AI-Driven Stroke Prediction and Outcome Forecasting Using Neuroimaging and Clinical Stream Data

Stroke is a leading cause of mortality and long-term disability worldwide, necessitating early prediction and outcome forecasting to improve treatment decisions. This study proposes an end-to-end artificial intelligence framework integrating neuroimaging (CT/MRI) and time-series clinical stream data such as vital signs, laboratory results, and electronic health records. The framework employs a 3D convolutional neural network for imaging, a temporal transformer for clinical data, and an attention-based multimodal fusion module. The proposed approach enables accurate prediction of acute ischemic stroke and forecasting of patient outcomes including functional independence, mortality, and complications. The model demonstrates the potential to enhance clinical decision-making, patient triage, and rehabilitation planning in real-world healthcare systems.