Development and validation of a predictive model for chronic or persistent immune thrombocytopenia in children incorporating anti-glycoprotein IIb antibody: a retrospective cohort study utilizing LASSO regression and bootstrap stability analysis

Description

Ximei Yang and colleagues published a 2026 study building a prediction model for which children with immune thrombocytopenia will develop chronic disease. In this autoimmune condition the body destroys its own platelets, and while most childhood cases resolve, a fifth to a third persist. The team reviewed 381 newly diagnosed children treated from 2022 to 2024, splitting them into training and test groups, and screened candidate predictors with methods that favor stable variables.

Four factors survived: age, recent infection or vaccination, platelet count, and antibody status against platelet protein GPIIb. The model reached an AUC of 0.743 on the test set, with its strength lying in reassurance; children classed as low risk had a 94.7 percent chance of avoiding chronic disease. Calibration was good and decision analysis showed net benefit across many thresholds. The authors present the tool as an early rule-out aid, not a firm forecast.

Catalog record

Source: https://europepmc.org/articles/PMC13279312