Abstract 8: Early MRI in TIA And Minor Stroke: Do it or Lose it

Abstract

Background: MRI is not always completed early after TIA or minor stroke and this may affect its utility. We measured the impact of scanning an individual patient late versus early in the investigation of TIA and minor stroke. Methods: 263 patients with a TIA or minor stroke (NIHSS <4) from the CATCH study were included in this analysis. To be included in this sub study patients needed to have had a baseline MRI completed within 24 hours of symptom onset and a follow-up MRI at 90 days. All MRI images were acquired on a 3.0 Tesla GE scanner. Baseline and 90 day scans were assessed independently for the presence of any stroke lesion. The presence and pattern of any stroke lesion was then compared at the two time points. Lesion patterns were classified as: no definite stroke, single territory cortical stroke(s), multiple territory cortical strokes, single territory subcortical only stroke(s), multiple territory subcortical only strokes, and multiple strokes in one territory including a cortical stroke Results: Stroke of any age, in any location was more common on the baseline MRI versus 90day MRI (68% versus 58%, p=0.005). A substantial proportion of the negative scans at 90 days had a clearly identifiable stroke on the baseline scan (35/115: 30%) that was missed on the 90day scan. All of these lesions were acute or subacute DWI lesions on the baseline scan showing non-specific white matter hyperintensity or no abnormality on the 90day scan. Among 104 patients with a stroke lesion on the 90 day MRI considered as a cause for the presenting symptoms, this lesion was the correct lesion in only 78 (53%) patients. 89 (34%) patients had a different lesion pattern on the baseline scan versus the 90day scan. The main difference observed was that patients with multiple DWI lesions on the baseline scan were either seen as a single or no lesions on the 90day MRI. Conclusion: Completing an MRI in a delayed fashion after TIA or minor stroke reduces the diagnostic yield of the imaging. Not only does it reduce lesion detection, but also the pattern of the lesions is missed. Conclusions regarding the original event may be false if based only on a delayed MRI. If minor stroke and TIA patients are going to be scanned with MRI this should be completed early after symptom onset.