Back

Development and Validation of an MGBP Nomogram for Functional Prognosis Prediction of Pediatric Intracerebral Hemorrhage

ZHOU, J.-h.; Zhang, Z.-Y.; Chen, Y.; Ming, M.-X.; Shen, Q.-L.; Zheng, J.-C.; Pan, G.; Zhang, Y.; Chen, W.-M.; Lu, G.-P.

2023-02-07 pediatrics
10.1101/2023.02.03.23285464 medRxiv
Show abstract

BackgroundPediatric intracerebral hemorrhage (pICH) remains a considerable cause of mortality. This study aimed to explore functional prognosis predictors in Chinese patients with pICH and attempted to develop and validate a nomogram for estimating individual risk probability of short-term unfavorable functional prognosis. MethodsA retrospective case-control study through records reviewing was conducted, non-neonate patients with nontraumatic pICH discharged from the Childrens Hospital of Fudan University between January 2012 and December 2021 were all originally included. The primary outcome was unfavorable functional prognosis at 3 months post pICH defined as a score of 3-6 as measured using the modified Rankin Scale (ranging from 0 [asymptomatic] to 6 [death]). Multivariate logistic regression analysis was performed to screen prognosis predictors, a prediction model-based nomogram was developed. Internal validation was assessed and quantified as receiver operating characteristic (ROC) curve and bootstrapped calibration curve. ResultsA total of 269 pICH patients were enrolled, the median age was 57.2 months (interquartile range, 5.4-115.1), and 157 (58.4%) patients were male. The median follow-up time was 3.2 months (interquartile range, 2.8-3.6), 112 (41.6%) patients had unfavorable functional outcome. Cerebral vascular malformation as etiology, modified Glasgow Coma Score on admission, brainstem location, and intracerebral hemorrhage volume as percentage of total brain volume were identified as functional prognosis predictors. A nomogram was established comprising these four predictors, abbreviated to MGBP, the area under the ROC curve of the nomogram was 0.827 (95% Confidence Interval, 0.778-0.877) with good calibration (P=0.803 for the Spiegelhalters Z-test). ConclusionsThe MGBP nomogram is the first model developed and validated in a consecutive cohort to predict 3-month unfavorable functional prognosis post nontraumatic pICH in Chinese children, which may provide clinicians with a potentially effective approach for early prediction and timely management of pICH.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.