Diagnostic and prognostic value of multimodal MRI in transient ischemic attack


  • Conflict of interest: The authors declare no potential conflict of interest.
  • Funding: This study was supported by a grant from the Brain Research Centre of the 21st Century Frontier Research Program funded by Ministry of Science and Technology of Korea (M103KV010010 06 K2201 01010).



The clinical diagnosis of transient ischemic attack is highly subjective, and the risk prediction after transient ischemic attack using the clinical parameters still remains unsatisfactory.


We aimed to investigate the diagnostic and prognostic value of multimodal magnetic resonance imaging in transient ischemic attack patients.


We prospectively performed diffusion-weighted imaging, perfusion-weighted imaging, and intracranial and extracranial magnetic resonance angiogram within 72 h of symptom onset in 162 transient ischemic attack patients defined by the classical time-based definition. Follow-up diffusion-weighted imaging was obtained three-days later in patients who did not exhibit lesions on the initial diffusion-weighted imaging. The occurrence of clinical events (transient ischemic attack or stroke) three-months after the initial transient ischemic attack was recorded, and the ABCD2 and ABCD3-I scores were calculated. The clinical and imaging parameters were compared between patients with and without initial diffusion-weighted imaging lesion, clinical events, and follow-up diffusion-weighted imaging lesions.


Abnormalities were present on diffusion-weighted imaging, perfusion-weighted imaging, and magnetic resonance angiogram in 38·9%, 44·1%, and 51·9% of patients, respectively. Diffusion-weighted imaging plus perfusion-weighted imaging explained 64·8%, and the addition of magnetic resonance angiogram explained 74% of the transient ischemic attack symptoms. The initial diffusion-weighted imaging positivity was associated with longer time from symptom onset to magnetic resonance imaging examination (odds ratio, 1·039; 95% confidence interval, 1·008–1·071; P = 0·013). On follow-up diffusion-weighted imaging, new lesions were found in 46·7% of the patients who initially showed normal diffusion-weighted imaging findings. Initial perfusion-weighted imaging abnormality predicted the appearance of follow-up diffusion-weighted imaging lesion (chi-square = 7·774, P = 0·005). During the three-months follow-up, 23 patients (14·2%) experienced subsequent transient ischemic attack (n = 16) or stroke (n = 7). Symptomatic magnetic resonance angiogram abnormality (odds ratio, 12·667; 95% confidence interval, 2·859–56·110; P = 0·001) was the only independent factor associated with clinical events with a sensitivity of 91·3% and specificity of 54·7% (C statistics, 0·73). None with initially normal multimodal magnetic resonance imaging findings developed subsequent clinical events.


Approximately three-quarter of transient ischemic attack is associated with multimodal magnetic resonance imaging abnormality. Initial perfusion-weighted imaging abnormality predicts newly developed diffusion-weighted imaging lesions, and symptomatic magnetic resonance angiogram abnormality seems to be the most important predictor for subsequent clinical events. Multimodal magnetic resonance imaging appears to be useful in assessing transient ischemic attack and predicting outcome in these patients.