Deep learning to identify stroke within 4.5 h using DWI and FLAIR in a prospective multicenter study

To enhance thrombolysis eligibility in acute ischemic stroke, we developed a deep learning model to estimate stroke onset within 4.5 h using diffusion-weighted imaging (DWI) and fluid-attenuated inversion recovery (FLAIR) images. Given the variability in human interpretation, our multimodal Res-U-Net (mRUNet) model integrates a modified U-Net and ResNet-34 to classify stroke onset as < 4.5 or ≥ 4.5 h. Using DWI and FLAIR images from patients scanned within 24 h of symptom onset, the modified U-Net generated a DWI-FLAIR mismatch image, while ResNet-34 performed the final classification. mRUNet was evaluated against ResNet-34 and DenseNet-121 on an internal test set (n = 123) and two external test sets: a single-center (n = 468) and a multi-center (n = 1151). mRUNet achieved an area under the receiver operating characteristic curve (AUC-ROC) of 0.903 on the internal set and 0.910 and 0.868 on external sets, significantly outperforming ResNet-34 and DenseNet-121. Our mRUNet model demonstrated robust and consistent classification of the 4.5-h onset-time window across datasets. By leveraging DWI and FLAIR images as a tissue clock, this model may support timely and individualized thrombolysis in patients with unclear stroke onset, such as those with wake-up stroke, in clinical settings.

Authors

Institutions

Publication Details

Journal
Scientific Reports
Published
2025-07-19
DOI
https://doi.org/10.1038/s41598-025-10804-6
Citations
2
Primary Topic
Acute Ischemic Stroke Management
Type
article
Field-Weighted Citation Impact
2.71

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Deep learning to identify stroke within 4.5 h using DWI and FLAIR in a prospective multicenter study

Seong Hwan Ahn, Dae‐Il Chang, Julian C. Hong, Kyung‐Pil Park et al.
2 citations
Scientific Reports
Acute Ischemic Stroke Management
2.71
article

Deep learning to identify stroke within 4.5 h using DWI and FLAIR in a prospective multicenter study

Seong Hwan Ahn, Dae‐Il Chang, Julian C. Hong, Kyung‐Pil Park, Keun‐Sik Hong, Jun Young Chang, Ju Hun Lee, Jin‐Man Jung, Han‐Jin Cho, Sung‐Il Sohn, Dong‐Eog Kim, Jae‐Kwan Cha, Eun‐Jae Lee, Eun Namgung, Dong‐Wha Kang, Jun Lee, Beom Joon Kim, Sun Uck Kwon, Gayoung Park, Hye-Soo Jung, Man-Seok Park, Young Sun Kim, Kyung Ho Yu
article en
2 citations

Abstract

To enhance thrombolysis eligibility in acute ischemic stroke, we developed a deep learning model to estimate stroke onset within 4.5 h using diffusion-weighted imaging (DWI) and fluid-attenuated inversion recovery (FLAIR) images. Given the variability in human interpretation, our multimodal Res-U-Net (mRUNet) model integrates a modified U-Net and ResNet-34 to classify stroke onset as < 4.5 or ≥ 4.5 h. Using DWI and FLAIR images from patients scanned within 24 h of symptom onset, the modified U-Net generated a DWI-FLAIR mismatch image, while ResNet-34 performed the final classification. mRUNet was evaluated against ResNet-34 and DenseNet-121 on an internal test set (n = 123) and two external test sets: a single-center (n = 468) and a multi-center (n = 1151). mRUNet achieved an area under the receiver operating characteristic curve (AUC-ROC) of 0.903 on the internal set and 0.910 and 0.868 on external sets, significantly outperforming ResNet-34 and DenseNet-121. Our mRUNet model demonstrated robust and consistent classification of the 4.5-h onset-time window across datasets. By leveraging DWI and FLAIR images as a tissue clock, this model may support timely and individualized thrombolysis in patients with unclear stroke onset, such as those with wake-up stroke, in clinical settings.

Scientific ReportsVol. 15(1)
Ulsan College (KR), Korea University (KR), Dong-A University Hospital (KR), Asan Medical Center (KR), Ansan University (KR), University of Ulsan (KR), Chonnam National University Hospital (KR), Hallym University Kangnam Sacred Heart Hospital (KR), Chosun University Hospital (KR), Dongguk University Ilsan Hospital (KR), Inje University Ilsan Paik Hospital (KR), Kyung Hee University Medical Center (KR), Pusan National University Hospital (KR), Keimyung University Dongsan Medical Center (KR), Korea Research Council for Industrial Science and Technology (KR), Kangdong Sacred Heart Hospital (KR), Pusan National University Yangsan Hospital (KR), Yeungnam University Medical Center (KR), Hallym University Sacred Heart Hospital (KR), Keimyung University (KR)
National Research Foundation, National IT Industry Promotion Agency, Korea Health Industry Development Institute, National Research Foundation of Korea, Ministry of Science and ICT, South Korea
Openalex Percentile: Top 9%
Acute Ischemic Stroke Management
2.71
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.