The impact of MRI on stroke management and outcomes: a systematic review

Rationale, aims and objectives: Magnetic resonance imaging (MRI) is widely used in stroke evaluation and is superior to computed tomography for the detection of acute ischaemia. We sought to evaluate the evidence that conventional MRI influences doctor management or patient outcomes in routine care.

Methods: We systematically searched PubMED, EMBASE and proceedings of the International Stroke Conference. Studies were included if they included patients presenting with possible stroke syndromes and they reported MRI results and resulting changes in management or outcome. Multiple reviewers determined inclusion/exclusion for each study, abstracted study characteristics and assessed study quality.

Results: Of 1813 articles screened, nine studies met inclusion criteria. None were randomized controlled trials, cohort studies or case-control studies. We found little evidence that MRI affects outcomes - one single-centre case series presented three patients. The remaining articles were studies of diagnostic tests or vignette-based studies that described changes in doctor management attributed to MRI. In the studies that suggested MRI influenced management, it did so in two ways. First, MRI distinguished stroke from mimics (e.g. brain tumours), thus enabling more appropriate selection of therapies. Second, even when MRI confirmed a suspected stroke diagnosis, it sometimes provided information (on stroke mechanism, localization, timing or pathophysiology) that influenced management.

Conclusions: The impact of MRI on management and outcomes in stroke patients has been inadequately studied. Further research is needed to understand how MRI may productively affect stroke management and outcomes.