Assessing the eligibility of Brainomix e-ASPECTS for acute stroke imaging

Area of Science:

• Neurology
• Radiology
• Medical Imaging

Background:

• Accurate assessment of acute ischemic stroke is critical for mechanical thrombectomy eligibility.
• Alberta Stroke Program Early CT Score (ASPECTS) is a standard tool but has interobserver variability.
• Brainomix e-ASPECTS offers automated, standardized assessment of early ischemic changes on non-contrast CT (NCCT).

Purpose of the Study:

• To evaluate the clinical utility and diagnostic performance of Brainomix e-ASPECTS software.
• To assess the software's accuracy in an unselected, real-world cohort of acute ischemic stroke patients.
• To compare automated e-ASPECTS scores with traditional radiologist-assigned ASPECTS.

Main Methods:

• Retrospective analysis of 1,029 NCCT studies from 954 patients.
• Comparison of e-ASPECTS scores against radiologist-assigned ASPECTS as the reference standard.
• Assessment of diagnostic accuracy, sensitivity, specificity, and correlation between scoring methods.

Main Results:

• Strong correlation (ρ = 0.953) between e-ASPECTS and radiologist ASPECTS.
• High diagnostic accuracy: 96.3% overall, with sensitivity 95.8% and specificity 96.9%.
• Excellent score concordance: 92.3% exact matches, 97.7% within a 1-point difference; minimal misclassification for thrombectomy eligibility (0.4%).

Conclusions:

• E-ASPECTS demonstrates high diagnostic accuracy and agreement with expert interpretation.
• The software serves as a valuable decision support tool for acute stroke imaging.
• E-ASPECTS should complement, not replace, expert review, especially for critical low ASPECTS scores.