Preoperative integration of spectral CT-derived extracellular volume and skeletal muscle index for disease-free survival prediction in adenocarcinoma of the esophagogastric junction

To develop and internally validate a preoperative model integrating spectral computed tomography (CT)-derived extracellular volume (ECV) fraction and skeletal muscle index (SMI) for predicting disease-free survival (DFS) after curative resection of adenocarcinoma of the esophagogastric junction (AEG), beyond conventional clinical factors. A total of 139 patients with AEG were included in this retrospective study, of whom 65 patients experienced DFS events and 74 patients did not experience DFS events. Clinical factors, body composition, and spectral CT parameters were analyzed. Cox regression identified independent predictors for DFS. Three nested models were constructed and assessed by time dependent ROC, calibration curves, and decision curve analysis. Subgroup analyses by body composition and ECV were performed. CEA > 5 ng/mL (HR = 3.31, p < 0.001), cT3-4 stage (HR = 2.63, p = 0.002), low SMI (HR = 2.18, p = 0.006) and high ECV grade (HR = 1.98, p = 0.026) were independent predictors of worse DFS. The AUCs of Model 1, Model 2, and Model 3 were 0.715, 0.747, 0.762 at 1 year; 0.738, 0.759, 0.773 at 2 years and 0.768, 0.810, 0.818 at 3 years. Calibration curves showed acceptable agreement between predicted and observed DFS probabilities. In exploratory subgroup analysis, patients with high ECV grade and low SMI showed the worst DFS prognosis. A preoperative model integrating CEA, cT stage, SMI, and ECV improved postoperative DFS prediction in patients with AEG. The combination of low SMI and high ECV identified the subgroup with the poorest DFS and may support risk-adapted perioperative assessment and postoperative surveillance.