Abstract Body: Background: Ischemic stroke is a leading causes of death and disability and imaging is essential when determining treatment. Currently, both computed tomography (CT) and magnetic resonance (MR) are accepted as options for first imaging of stroke. Whether MR or CT is more advantageous for first stroke imaging has yet to be determined in a randomized study. The goal of this study was to determine feasibility of randomizing code stroke patients to MR or CT.

Methods: Multisite, randomized, prospective study of code stroke patients presenting within a 12-week window to 4 certified stroke centers. Hospital-level cluster randomization assigned each site 6 CT-first weeks, and 6 MR-first weeks. Patients ≥18 years presenting with stroke symptoms < 24 hours with active code strokes at time of first imaging were included. Patients transferred from another hospital or who received prior imaging at an outside facility were excluded. Demographics, clinical stroke variables, and workflow metrics were extracted from the local stroke database or patient electronic health records. A univariate logistic regression model was used to evaluate the primary outcome: compliance (i.e. proportion of patients scanned according to assigned imaging). We hypothesized compliance would be comparable to that seen when MR-first was the preferred standard of care, ≥60%, demonstrating feasibility.

Results: 406 patients (199 females; mean age 67 years, range 24 - 103) were included in the analysis (Table 1). Compliance with assignment to CT was 90%, compliance with MR was 66%. Those assigned to MR were significantly less likely to be scanned as assigned (OR: 0.21, 95% CI [0.12-0.36]). Reasons for non-compliance included both process-related (e.g. MR scanner in use) and patient-related reasons (e.g. medically unstable). Most frequently, the reason for non-compliance was not documented.

Conclusion: This study is the first step in evaluating feasibility for a large-scale randomized clinical trial to determine whether MR or CT is preferable as the first stroke imaging modality. Compliance with assignment (MR or CT) our preset threshold of 60%, with significantly higher compliance when CT was assigned compared to MR. With mitigation of process-related barriers to randomization compliance, these results inform next steps in optimizing a future trial.