Applications of artificial intelligence for DWI and PWI data processing in acute ischemic stroke: Current practices
Ines Ben Alaya1, Hela Limam2, Tarek Kraiem1
1Tunis El Manar University, Higher Institute of Medical Technology of Tunis, Laboratory of Biophysics and Medical Technology, 1006 Tunis, Tunisia.
Clinical Imaging
|October 14, 2021
Summary
Artificial Intelligence (AI) can automate the analysis of Perfusion-Weighted Imaging (PWI) and Diffusion-Weighted Imaging (DWI) for acute ischemic stroke (AIS). This AI-driven approach improves the speed and accuracy of identifying stroke regions, aiding treatment decisions.
Area of Science:
- Neuroimaging
- Artificial Intelligence in Medicine
- Stroke Diagnostics
Background:
- Multimodal Magnetic Resonance Imaging (MRI) using Perfusion-Weighted Imaging (PWI) and Diffusion-Weighted Imaging (DWI) is crucial for acute stroke diagnosis.
- Visual interpretation of PWI/DWI data helps triage Acute Ischemic Stroke (AIS) patients for reperfusion therapy, especially beyond 6 hours.
- This interpretation identifies the ischemic core (DWI) and penumbra (PWI), guiding treatment decisions for patients who may benefit from endovascular therapy.
Purpose of the Study:
- To provide an overview of Artificial Intelligence (AI) applications for automating PWI/DWI data processing in acute ischemic stroke.
- To highlight the potential contributions of AI in clinical practices for stroke assessment.
- To compare current AI approaches based on key requirements for automated stroke analysis.
Main Methods:
- Review of Artificial Intelligence (AI) applications for automated post-processing of PWI and DWI data.
- Focus on AI's ability to identify stroke slices, stroke hemisphere, segment DWI regions, and measure hypoperfused PWI tissue.
- Comparison of different AI methodologies against essential clinical criteria.
Main Results:
- Automated procedures for PWI/DWI analysis can significantly enhance the reproducibility and accuracy of stroke assessment.
- AI offers a solution to the time-consuming and variable nature of manual PWI/DWI interpretation.
- AI-driven automation is essential for improving clinical diagnosis and therapeutic decision-making in AIS.
Conclusions:
- AI holds significant promise for revolutionizing the analysis of neuroimaging data in acute stroke.
- Automated PWI/DWI processing via AI can overcome limitations of manual interpretation, reducing inter- and intra-observer variability.
- Implementing AI in clinical workflows can lead to more efficient and reliable stroke treatment decisions.
