Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
May 8, 2026European Stroke JournalOpen Access

Abstract Number: Esoc2026a1608 Limited Added Value of Brain Mri for Prediction of Atrial Fibrillation Detected After Ischemic Stroke - Insights From Interpretable Machine Learning

View Full Paper
Ask AI
Bookmark
Share

Key result

Adding brain MRI features improves post-stroke AF prediction over clinical models but not age-HRV models.

  • P<0.01
  • n=1,689

Why the study?

Improved risk stratification for post-stroke AF is needed, and whether brain MRI provides additional predictive value over clinical variables or age and heart rate variability remains controversial.

Does the addition of brain MRI-derived features improve the prediction of atrial fibrillation detected after ischemic stroke in patients with ischemic stroke?

Population

1,227 patients in the primary cohort and 462 in the external cohort

Comparison

ML models with MRI-derived features vs models with clinical variables or age and HRV alone

Design

Cohort study with external validation

Authors

MSM. SchölsANAlexander NeldeMKMarkus Klammer

Discussion

Loading...

Member takes

Overview

Supports adding MRI to post-stroke AF risk models; extends prior age- and HRV-based ML predictions.

Key Points

  • This research aims to determine if MRI provides additional predictive value for atrial fibrillation detected after ischemic stroke (AFDAS).
  • Extracted MRI-derived features like lesion volume and central autonomic network involvement.
  • Developed tree-based machine-learning models for AFDAS and known atrial fibrillation (KAF).
  • Performed external validation in an independent cohort with 1,227 primary and 462 external patients.
  • MRI-derived features modestly discriminated AFDAS (ROC-AUC 0.75; p<0.01).
  • Key predictors included insular involvement and lesion volume for AFDAS.
  • Age and heart rate variability alone performed comparably (ROC-AUC 0.83 vs. 0.81) without MRI features.

Study Design

Type

Cohort (n=1,689)

Structured PICO

Does the addition of brain MRI-derived features improve the prediction of atrial fibrillation detected after ischemic stroke in patients with ischemic stroke?

P
Population
1,689 patients with ischemic stroke (primary cohort n=1,227; external cohort n=462)
I
Intervention
Machine learning models incorporating interpretable brain MRI-derived features (lesion volume, infarct distribution, insular involvement, and central autonomic network involvement)
C
Comparator
Machine learning models using clinical variables alone, or a previously published model using age and heart rate variability (HRV) alone
O
Outcome
Prediction of atrial fibrillation detected after stroke (AFDAS) and previously known AF (KAF) measured by ROC-AUCsurrogate

Main Result

Absolute Event Rate: 0.75% vs 0.67%

p-value: p=<0.01

Brain MRI-derived features do not add significant predictive value for post-stroke atrial fibrillation beyond simple clinical variables like age and heart rate variability.

Cite This Study

Schöls et al. (2026) conducted a cohort in Ischemic stroke (n=1,689). Brain MRI-derived features vs. Clinical variables or age and heart rate variability (HRV) alone was evaluated on Prediction of atrial fibrillation detected after stroke (AFDAS) (p=<0.01). Adding brain MRI-derived features to a clinical model improved prediction of atrial fibrillation detected after stroke (ROC-AUC 0.75 vs 0.67; p<0.01), but did not improve an age and HRV model.

synapsesocial.com/papers/69fd7e5cbfa21ec5bbf068c1https://doi.org/10.1093/esj/aakag023.768
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Development of an MRI based artificial intelligence model for the identification of underlying atrial fibrillation after ischemic stroke: a multicenter proof-of-concept analysis2025 · 3 citations
  2. 2Explainable machine learning models to improve prediction of incident stroke in atrial fibrillation patients using health records, medical imaging and ECG derived metrics2025
  3. 3Machine learning for stroke in heart failure with reduced ejection fraction but without atrial fibrillation: A post‐hoc analysis of the <scp>WARCEF</scp> trial2024 · 2 citations
  4. 4Decoding Stroke Patterns: A Novel Deep Learning Approach to Atrial Fibrillation Risk Stratification2024 · 1 citations
  5. 5Evaluating Machine Learning Models for Stroke Prognosis and Prediction in Atrial Fibrillation Patients: A Comprehensive Meta-Analysis2024 · 16 citations