2165 Radiomics in Stroke Severity Scoring and Aphasia Prediction

View at publisher Summarise with VIRAT
Authors: Rishi Hoskeri, Prateek Dullur, Alexander D. Smith, Meenakshi Singhal, Brian Ellis, Aaron Anderson, Tony Tu
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

INTRODUCTION: Acute ischemic stroke requires immediate and accurate assessment of both imaging (CT/MRI) and clinical (NIH stroke scale) pathology to plan surgical intervention. Correlating objective textural features from imaging with neurological exam findings can enable improved surgical targeting (e.g., thrombectomy), prophylaxis standards, and screening for early rehabilitation strategies (e.g., neurostimulation). METHODS: Radiomic texture analysis was performed on trace DWI MRI (n = 955, SOOP) and T2w MRI (n = 245, ARC) images, with each dataset analyzed separately. Statistical radiomic features were extracted from radiologist-defined stroke lesions using the MATLAB Radiomics Toolbox. Random forest models were trained to predict clinical severity thresholds: NIHSS >5 in SOOP and WAB score

Fields of Study & Subjects

Radiomics and Machine Learning in Medical Imaging Acute Ischemic Stroke Management Artificial Intelligence in Healthcare

Publication details

Journal / source
Neurosurgery
Volume
72
Issue
Supplement_1
Pages
291–291
Publication date
26 March 2026
Language
EN