Related Experiment Video
Updated: May 6, 2026
06:01
A Thrombotic Stroke Model Based On Transient Cerebral Hypoxia-ischemia
Published on: August 18, 2015
14.8K
Hybrid Clot Histomic-Transcriptomic Models Predict Functional Outcome After Mechanical Thrombectomy in Acute Ischemic
Briana A Santo1,2,3, Kerry E Poppenberg1,2, Shiau-Sing K Ciecierska1
1Canon Stroke and Vascular Research Center, University at Buffalo, Buffalo , New York , USA.
Neurosurgery
|December 5, 2024
Summary
Combining clot histomics and mRNA expression significantly improves prediction of stroke recovery. These hybrid models offer better prognostication for early neurological improvement and long-term outcomes after stroke.
Area of Science:
- Computational pathology
- Genomics
- Neurology
Background:
- Stroke clot analysis reveals features linked to patient outcomes.
- Previous studies analyzed histology or gene expression separately.
- Few studies integrated both histomic and transcriptomic data for stroke prognostication.
Purpose of the Study:
- To investigate if combined clot histomics and mRNA expression predict early neurological improvement (ENI) and 90-day functional outcome (modified Rankin Scale Score, mRS).
- To compare the predictive performance of hybrid models against models using only histology or transcriptomics.
Main Methods:
- Paired histological and transcriptomic analysis of 32 stroke clots.
- Extraction of 237 histomic features and mRNA expression profiling.
- Machine learning (recursive feature elimination) to build predictive models for ENI and mRS.
Main Results:
- Hybrid models integrating histomic and transcriptomic features achieved high accuracy (90.8% for ENI, 93.7% for mRS).
- These models outperformed those based on histomics, transcriptomics, or clot composition alone.
- Optimal models identified 9 features for ENI and 7 for mRS.
Conclusions:
- Hybrid computational models enhance stroke outcome prognostication.
- Digital histology and mRNA signatures show potential as biomarkers for post-thrombectomy functional outcomes.
