Cerebral Infarction, Cerebrovascular Disorders, Posterior Ischemic Stroke
Conditions
Keywords
Machine Learning, Posterior Cerebral Artery, Basilar Artery, Vertebral Artery, Endovasculat treatment, Thrombectomy, Prediction Algorithm
Brief summary
This study aims to develop and externally validate machine learning prediction models (FORECAST-PC) to determine 90-day functional outcomes for patients suffering from acute ischemic stroke in the posterior circulation who are selected for endovascular thrombectomy (EVT)
Detailed description
Outcome prediction after endovascular treatment (EVT) for posterior circulation (PC) acute ischemic stroke remains challenging, as current prognostic tools are often limited strictly to basilar artery occlusions or rely heavily on anterior circulation data. The FORECAST-PC study is an investigator-initiated, retrospective, international multicenter cohort study designed to develop and validate comprehensive PC outcome prediction models across key stages of the clinical pathway: Baseline, Pre-EVT, Immediately Post-EVT, and 24-hours Post-EVT. The models are derived from patients across three European centers and externally validated on an unseen cohort from 11 international centers to ensure robustness against geographic domain shifts and variations in clinical workflows. The clinical utility of these models is assessed using Decision Curve Analysis to demonstrate the net benefit of model-guided prognostication over standard clinical assumptions.
Interventions
OTHERFORECAST-PC Prediction Models
Application of machine learning gradient-boosted decision tree models (Full-Feature and Light versions) at admission and 24-hours to predict 90-day functional outcome.
Sponsors
Dr. med. Alexander Salerno, MD-PhD
Study design
Observational model
COHORT
Time perspective
RETROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* Age \> 18 years * Presentation with acute ischemic stroke (AIS) attributed to a radiologically confirmed posterior circulation occlusion * Selected for Endovascular Thrombectomy (EVT) according to the treating physician * Pre-stroke modified Rankin Scale (mRS) 0-3 * EVT initiated or attempted within 24-hours of the last known well (LKW) time * Includes cases with/without prior intravenous thrombolysis (IVT), mothership, and drip-and-ship pathways
Exclusion criteria
* EVT carried out beyond 24 hours from stroke onset * EVT performed secondarily after in-hospital worsening or stroke recurrence * In-hospital strokes (to avoid the confounding influence of prior systemic comorbidities and the acute event that necessitated initial admission) * Missing primary outcome data * Inconsistencies between descriptive dependence status and raw pre-stroke mRS * Stroke cases occurring prior to 2015 (to account for major changes in endovascular treatment techniques)
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Unfavorable Functional Outcome at 90 Days | 90 days | Defined as a dichotomized 90-day modified Rankin Scale (mRS) score of 3 to 6 (where 0-2 is favorable and 3-6 is unfavorable). |
Countries
Switzerland
Outcome results
None listed