Objective
To determine if sufficient evidence exists
to support the use of magnetic resonance angiography as a means of
selecting patients with recently symptomatic high grade carotid
stenosis for surgery.
Design
Systematic review of published research
on the diagnostic performance of magnetic resonance angiography,
1990-9.
Main outcome measures
Performance
characteristics of diagnostic test.
Results
126 potentially relevant articles were
identified, but many articles failed to examine the performance of
magnetic resonance angiography as a diagnostic test at the surgical
decision thresholds used in major clinical trials on endarterectomy. 26
articles were included in a meta-analysis that showed a maximal joint
sensitivity and specificity of 99% (95% confidence interval 98% to
100%) for identifying 70-99% stenosis and 90% (81% to 99%) for
identifying 50-99% stenosis. Only four articles evaluated
contrast enhanced magnetic resonance angiography.
Conclusions
Magnetic resonance angiography is
accurate for selecting patients for carotid endarterectomy at the
surgical decision thresholds established in the major endarterectomy
trials, but the evidence is not very robust because of the
heterogeneity of the studies included. Research is needed to determine
the diagnostic performance of the most recent developments in magnetic
resonance angiography, including contrast enhanced techniques, as well
as to assess the impact of magnetic resonance angiography on surgical
decision making and outcomes.
What is already known on this topicCarotid endarterectomy for recently symptomatic carotid stenosis is
beneficial in patients with 70-99% stenosis as measured by
conventional angiographyIt is not known whether the less invasive imaging technique of magnetic
resonance angiography can accurately identify patients who will benefit
from surgeryWhat this study addsMagnetic resonance angiography is highly sensitive and specific in
diagnosing 70-99% carotid stenosisHowever, the studies on which this conclusion is based are of low
quality and high heterogeneity