Survival prediction after upfront surgery in patients with pancreatic ductal adenocarcinoma: Radiomic, clinic-pathologic and body composition analysis.

摘要: Abstract Objective To investigate the value of radiomic features at contrast-enhanced CT integrated with clinic-pathologic and body composition measures for predicting survival after upfront surgery in patients pancreatic ductal adenocarcinoma (PDAC). Methods Two hundred ninety-nine PDAC who underwent surgical resection were included allocated to training set (210 patients) validation (89 patients). The radiomics signature was constructed by using least absolute shrinkage selection operator Cox regression. Multivariable regression analysis used construct a model based on signature, measures. A clinical without also developed. Model performance analyzed Harrell’s concordance index (C-index) time-independent receiver operating characteristic (ROC) analysis. Kaplan-Meier (KM) method Results Five independent variables selected model: carbohydrate antigen 19–9, skeletal muscle index, histologic grade postoperative chemotherapy. radiomics-based provided better predictive (C-index = 0.73; all p  Conclusion radiovdmics-based integrating could predict patients.