Related Experiment Video
Updated: Nov 26, 2025
14:27
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
16.0K
Predictors of Stroke Outcome Extracted from Multivariate Linear Discriminant Analysis or Neural Network Analysis
Tomohisa Nezu1, Naohisa Hosomi2,3, Kazumasa Yoshimura4
1Department of Clinical Neuroscience and Therapeutics, Hiroshima University Graduate School of Biomedical and Health Sciences.
Journal of Atherosclerosis and Thrombosis
|December 10, 2020
Summary
Predicting stroke recovery is key. Higher alkaline phosphatase (ALP) levels and poor nutritional status in acute ischemic stroke patients independently predict worse functional outcomes, aiding in better patient management.
Area of Science:
- Neurology
- Biochemistry
- Medical Informatics
Background:
- Accurate prediction of functional outcome is crucial for managing acute ischemic stroke (AIS) patients.
- Identifying reliable prognostic factors aids in optimizing treatment strategies and patient care.
Purpose of the Study:
- To explore prognostic factors for functional outcome in AIS patients using advanced analytical models.
- To evaluate the association between candidate factors, baseline characteristics, and stroke outcome.
Main Methods:
- Analysis of 1,916 AIS patients with premorbid modified Rankin Scale (mRS) scores of 0-2.
- Utilized multivariate linear discriminant analysis and neural network analysis to predict poor functional outcome (mRS 3-6 at 3 months).
- Incorporated age, sex, initial neurological severity, nutritional status, and serum alkaline phosphatase (ALP) levels into prediction models.
Main Results:
- Both predictive models demonstrated that nutritional status and serum ALP levels significantly improved prediction accuracy.
- Higher serum ALP levels were observed in patients with poor outcomes (294.3±259.5 U/l vs. 246.3±92.5 U/l, P＜0.001).
- Multivariable logistic analysis confirmed higher ALP levels (1-SD increase) as an independent predictor of poor stroke outcomes (OR 1.21, 95% CI 1.02-1.49), even after adjusting for confounders.
Conclusions:
- Multivariate linear discriminant and neural network analyses consistently identified nutritional status and serum ALP levels as key indicators.
- These factors are independently associated with functional outcomes in acute ischemic stroke patients, offering valuable prognostic information.
