AI-enhanced predictive modeling for acute ischemic stroke: Advancing diagnosis accuracy and patient outcomes

Acute ischemic stroke (AIS) requires rapid and accurate diagnosis to enable timely treatment and improve patient outcomes. This study presents an AI-enhanced predictive modeling approach for AIS that integrates advanced machine learning algorithms to improve diagnostic accuracy and provide reliable outcome predictions. We retrospectively collected clinical and imaging data from AIS patients and developed a predictive model combining a convolutional neural network (CNN) for early stroke detection on brain imaging with gradient boosting machine learning for prognostic outcome prediction. The model was trained and validated on separate cohorts and evaluated against standard clinical assessment and risk scores. Key results demonstrate that the AI-enhanced model achieved 96% sensitivity and 94% specificity for AIS detection, outperforming conventional clinical assessment (85% sensitivity, 88% specificity). It also accurately predicted 90-day functional outcomes with an area under the ROC curve (AUC) of 0.90, significantly higher than a baseline logistic model (AUC 0.82, p<0.01). These results indicate a substantial improvement over traditional methods. The integrated approach not only expedited stroke diagnosis but also provided robust prognostic insights, which together can support clinicians in making timely, informed treatment decisions. As a whole, the proposed AI-driven model significantly advances stroke diagnostic accuracy and outcome prediction, showcasing its potential to enhance acute stroke care and patient outcomes.