Impact of Lesion Load Thresholds on Alberta Stroke Program Early Computed Tomographic Score in Diffusion-Weighted...

Area of Science:

• Neurology
• Radiology
• Medical Imaging

Background:

• Alberta Stroke Program Early Computed Tomography Score (ASPECTS) on diffusion-weighted imaging (DWI) guides acute stroke treatment.
• The precise lesion load threshold for defining ischemic regions in DWI-ASPECTS remains undefined.
• This study investigates the impact of varying lesion load thresholds on DWI-ASPECTS and compares automated analysis to visual assessment.

Purpose of the Study:

• To determine the effect of different lesion load thresholds on DWI-ASPECTS.
• To compare automated DWI-ASPECTS calculations with visual ratings by stroke neurologists.
• To evaluate how lesion load thresholds influence patient selection for acute stroke interventions.

Main Methods:

• Analysis of DWI lesions from 315 stroke patients against a probabilistic ASPECTS template.
• Application of multiple lesion load thresholds (>0%, >1%, >10%, >20%) per DWI-ASPECTS region.
• Comparison of automated DWI-ASPECTS scores with visual ratings by an experienced stroke neurologist.

Main Results:

• Automated DWI-ASPECTS varied significantly with lesion load thresholds, yielding median scores of 1, 5, 8, and 10 for >0%, >1%, >10%, and >20% thresholds, respectively.
• Overall agreement between automated and visual scoring was low across all thresholds (κw range: 0.020–0.386).
• Dichotomized scoring (≤7 or ≤5) showed fair to substantial agreement (κ range: 0.45–0.695), particularly with higher thresholds (>10%, >20%).

Conclusions:

• Automated DWI-ASPECTS assessment shows low agreement with visual scoring, irrespective of the lesion load threshold.
• Dichotomized scoring approaches yield more comparable results to visual assessment.
• The choice of lesion load threshold critically impacts patient selection for acute stroke treatment, highlighting the need for standardization.