A modified alternative fistula risk score (a-FRS) obtained from the computed tomography enhancement pattern of the pancreatic parenchyma predicts pancreatic fistula after pancreatoduodenectomy

Background: Alternative fistula risk score (a-FRS) is useful to predict clinically relevant postoperative pancreatic fistula (CR-POPF) after pancreatoduodenectomy (PD). Methods: Clinical data from 239 patients undergoing PD were collected. The CT value of the pancreatic parenchyma was measured in the nonenhanced (N), arterial (A), portal venous (P), and late (L) phases. The A/N, A/P, P/L and A/L ratios were calculated and their correlation with CR-POPF were analyzed. By replacing pancreatic texture with the best CT attenuation ratio, a modified a-FRS was developed. Results: Forty-seven patients developed CR-POPF. The A/P ratio (P < 0.001), P/L ratio (P = 0.002) and A/L ratio (P < 0.001) were significantly higher in the CR-POPF group. The A/L ratio performed best in predicting CR-POPF (AUC: 0.803) and the cut-off value is 1.36. A/L ratio >1.36 (P < 0.001), body mass index (P = 0.005) and duct diameter (P = 0.037) were independently associated with CR-POPF. By replacing soft texture with an A/L ratio >1.36, a modified a-FRS was developed and performed better than the a-FRS (AUC: 0.823 vs 0.748, P = 0.006) in predicting CR-POPF. Conclusions: The modified a-FRS is an objective and preoperative model for predicting the occurrence of CR-POPF after PD.