Development and external validation of a clinical prediction model using routine clinical indicators for imaging-defined intracranial complications in term infants with bacterial meningitis
Translational Pediatrics · 2026-11 · 15 (11) · 437-437
DOI: 10.21037/tp-2026-0763 ↗Kurz gefasst
Fachartikel in der Zeitschrift Translational Pediatrics, 2026-11. Verfasst von Ying Chen und 7 weiteren Autor:innen.
Abstract (Originalsprache)
Background
Neonatal bacterial meningitis (NBM) may cause imaging-defined intracranial complications detected during hospitalization. We aimed to develop and externally validate a model using routine early clinical indicators in term infants with NBM.
Methods
This retrospective two-center study included infants born at 37–42 weeks of gestation, symptom onset at ≤28 days of age, an NBM diagnosis, a hospital stay of ≥14 days, and qualifying cranial imaging from day 14 after symptom onset until discharge. The development cohort was assembled at the Capital Institute of Pediatrics (2013–2026), and the external validation cohort at the Seventh Medical Center of PLA General Hospital (2023–2026). Thirty routinely collected candidate predictors recorded during the first 3 days after symptom onset underwent stability-oriented selection using repeated cross-validation and bootstrap resampling. Logistic regression and five machine-learning algorithms were compared using common validation partitions. The composite outcome comprised prespecified intracranial abnormalities documented by cranial ultrasonography, computed tomography, or magnetic resonance imaging. Model discrimination was evaluated by the area under the receiver operating characteristic curve (AUROC); calibration, overall prediction error, and exploratory threshold-dependent performance were also assessed.
Results
The development cohort included 565 infants, of whom 138 (24.4%) had imaging-defined complications; the external cohort included 102 infants, of whom 28 (27.5%) had complications. The final logistic regression model included six predictors: cerebrospinal fluid (CSF) white blood cell count, altered muscle tone, hypotension requiring vasoactive support, gaze deviation, seizure, and C-reactive protein. Nested internal validation of the complete model-building procedure yielded an AUROC of 0.806, a Brier score of 0.137, and a calibration slope of 0.838. In external validation, the AUROC was 0.637 [95% confidence interval (CI), 0.510–0.762], the Brier score was 0.233 (95% CI, 0.175–0.297), and the calibration slope was 0.286 (95% CI, 0.033–0.561). At an illustrative 20% predicted-risk threshold (not selected for clinical use), external sensitivity was 60.7% (95% CI, 42.4–76.4%) and specificity was 54.1% (95% CI, 42.8–64.9%).
Conclusions
The six-predictor logistic regression model summarizes routinely obtainable early prognostic information, but independent external evaluation showed modest discrimination and substantial miscalibration. It remains a research model and should not guide individual clinical decisions without further validation and prospective evaluation.
Zitieren
APA
Chen, Y., Cheng, Q., Zhang, N., Jia, S., Wang, C., Li, Y., He, B., & Wu, J. (2026). Development and external validation of a clinical prediction model using routine clinical indicators for imaging-defined intracranial complications in term infants with bacterial meningitis. Translational Pediatrics, 15(11), 437-437. https://doi.org/10.21037/tp-2026-0763
Vancouver
Chen Y, Cheng Q, Zhang N, Jia S, Wang C, Li Y, et al. Development and external validation of a clinical prediction model using routine clinical indicators for imaging-defined intracranial complications in term infants with bacterial meningitis. Translational Pediatrics. 2026;15(11):437-437. doi:10.21037/tp-2026-0763
BibTeX
@article{Chen2026,
author = {Chen, Ying and Cheng, Qianyu and Zhang, Na and Jia, Shenghua and Wang, Chi and Li, Ying and He, Bingxi and Wu, Jing},
title = {Development and external validation of a clinical prediction model using routine clinical indicators for imaging-defined intracranial complications in term infants with bacterial meningitis},
journal = {Translational Pediatrics},
year = {2026},
volume = {15},
number = {11},
pages = {437--437},
doi = {10.21037/tp-2026-0763}
}
EUVIMED