Construction of a model based on multimodal computed tomography parameters for predicting post-thrombectomy cerebral edema in acute cerebral infarction
10.13491/j.issn.1004-714X.2026.02.005
- VernacularTitle:CT多模态参数预测急性脑梗死取栓后脑水肿的模型构建
- Author:
Manman LIANG
1
;
Peng XU
1
;
Yichao LIU
1
Author Information
1. Department of Imaging Center, The People's Hospital of Bozhou, Bozhou 236800, China.
- Publication Type:OriginalArticles
- Keywords:
Acute cerebral infarction;
Multimodal computed tomography parameters;
Mechanical thrombectomy;
Cerebral edema
- From:
Chinese Journal of Radiological Health
2026;35(2):187-192
- CountryChina
- Language:Chinese
-
Abstract:
Objective To establish a risk prediction model for cerebral edema after mechanical thrombectomy in acute cerebral infarction (ACI) based on multimodal computed tomography (CT) parameters. Methods A total of 220 ACI patients who underwent mechanical thrombectomy at The People’s Hospital of Bozhou between January 2022 and June 2025 were enrolled. All patients underwent preoperative conventional non-contrast CT, CT perfusion imaging, and multiphase CT angiography. Based on the presence or absence of cerebral edema at 72 hours postoperatively, patients were divided into cerebral edema and non-cerebral edema groups. General data and multimodal CT parameters were compared between the two groups. Multivariable logistic regression model was used to analyze the influencing factors of postoperative cerebral edema. A nomogram prediction model was constructed based on the multivariable analysis results, and the efficacy of the model was validated. Results The incidence of postoperative cerebral edema was 23.64% (52/220). Multivariable logistic regression analysis showed that time from onset to recanalization (OR=1.081, 95%CI=1.017-1.149), cerebral blood flow (OR=0.903, 95%CI=0.828-0.984), time to peak (OR=1.171, 95%CI=1.041-1.317), mean transit time (OR=2.815, 95%CI=1.823-6.178), and collateral circulation score (OR=0.960, 95%CI=0.930-0.990) were influencing factors for cerebral edema after mechanical thrombectomy in ACI patients (P<0.05). A nomogram model was constructed based on these factors. The calibration curve indicated good model fit, and the decision curve demonstrated net benefit. The area under the receiver operating characteristic curve for predicting cerebral edema after mechanical thrombectomy in patients with ACI was 0.961, with a sensitivity of 88.46% and a specificity of 91.07%. The Hosmer–Lemeshow goodness-of-fit test showed good model fit (χ2=1.766, P=0.962). Conclusion A risk prediction model for cerebral edema after mechanical thrombectomy in ACI patients was constructed using multimodal CT parameters. The model demonstrated high sensitivity and specificity.
