Etiology-based scoring for pediatric secondary intussusception: a retrospective analysis of clinical heterogeneity

Purpose: To investigate clinical heterogeneity in pediatric secondary intussusception and to develop four simplified, etiology-specific scoring systems to facilitate preoperative etiologic prioritization after intussusception has been confirmed. Methods: This retrospective study analyzed 92 pediatric patients diagnosed with secondary intussusception from 2018 to 2023. Disease patterns across PLP subtypes were analyzed using a four-dimensional framework including etiology, age, sex, and clinical features. Candidate variables were selected based on clinical plausibility and univariable screening before being entered into etiology-specific OvR models. Etiology-specific scoring models for IgA vasculitis (IgAV), Meckel's diverticulum (MD), intestinal polyps (IP), and intestinal duplication (ID) were constructed using binary logistic regression and a one-vs.-rest strategy. Each score was internally validated according to ROC curves, with discrimination assessed by AUC as well as sensitivity and specificity. Results: The IgAV score incorporated hematochezia (+2), age ≥6 years (+1), and absence of abdominal mass (+1), with an AUC of 0.85, sensitivity of 80%, and specificity of 75.4%. The AUCs for Meckel's diverticulum, intestinal polyps, and intestinal duplication were 0.780, 0.925, and 0.851, respectively. The remaining etiology-specific scores similarly comprised two to three bedside-available variables (e.g., demographic features and key clinical manifestations), allowing for practical use in the preoperative setting. Conclusion: These simple, non-invasive scoring systems may assist early etiologic prioritization and support more targeted diagnostic assessment in pediatric intussusception. Their simplicity and internal performance characteristics suggest potential utility in acute care settings. Future studies are warranted to validate their generalizability and to explore integration into routine clinical workflows.