Related Experiment Video
Updated: May 27, 2026
05:07
Dynamic Assessments of Coronary Flow Reserve after Myocardial Ischemia Reperfusion in Mice
Published on: August 25, 2023
DSA-NRP: No-Reflow Prediction From Angiographic Perfusion Dynamics in Stroke EVT
IEEE Transactions on Medical Imaging
|May 25, 2026
Summary
A new machine learning model predicts the no-reflow complication after endovascular thrombectomy for acute ischemic stroke using digital subtraction angiography. This enables immediate risk assessment, unlike delayed MRI, improving patient management.
Area of Science:
- Neurology
- Medical Imaging
- Artificial Intelligence
Background:
- No-reflow is a complication after endovascular thrombectomy (EVT) for acute ischemic stroke (AIS), leading to poor outcomes.
- Current identification relies on delayed perfusion MRI (24 hours post-procedure), hindering timely intervention.
Purpose of the Study:
- To develop the first machine learning (ML) framework for immediate no-reflow prediction post-EVT.
- To utilize intra-procedural digital subtraction angiography (DSA) and clinical data for real-time prediction.
Main Methods:
- Retrospective analysis of AIS patients treated with EVT at UCLA Medical Center (2011-2024).
- No-reflow defined as >15% reduction in cerebral blood volume/flow within the infarct core.
- Extracted statistical and temporal perfusion features from DSA (anteroposterior and lateral views) to train ML classifiers.
Main Results:
- The ML framework significantly outperformed a clinical-features baseline (AUROC: 0.9330 vs. 0.7768, p=0.006).
- Intra-procedural DSA perfusion dynamics show potential for predicting microvascular integrity.
Conclusions:
- This novel ML approach enables immediate, accurate no-reflow prediction post-EVT.
- Facilitates proactive management of high-risk patients without relying on delayed imaging.
- Warrants validation in larger, independent cohorts.
