Data Management And Sharing System of China National Clinical Research Center for Neurological Diseases
Predicting functional outcome in patients with acute brainstem infarction using deep neuroimaging features
First author：Lingling Ding
Corresponding author：Yongjun Wang,Zixiao Li
essential_information
- Person in charge：Lingling Ding
- Data Sources： CNSR-Ⅲ
- Number of papers：CNSR-Ⅲ20210038
- Published time：2021-11-12
2021-11-12
Published articles
Journal： European Journal of Neurology
Publication type: Original Article
IF(IF): 6.089
PMID: 34773321 Link To Pubmed Full text Download
Background And Purpose
This study aims to utilize deep convolutional neural networks (CNN) to automatically extract high-level neuroimaging features, in combination with clinical and laboratory data, for predicting functional outcomes in patients with acute brainstem infarction.
Method
This nationwide multicenter stroke registry study included 1482 patients with acute brainstem infarction. We applied CNNs to automatically extract deep neuroimaging features from diffusion-weighted imaging. Deep learning models based on clinical features, laboratory features, conventional imaging features (infarct volume, number of infarctions), and deep neuroimaging features were trained to predict functional outcomes at 3 months poststroke. Unfavorable outcome was defined as modified Rankin Scale score of 3 or higher at 3 months. The models were evaluated by comparing the area under the receiver operating characteristic curve (AUC).
Conclusion
Deep learning models can successfully extract objective neuroimaging features from the routine radiological data in an automatic manner and aid in predicting the functional outcomes in patients with brainstem infarction at 3 months with very high accuracy.
Main results
A model based solely on 14 deep neuroimaging features from CNNs achieved an extremely high AUC of 0.975 (95% confidence interval [CI] = 0.934-0.997) and significantly outperformed the model combining clinical, laboratory, and conventional imaging features (0.772, 95% CI = 0.691-0.847, p < 0.001) in prediction of functional outcomes. The deep neuroimaging model also demonstrated significant improvement over traditional prognostic scores. In an interpretability analysis, the deep neuroimaging features displayed a significant correlation with age, National Institutes of Health Stroke Scale score, infarct volume, and inflammation factors.