0.0067), with an odds ratio of 10.4. The SIR on 3D-FFE T1-weighted images was significantly higher in patients with non-PC (p=0.0001) and those with POPF (p=0.017) than in those with PC and those without POPF, respectively. Multiple regression analysis demonstrated that the SIR on 3D-FFE T1-weighted image was independently associated with the grade of pancreatic fibrosis (p<0.0001). CONCLUSION The frequency of POPF was significantly higher in patients with non-PC than in those with PC was inversely related to the grade of pancreatic fibrosis. The SIR on 3D-FFE T1-weighted image might be a potential imaging biomarker for predicting POPF.