Related Experiment Video
Updated: Oct 20, 2025
A Magnetic Resonance Imaging Protocol for Stroke Onset Time Estimation in Permanent Cerebral Ischemia
Published on: September 16, 2017
Sensitivity of diffusion-weighted magnetic resonance imaging in transient global amnesia as a function of time from
Matthew L Wong1, Lucas Oliveira J E Silva2, Danielle J Gerberi3
1Department of Emergency Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts, USA.
Background:
The objective was to systematically evaluate the sensitivity of diffusion-weighted magnetic resonance imaging (DW-MRI) for transient global amnesia (TGA) across various time frames compared to the reference-standard clinical criteria.
Methods:
All indexed publications related to TGA and MRI through June 2020 were retrieved by a medical librarian. Two independent reviewers identified original research studies of adults with a clinical diagnosis of TGA using Caplan and Hodges and Warlow criteria (reference standard) who were evaluated with DW-MRI. Pooled estimates and its 95% confidence intervals (CI) for the proportion of acute TGA patients with positive DW-MRI (i.e., sensitivity) were obtained using random-effects meta-analysis for various time frames. Quality assessment was performed using the revised Quality of Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool.
Results:
After screening 665 reports, we identified 81 potentially relevant studies. Twenty-three studies representing 1688 patients met eligibility criteria, but not all studies had data available for meta-analysis. The pooled sensitivity (also described as positivity rate) of DW-MRI was 15.6% (95% CI = 2.6%-35.0%) between 0 and 12 h from symptom onset, 23.1% (95% CI = 6.1%-45.7%) at 0-24 h, 72.8% (95% CI = 40.8%-96.3) at 12-24 h, 68.8% (95% CI = 44.8%-88.8%) at 24-36 h, 72.4% (95% CI = 59.8%-83.5%) at 36-48 h, 82.8% (95% CI = 54.7%-99.6%) at 48-60 h, 66.9% (95% CI = 47.5%-83.9%) at 60-72 h, and 72.0% (95% CI = 30.1%-100.0%) at 72-96 h. There was significant concern for risk of bias in the QUADAS-2 domains of patient selection and index test, yielding a low level of certainty in the pooled estimates.
Conclusion:
DW-MRI lesions are uncommon in patients with TGA early after symptom onset, but the sensitivity (i.e., positivity rate) of DW-MRI increases with time. Despite the limited quality of existing evidence, obtaining an early DW-MRI in patients with clinical diagnosis of TGA in the acute setting is likely a low-yield test.
