Diffusion MR Imaging and Transient Ischemic Attacks
To the Editor:
With great interest we read the excellent study of Kidwell et al1 about diffusion MRI in patients with transient ischemic attacks (TIAs). In 48% of their 42 TIA patients, they found abnormal diffusibility on diffusion-weighted imaging (DWI). The frequency of positive DWI findings increased with increasing TIA duration. In DWI-positive TIA patients, diffusional abnormalities were smaller and alterations of apparent diffusion coefficients (ADC) less pronounced than in stroke patients.
We are pleased to present a study with similar results. We studied 40 consecutive patients (mean±SD age 61±10.5 years; 20 men) eligible for MRI scanning, in whom TIA diagnosis was assumed based on patient history. All included patients reported an acute focal neurological deficit that lasted <24 hours. Because symptoms had completely resolved before they reached the hospital, clinical deficits had not been witnessed by a physician. DWI was performed within 7 to 72 hours after symptom onset (mean 36.5 hours), on a 1.5-T MR imaging system (General Electric Medical Systems) using a single-shot, multislice spin-echo echo planar imaging sequence. (TR 12000 ms, TE 101 ms, flip angle 90°, field of view 40×20 cm, matrix 128×64 mm, axial slices, thickness 5 mm, interslice gap 2.5 mm, inversion recovery pulse TI 2200 ms). Diffusion gradients were applied in 3 orthogonal directions, with a maximum b value of 1000 s/mm2.
In 14 patients (35%), DWI revealed a hyperintense lesion indicative of acute ischemic compromise. Five of those patients had multiple lesions on T2-weighted images, which suggested small-vessel disease. In these patients DWI was useful to identify the acute ischemic lesion among preexisting white matter lesions. For DWI-positive TIA patients, the mean relative ADC value of the entire DWI lesion was calculated (ADC=−log (signal intensityb=1000/signal intensityb=0)/1000) and expressed as relative ADC compared with contralateral normal-appearing tissue. Relative ADC values of TIA-associated DWI lesions ranged from 0.67 to 0.90, with a mean±SD of 0.81±0.07, which was significantly higher compared with values in 51 stroke patients, who had a mean relative ADC value of 0.62±0.1 when measured within the same time window of 72 hours (P<0.001, Mann-Whitney test).
TIA duration differed significantly between DWI-positive patients, who had a mean TIA duration of 7.1±9.2 hours, and DWI-negative patients, with a mean TIA of 3.2±6.7 hours (P<0.05, Mann-Whitney test). When patients were sorted into groups according to TIA duration, the frequency of positive DWI findings ranged from 0% to 67% (Figure⇓). These results are similar to data from Kidwell et al,1 who found that among patients with TIA lasting for <1 hour, 33% had DWI lesions, compared with 71% among patients with TIA duration of 12 to 24 hours. However, in their study no details were mentioned about patients with a very short TIA duration of only several minutes. All of our patients who reported symptoms lasting for 5 minutes or less had normal DWI results. On the other hand, among patients with positive DWI findings the shortest TIA duration was 10 minutes, and only 2 of 26 patients (8%) had TIA lasting for <30 minutes. These data may indicate pathophysiological differences between the very brief ischemia lasting for a few minutes and ischemic events lasting for hours. However, TIA duration should be interpreted cautiously, because it represents the patient’s estimate rather than an exactly measured time period. Nevertheless, the DWI studies on TIA seems to indicate that ischemic tissue injury may occur if clinical symptoms last longer than a few minutes. The likelihood of ischemic injury seems to increase with advancing duration of clinical symptoms, even if they eventually disappear within 24 hours (the threshold chosen to discriminate TIA from stroke).
For patients who present in clinical practice with a history of acute focal neurological deficits which have already resolved rather than with identifiable clinical symptoms, DWI might be useful to verify the suspected ischemic etiology. Furthermore, according to a recent abstract,2 in DWI-positive TIA patients further diagnostic evaluation is more likely to reveal an underlying cardiac or cerebrovascular etiology.
- Copyright © 1999 by American Heart Association
Kidwell CS, Alger JR, Di Salle F, Starkman S, Villablanca P, Bentson J, Saver JL. Diffusion MRI in patients with transient ischemic attacks. Stroke. 1999;30:1174–1180.
Ay H, Buonanno FS, Schaefer PW, Furie KL, Rordorf G, Gonzalez RG, Kistler JP, Koroshetz WJ. Clinical and diffusion-weighted characteristics of an identifiable subset of TIA patients with acute infarction. Stroke. 1999;30:1:235. Abstract.
We thank Dr Engelter and colleagues for their thoughtful comments on our study. Their observations from an additional 40 TIA patients undergoing diffusion-weighted MR imaging confirm and extend our report.
With regard to TIAs of very brief duration, the briefest TIA in our cohort was 10 minutes, which occurred in 3 patients. Only 1 of these exhibited a DWI abnormality. Ten patients in our cohort had TIAs ≤30 minutes in duration. Thirty percent of these exhibited DWI lesions.
The convergent results of Engelter and colleagues and ourselves demonstrate that the likelihood of DWI abnormalities developing in TIA patients increases with the duration of ischemia. However, as brief an ischemic episode as 10 minutes may produce DWI alteration. The severity of ischemia is likely also an important determinant of diffusion abnormality. Clinicians should be aware that in a substantial minority of patients with TIAs as brief as 10 to 60 minutes in length, diffusion MR can demonstrate abnormalities, and these findings can alter the diagnosis of stroke mechanism and the treatment plan.R1
Kidwell CS, Alger JR, Di Salle F, Starkman S, Villablanca P, Bentson J, Saver JL. Diffusion magnetic resonance imaging in patients with transient ischemic attacks. Stroke.. 1999;30:1174–1180.