A simple practical classification of cerebral infarcts on CT and its interobserver reliability

Area of Science:

• Neuroradiology
• Medical Imaging Analysis

Background:

• Cerebral infarct classification is crucial for understanding stroke progression and treatment efficacy.
• Accurate differentiation of infarct characteristics, including swelling and hemorrhage, is essential for patient management.

Purpose of the Study:

• To evaluate the interobserver reliability of a straightforward classification system for cerebral infarcts on CT scans.
• To determine if this method can distinguish infarcts based on site, size, swelling, and hemorrhagic transformation.

Main Methods:

• Two experienced neuroradiologists independently classified 119 CT brain scans of recent infarcts.
• Six less experienced radiologists classified 33 scans.
• Interobserver agreement was assessed using unweighted kappa statistics.

Main Results:

• Experienced neuroradiologists showed good to excellent agreement for infarct site/size (kappa=0.78) and swelling (kappa=0.8).
• Agreement for hemorrhagic transformation was fair (kappa=0.3).
• Less experienced radiologists demonstrated fair to excellent agreement.

Conclusions:

• The proposed CT-based cerebral infarct classification is simple, quick, and reliable.
• It effectively distinguishes infarcts with varying swelling and hemorrhagic transformation, aiding research into treatment effects.
• The method is practical for clinical research and practice, applicable to both CT and MRI.