Deep Learning-Based Extraction of Biomarkers for the Prediction of the Functional Outcome of Ischemic Stroke Patients
See where this sits in the topic map →Summary AI-generated
- TL;DR
- Accurately predicting functional outcomes in ischemic stroke patients is clinically vital, but the practical value of brain CT scans for this task remains unclear.
- Problem
- It is challenging to accurately predict how ischemic stroke patients will functionally recover using standard clinical data and brain CT scans.
- Method
- We developed deep learning models to predict 3-month patient outcomes using only CT data, as well as a model that combines imaging features (such as detected vessel occlusions) with clinical tabular data.
- Results
- Image-only deep learning models achieved an AUC of 0.779, while the fused model incorporating both imaging and tabular data reached an AUC of 0.806.
- Contributions
- Not specified in the abstract.
- Limitations
- Not specified in the abstract.
- Takeaways
- Refining prognostic models by combining image biomarkers with clinical data enables more accurate functional outcome predictions for ischemic stroke patients.
- Applications
- Analyzing multi-center admission brain scans (NCCT and CTA) and clinical data from 743 ischemic stroke patients to predict their 3-month modified Rankin Scale (mRS) outcomes.
- Topics
- Diagnostics; Deep Learning; Medical Imaging; Stroke Outcome Prediction
- For industry
- Healthcare and clinical decision support.
- Why it matters
- Advances machine learning methods for medical prognosis, potentially aiding clinical decision-making in stroke care.
Abstract
Accurately predicting functional outcomes in stroke patients remains challenging yet clinically relevant. While brain CTs provide prognostic information, their practical value for outcome prediction is unclear. We analyzed a multi-center cohort of 743 ischemic stroke patients (<72 h onset), including their admission brain NCCT and CTA scans as well as their clinical data. Our goal was to predict the patients' future functional outcome, measured by the 3-month post-stroke modified Rankin Scale (mRS), dichotomized into good (mRS ≤ 2) and poor (mRS > 2). To this end, we developed deep learning models to predict the outcome from CT data only, and models that incorporate other patient variables. Three deep learning architectures were tested in the image-only prediction, achieving 0.779 ± 0.005 AUC. In addition, we created a model fusing imaging and tabular data by feeding the output of a deep learning model trained to detect occlusions on CT angiograms into our prediction framework, which achieved an AUC of 0.806 ± 0.082. These findings highlight how further refinement of prognostic models incorporating both image biomarkers and clinical data could enable more accurate outcome prediction for ischemic stroke patients.