Efficiency evaluation of dual-energy CT to predict the postoperative early recurrence of pancreatic ductal adenocarcinoma

Abstract

Objectives: To evaluate the efficacy of quantitative parameters from dual-energy CT (DECT) and basic CT features in predicting the postoperative early recurrence (ER) of pancreatic ductal adenocarcinoma (PDAC).

Methods: In this study, patients with PDAC who underwent radical resection and DECT from 2018 to 2022 were enrolled and categorised into ER and non-ER groups. The clinical data, basic CT features and DECT parameters of all patients were analyzed. Independent predictors of ER were identified with Logistic regression analyses. Three models (model A: basic CT features; model B: DECT parameters; model C: basic CT features + DECT parameters) were established. Receiver operating characteristic curve analysis was utilized to evaluate predictive performance.

Results: A total of 150 patients were enrolled (ER group: n = 63; non-ER group: n = 87). Rim enhancement (odds ratio [OR], 3.32), peripancreatic strands appearance (OR, 2.68), electron density in the pancreatic parenchymal phase (P-Rho; OR, 0.90), arterial enhancement fraction (AEF; OR, 0.05) and pancreatic parenchyma fat fraction in the delayed phase (OR, 1.25) were identified as independent predictors of ER. Model C showed the highest area under the curve of 0.898. In addition, the corresponding ER risk factors were identified separately for resectable and borderline resectable PDAC subgroups.

Conclusions: DECT quantitative parameters allow for the noninvasive prediction of postoperative ER in patients with PDAC, and the combination of DECT parameters and basic CT features shows a high prediction efficiency. Our model can help to identify patients with high-risk factors to guide preoperative decision making.

Keywords: Computed tomography; Logistic models; Pancreatic neoplasms; Recurrence.