Advances in the Application of Machine Learning in Magnetic Resonance Imaging for the Assessment of Acute Ischemic Stroke
Abstract：Stroke is a severe cerebrovascular disease with high disability and mortality rates, accounting for 87% of cases as ischemic stroke [1]. Acute ischemic stroke (AIS) refers to ischemic stroke occurring within 2 weeks of onset, with a disability rate of 14.6% to 23.1% at 3 months post-onset [2]. Currently, MRI has been regarded as a key imaging modality for evaluating AIS, especially showing significantly higher sensitivity and accuracy in detecting acute infarct lesions compared to CT. However, MRI has longer scan times and requires higher patient cooperation. Machine learning (ML) is a technique that uses algorithms to learn patterns and features from large datasets for prediction, including traditional ML (such as logistic regression, random forest, etc.) and deep learning (DL). ML is widely applied in image classification, segmentation, and detection, and has the potential to overcome the limitations of MRI, offering new opportunities for the diagnosis, prognosis, and complication prediction of AIS. This review aims to summarize recent advances in the application of ML in MRI evaluation of AIS, in order to provide decision support for clinicians and offer references for future research directions.
Keywords：StrokeMagnetic Resonance ImagingMachine LearningDeep Learning
Publication Date：2025-12-15
Online Publishing Date：2025-12-29(First online date of this platform, not the publication date of the document)
Pages：3( 1755-1757 )