Related Experiment Video
Updated: May 27, 2025
09:33
Diffusion Tensor Magnetic Resonance Imaging in the Analysis of Neurodegenerative Diseases
Published on: July 28, 2013
28.2K
Deep learning-based classification of diffusion-weighted imaging-fluid-attenuated inversion recovery mismatch.
Pum Jun Kim1, Dongyoung Kim1, Joonwon Lee2
1Graduate School of Artificial Intelligence, Ulsan National Institute of Science and Technology, Ulsan, Republic of Korea.
Scientific Reports
|February 18, 2025
Summary
A deep learning model accurately identifies diffusion-weighted imaging (DWI) and fluid-attenuated inversion recovery (FLAIR) mismatch in acute stroke patients. This AI tool reduces subjectivity in assessing mismatch, aiding treatment decisions for recanalization therapy.
Area of Science:
- Medical Imaging and Diagnostics
- Artificial Intelligence in Medicine
- Neurology and Stroke Management
Background:
- Diffusion-weighted imaging (DWI)-fluid-attenuated inversion recovery (FLAIR) mismatch is crucial for identifying acute ischemic stroke patients eligible for recanalization treatment.
- Visual assessment of DWI-FLAIR mismatch suffers from inter-rater variability, impacting diagnostic accuracy and treatment consistency.
- Objective and automated methods are needed to overcome the limitations of subjective visual assessment in stroke imaging.
Purpose of the Study:
- To develop and validate a deep learning-based classifier for automated DWI-FLAIR mismatch categorization in acute ischemic stroke.
- To improve the objectivity and consistency of DWI-FLAIR mismatch assessment for better patient selection in recanalization therapy.
Main Methods:
- Convolutional Neural Network (CNN) models were developed for two binary classifications: DWI-FLAIR match vs. non-match and match vs. mismatch.
- Data from 2369 patients (derivation cohort) and 679 patients (external validation cohorts) across four stroke centers were utilized.
- Model performance was evaluated using Area Under the Curve (AUC) for internal and external validation sets.
Main Results:
- The classifier for DWI-FLAIR match vs. non-match achieved an internal AUC of 0.862 and external AUCs of 0.829 and 0.835.
- The classifier for DWI-FLAIR match vs. mismatch demonstrated higher performance with an internal AUC of 0.934 and external AUCs of 0.883 and 0.913.
- The deep learning approach significantly reduced subjectivity in DWI-FLAIR mismatch assessment.
Conclusions:
- A deep learning-based classifier can accurately and reliably categorize DWI-FLAIR mismatch in acute ischemic stroke patients.
- This AI tool has the potential to diminish subjectivity and support targeted clinical decision-making for recanalization treatments.
- Automated DWI-FLAIR mismatch assessment using deep learning can enhance the efficiency and effectiveness of acute stroke care.
