Development and validation of a preoperative predictive scoring system for esophageal perforation induced by fishbones

Fishbones are the most common esophageal foreign bodies, with perforation representing their most severe complication. However, reliable tools to preoperatively predict perforation risk are lacking. This study aimed to develop and validate a practical scoring system for this purpose. A multi-center retrospective study enrolled 1,248 patients with fishbone-related esophageal foreign bodies, randomly divided into training (n = 512), internal validation (n = 220), and external validation cohorts (n = 516). Univariate and multivariate logistic regression identified independent risk factors for perforation. A scoring system was constructed by weighting regression coefficients. Performance was assessed using the area under the curve (AUC), sensitivity, specificity, accuracy, and predictive values. Esophageal perforation occurred in 12.7% (159/1,248) of patients. Three independent predictors were identified: fishbone length ≥ 2.0 cm (2 points), lodgment duration ≥ 24 h (1 point), and forced swallowing (1 point). The system demonstrated strong discrimination, with AUCs of 0.863 (internal) and 0.829 (external). Risk stratification revealed perforation rates of 9.9% (low-risk: 0–2 points), 83.3% (intermediate-risk: 3 points), and 100.0% (high-risk: 4 points) in internal validation, and 8.6%, 63.6%, and 100.0% in external validation. This scoring system, incorporating fishbone length, lodgment duration, and forced swallowing, shows promising performance in predicting esophageal perforation risk and may assist in preoperative risk stratification. Further validation in broader populations is warranted to confirm its clinical utility.