Developing a Predictive Model for Ischemic Stroke Onset Time Using Transfer Learning
European Neurology · 2025
DOI: 10.1159/000549892 ↗Kurz gefasst
Fachartikel in der Zeitschrift European Neurology, 2025. Verfasst von Yang Du und 8 weiteren Autor:innen. Beteiligte Einrichtung: Department of Neurology, West China School of Medicine, Sichuan University, Sichuan University Affiliated Chengdu Second People's Hospital, Chengdu, …. Hauptthemen laut MEDLINE: Ischemic Stroke. Die Arbeit nennt 29 Literaturangaben.
Themen: Schlaganfall
Abstract (Originalsprache)
Introduction
Identification of acute ischemic stroke (AIS) patients within the 4.5-h therapeutic window is critical for therapy. Diffusion-weighted imaging (DWI) and fluid-attenuated inversion recovery (FLAIR) sequences are an approach to determine whether the time since stroke (TSS) is within 4.5 h. However, interobserver variability and limited accuracy are observed in visual assessments. We aimed to develop a transfer learning model for predicting AIS onset within 4.5 h. Materials and Methods: A total of 266 AIS patients with known TSS who underwent imaging scans before treatment were retrospectively analyzed, divided into a training set (n = 211) and a validation set (n = 55). The model was built using DWI and FLAIR sequences. After image preprocessing and data augmentation, a 3D ResNet-18 pretrained on the Kinetics dataset was selected and adapted via transfer learning with DWI-FLAIR input. The model performance was compared with human visual assessment, which was based on the DWI-FLAIR mismatch principle. Partial mismatch was defined as hyperintense infarct on DWI with a smaller corresponding hyperintense area on FLAIR.
Results
Baseline characteristics did not differ between the training and validation sets. On the validation set, the model achieved sensitivity of 0.833 (0.703–0.941), specificity of 0.880 (0.737–1.000), and AUC of 0.929 (0.758–0.935), outperforming human visual assessment (sensitivity 0.767 [0.613–0.903]; specificity 0.360 [0.185–0.560]; AUC 0.563 [0.451–0.693]). For partial DWI-FLAIR mismatch cases, the model correctly classified all 15 cases, whereas humans classified 4.
Conclusion
The 3D ResNet-18 model shows promise in identifying AIS within 4.5 h, including partial DWI-FLAIR mismatch, but requires multicenter validation before use.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Neurology, West China School of Medicine, Sichuan University, Sichuan University Affiliated Chengdu Second People's Hospital, Chengdu, China.
- School of Clinical Medicine, Chengdu Medical College, Chengdu, China.
- Department of Outpatient, General Hospital of Western Theater Command, Chengdu, China.
- Department of Radiology, West China School of Medicine, Sichuan University, Sichuan University Affiliated Chengdu Second People's Hospital, Chengdu, China.
- School of Clinical Medicine, The First Affiliated Hospital of Chengdu Medical College, Chengdu, China, zhanglili201908@163.com.
- Department of Neurology, West China School of Medicine, Sichuan University, Sichuan University Affiliated Chengdu Second People's Hospital, Chengdu, China, dingxin105@163.com.
Zitieren
APA
Du, Y., Wang, S., Wang, W., Zhang, W., Chen, X., Li, Y., Li, J., Zhang, L., & Ding, X. (2025). Developing a Predictive Model for Ischemic Stroke Onset Time Using Transfer Learning. European Neurology. https://doi.org/10.1159/000549892
Vancouver
Du Y, Wang S, Wang W, Zhang W, Chen X, Li Y, et al. Developing a Predictive Model for Ischemic Stroke Onset Time Using Transfer Learning. European Neurology. 2025. doi:10.1159/000549892
BibTeX
@article{Du2025,
author = {Du, Yang and Wang, Shuai and Wang, Weidong and Zhang, Wenming and Chen, Xiang and Li, Yuan and Li, Jie and Zhang, Lili and Ding, Xin},
title = {Developing a Predictive Model for Ischemic Stroke Onset Time Using Transfer Learning},
journal = {European Neurology},
year = {2025},
doi = {10.1159/000549892}
}
EUVIMED