Can we identify stroke sub-type without imaging? A multidimensional analysis

Journal article

Highlights: The study investigates stroke sub-type identification (AIS and ICH) using non-imaging hemodynamic parameters, providing an alternative to traditional imaging. The PCA-LR model showed high accuracy, with AUCs of ≥0.92 in two-fold and ≥0.79 in six-fold cross-validation for stroke sub-type differentiation. Baroreceptor sensitivity (BRS), cerebral hemodynamics, and peripheral parameters were key in distinguishing AIS from ICH. This non-imaging model provides promising guidance for future research on early stroke diagnosis and management, especially in pre-hospital settings lacking imaging. Further validation is required to distinguish stroke from stroke mimics, with larger studies required to refine its clinical utility.

Authors: Alshehri, A; Panerai, RB; Lam, MY; Llwyd, O; Robinson, TG

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: IOP Publishing
• Journal: Medical Engineering & Physics
• Volume: 141
• Issue: 1
• Pages: 104364
• Article number: 104364
• Publication date: 2025-05-13
• Acceptance date: 2025-05-12
• DOI: 10.1016/j.medengphy.2025.104364
• EISSN: 1873-4030
• ISSN: 1350-4533
• Language: English
• Keywords: Baroreflex sensitivity; Transcranial doppler sonography; Cerebral autoregulation; Stroke; Blood flow velocity