Related Experiment Video
Updated: Jan 10, 2026
Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery
Published on: September 27, 2024
Development and validation of a nomogram based on preoperative contrast-enhanced computed tomography-derived
Yongkai Wang1, Zhiyuan Chen2, Hongpeng Chu1
1Department of Hepatobiliary Surgery, The First Affiliated Hospital of Wannan Medical College, Yijishan Hospital, Wuhu City, Anhui Province, China.
Background:
The primary goal of this study was to create a preoperative risk stratification tool for clinically relevant postoperative pancreatic fistula (CR-POPF) after pancreaticoduodenectomy (PD) using extracellular volume (ECV) data acquired from preoperative contrast-enhanced computed tomography scans.
Methods:
Patient records from PD procedures performed at 3 Chinese hospitals from January 2022 to December 2024 were retrospectively examined. Cohort stratification was based on CR-POPF occurrence. Preoperative risk factors independently associated with CR-POPF were identified using least absolute shrinkage and selection operator (LASSO) regression analysis augmented by logistic regression. A risk assessment model was developed incorporating these significant clinical and imaging predictors. The performance evaluation included discriminatory capacity (receiver operating characteristic curves), calibration (calibration plots), and net benefit estimation (decision curve analysis).
Results:
In the training and validation cohorts, CR-POPF incidences were 20.6% (32 patients) and 20.5% (26 patients), respectively. The LASSO logistic regression identified 4 independent predictors: body mass index, ECV, main pancreatic duct diameter, and pancreatic neck parenchymal thickness (short axis). The resulting graphical prediction tool showed excellent performance in external validation, achieving an area under the receiver operating characteristic curve of 0.883 (range, 0.806-0.960), outperforming 3 existing POPF risk models.
Conclusion:
Our findings suggest that the CR-POPF risk assessment construct, constructed based on the identified variables, can be reliably used for the preoperative assessment of fistula risk in patients scheduled for PD. This tool enables surgeons to anticipate, identify, and manage pancreatic fistulas more effectively. In addition, ECV, introduced as a novel and readily obtainable predictor, offers significant clinical value in assessing the likelihood of POPF after PD.
