Prognostic Modeling Based on Post-Endovascular Thrombectomy Systolic Blood Pressure Trajectories Using Explainable Artificial Intelligence: A Secondary Analysis of the OPTIMAL-BP Trial
An AI model using blood-pressure patterns after stroke treatment better predicts recovery than traditional clinical measures, guiding smarter BP management.
This secondary analysis of the OPTIMAL-BP RCT developed an explainable AI model incorporating post-EVT blood pressure trajectory metrics to predict 90-day functional independence in 288 stroke patients across 19 centers. The DNN with SBP features (AUC 0.86) significantly outperformed clinical-variable models alone, with SHAP analysis revealing actionable BP management targets for different treatment arms.
What the study was
- Study design
- Secondary analysis of RCT (OPTIMAL-BP); 19-center South Korea; retrospective ML modeling
- Population
- Acute ischemic stroke patients after successful endovascular thrombectomy; South Korea
- Sample size
- 288
- Category
- Diagnostics
- Maturity
- Validated
- Journal
- Journal of Medical Systems
Why it surfaced
Multi-center RCT-embedded ML model with SHAP explainability addresses acute stroke BP management — a high-volume, clinically impactful application; AUC improvement over standard is statistically confirmed.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.