A prognostic model for interventional thrombectomy in patients with acute ischemic stroke based on a BP neural network, random forest model and decision tree model.
Year: 2023
Type: article
Study design: Observational Study
Abstract: OBJECTIVE: To investigate the predictive effect of a Back propagation (BP) neural network model, a random forest (RF) model and a decision tree model on the prognosis of interventional thrombolectomy for acute ischemic stroke (AIS) patients. METHODS: A total of 255 patients with AIS admitted to the Department of Neurology, Beiliu People's Hospital of Guangxi from March 2018 to February 2022 were retrospectively included, all of whom received interventional thromposectomy. Patients' prognosis was determined by the modified Rankin Scale (mRs) at 3 months after surgery, including the good prognosis group (mRs≤2 points) and the poor prognosis group (mRs 3-6 points). Clinical data of the two groups were collected to explore and screen the factors affecting poor clinical prognosis. Based on the selected influencing factors, the BP neural network, RF model, and decision tree models were established respectively, and their predictive performances were verified. RESULTS: All the three models predicted the same verification set data. The prediction accuracy, sensitivity and specificity of the BP neural network model were 0.961, 0.983 and 0.875, respectively. The prediction accuracy, sensitivity and specificity of the RF model were 0.948, 0.952 and 0.933, respectively. The prediction accuracy, sensitivity and specificity of the decision tree model were 0.882, 0.953 and 0.667, respectively. CONCLUSION: The three prediction models have shown good diagnostic efficacy and stability in the preliminary study of the prognosis of AIS mediated thrombectomy, which has important guiding significance for clinical prognosis assessment and selection of appropriate surgical population. The prediction model can be selected according to the actual situation of patients to provide more efficient guidance for clinicians.
Source: PubMed
Authors: Senlin Zhou, Jiajun Wei, Lairong Tang, Caimian Gu, Jiong Liang
Language: English
Subfield: Epidemiology
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W4380290474