Multimodal machine learning for early risk stratification of post-stroke cognitive impairment - PubMed
an hour ago
- #stroke
- #cognitive impairment
- #machine learning
- Post-stroke cognitive impairment (PSCI) is a major vascular contributor to dementia, affecting recovery and quality of life.
- A stacking-based multimodal machine learning model integrating clinical, demographic, and neuroimaging features was developed for early PSCI prediction in acute ischemic stroke patients.
- The study involved 1070 patients, with 37.2% developing PSCI. The model achieved high internal validation accuracy (98.13%) and good external validation results (81.00% accuracy).
- Key predictors included infarct volume, cortical lesions, medial temporal lobe atrophy, and baseline NIHSS score.
- The model serves as a reliable tool for early detection, supporting personalized intervention strategies to prevent progression to post-stroke dementia.