Developing a predictive model for delayed healing of esophagojejunal anastomotic fistula following total gastrectomy based on imaging and clinical inflammatory-nutritional status.

Abstract

Background: Esophagojejunal anastomotic fistula (EJF) following radical total gastrectomy is a severe perioperative complication in patients with gastric cancer, particularly as delayed fistula healing increases hospitalization costs and leads to poor prognosis. Numerous factors influence the occurrence and progression of EJF, with inflammation and nutritional status being significant contributors to perioperative complications. Therefore, this study aims to investigate the prediction of delayed EJF healing based on postoperative clinical and imaging-related inflammation-nutrition status.

Methods: We retrospectively collected data on 315 cases of EJF following radical total gastrectomy for gastric cancer from two centers between 2015 and 2023 (training group: center one with 194 cases, validation group: center two with 121 cases). EJF was diagnosed based on clinical presentation, gastrointestinal imaging, or endoscopic findings. The healing time for EJF was defined as the period from diagnosis to the removal of the abdominal drainage tube, and patients were categorized into early healing and delayed healing groups based on the median healing time. Postoperative abdominal computed tomography(CT) scans and clinical characteristics at the time of EJF diagnosis were collected. Univariate and multivariable logistic regression analyses were performed on the training group data to construct a predictive model (nomogram). The model's performance in both the training and validation groups was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC), calibration curves, and decision curve analysis (DCA).

Result: The mean healing time for EJF was 16 ± 7 days (median time: 12 days, range: 4-43 days). Postoperative systemic immune-inflammation index (SII) > 521×10ˆ9/L, controlling nutritional status score (CONUT) > 4, nutritional support method, visceral fat index (VFI) < 74.42 cm²/m², and skeletal muscle index (SMI) < 41.25 cm²/m² were associated with delayed EJF healing times. A comprehensive model was developed, in the validation group, the model demonstrated an AUC of 0.838 (95% confidence interval (95% CI): 0.763-0.912). The DCA and calibration curves indicated a strong predictive consistency and clinical utility of the model.