Dual-Phase Computed Tomography for Assessment of Pancreatic Fibrosis and Anastomotic Failure Risk Following Pancreatoduodenectomy

Introduction Delayed or decreased computed tomography (CT) enhancement characteristics in pancreatic fibrosis have been described. Methods A review of 157 consecutive patients with preoperative dual-phase CT between 2004 and 2009 was performed. Pancreatic CT attenuation upstream from the tumor was measured in the pancreatic and hepatic imaging phases. The ratio of the mean CT attenuation value [hepatic to pancreatic phase; late/early (L/E) ratio] and histological grade of pancreatic fibrosis was correlated with the development of a clinically relevant pancreatic anastomotic failure (PAF) and other clinical parameters. Results A clinically relevant PAF was observed in 21 patients (13.4%) with morbidity and mortality of 39.5% and 0%, respectively. The PAF group showed maximum enhancement in the pancreatic and washout in the hepatic CT phase, while the no PAF group showed a delayed enhancement pattern. Degree of pancreatic fibrosis and L/E ratio were significantly lower for the PAF group than the no PAF group (0.86+/-0.14 vs. 1.09+/-0.24; P<0.0001 and 21.0+/-17.9 vs. 40.4+/-29.8; P<0.0001); fewer PAF patients showed an atrophic histological pattern (14% vs. 39%; P=0.046). The L/E ratio was positively correlated with pancreatic fibrosis. Pancreatic fibrosis and L/E ratio increased with larger duct size (P<0.001), the presence of diabetes (P<0.05), and the surgeon's assessment of pancreas firmness (P<0.001). In multivariate analyses, L/E ratio and body mass index were significant predictors for the development of a clinically relevant PAF; a 0.1-U increase of L/E ratio decreased the odds of a PAF by 54%. Conclusion Pancreatic CT enhancement pattern can accurately assess pancreatic fibrosis and is a powerful tool to predict the risk of developing a clinically relevant PAF following PD.