Related Experiment Videos
Interobserver agreement for diagnostic MRI criteria in suspected multiple sclerosis
F Barkhof1, M Filippi, J H van Waesberghe
1Department of Radiology, Academic Hospital of the Vrije Universiteit, Amsterdam, The Netherlands. f.barkhof@azvu.nl
Neuroradiology
|June 24, 1999
Summary
Experienced observers show good agreement on MRI criteria for multiple sclerosis (MS) diagnosis, especially when using dichotomized criteria. This validates MRI
Area of Science:
- Neurology
- Radiology
- Medical Imaging
Background:
- Magnetic Resonance Imaging (MRI) is crucial for diagnosing multiple sclerosis (MS).
- Evaluating interobserver agreement in applying diagnostic criteria to MRI scans is essential for consistent diagnosis.
- Standardized criteria application ensures reliable interpretation of imaging findings.
Purpose of the Study:
- To assess the agreement among experienced observers in applying diagnostic criteria to MRI scans at initial patient presentation.
- To evaluate the reliability of various MRI features and composite criteria for MS diagnosis.
- To determine if dichotomized criteria improve observer concordance.
Main Methods:
- Five experienced radiologists reviewed 25 sets of MRI scans (T2-weighted and gadolinium-enhanced T1-weighted).
- Observers scored specific lesion locations, counts, and characteristics (e.g., periventricular, juxtacortical, enhancing).
- Agreement was quantified using weighted kappa values and mean average difference to the median, with and without dichotomization.
Main Results:
- Poor agreement was noted for raw scores of total T2 lesions, occipital, oval, juxtacortical, and hypointense lesions.
- Moderate agreement was found for frontal, callosal, basal ganglia, and large T2 lesions.
- Good agreement was achieved for parietal, temporal, infratentorial, and periventricular lesions.
- Dichotomized criteria, particularly lesion count on T2-weighted images, significantly improved agreement (P < 0.05).
- The Paty and Fazekas criteria showed good agreement, while the Barkhof criteria demonstrated moderate agreement.
Conclusions:
- Despite variability in scoring individual lesion counts, experienced observers exhibit substantial agreement on dichotomized MRI criteria for MS diagnosis.
- Composite and dichotomized MRI criteria, especially lesion counts, are validated for reliable use in supporting MS diagnosis.
- The findings support the continued use of MRI as a key paraclinical tool in the diagnostic pathway for multiple sclerosis.