%0 Journal Article
%T Predictive factors influencing outcome of early cranioplasty after decompressive craniectomy: a outcome prediction model study.
%A Yan Z
%A Li X
%A Xia B
%A Xue C
%A Wang Y
%A Che H
%A Shen D
%A Guo S
%J Front Neurol
%V 15
%N 0
%D 2024
%M 38903167
%F 4.086
%R 10.3389/fneur.2024.1384320
%X <B>UNASSIGNED: </B>The timing of cranioplasty (CP) has become a widely debated topic in research, there is currently no unified standard. To this end, we established a outcome prediction model to explore the factors influencing the outcome of early CP. Our aim is to provide theoretical and practical basis for whether patients with skull defects after decompressive craniectomy (DC) are suitable for early CP.<BR><B>UNASSIGNED: </B>A total of 90 patients with early CP after DC from January 2020 to December 2021 were retrospectively collected as the training group, and another 52 patients with early CP after DC from January 2022 to March 2023 were collected as the validation group. The Nomogram was established to explore the predictive factors that affect the outcome of early CP by Least absolute shrinkage analysis and selection operator (LASSO) regression and Logistic regression analysis. Receiver operating characteristic (ROC) curve was used to evaluate the discrimination of the prediction model. Calibration curve was used to evaluate the accuracy of data fitting, and decision curve analysis (DCA) diagram was used to evaluate the benefit of using the model.<BR><B>UNASSIGNED: </B>Age, preoperative GCS, preoperative NIHSS, defect area, and interval time from DC to CP were the predictors of the risk prediction model of early CP in patients with skull defects. The area under ROC curve (AUC) of the training group was 0.924 (95%CI: 0.867-0.980), and the AUC of the validation group was 0.918 (95%CI, 0.842-0.993). Hosmer-Lemeshow fit test showed that the mean absolute error was small, and the fit degree was good. The probability threshold of decision risk curve was wide and had practical value.<BR><B>UNASSIGNED: </B>The prediction model that considers the age, preoperative GCS, preoperative NIHSS, defect area, and interval time from DC has good predictive ability.