Abstract Body: Introduction
The prevalence of Mismatch (MM) profiles on admission perfusion imaging in large core ischemic strokes, stratified by time since last known normal (LKN) and MM profile definition, is not well described.

Methods
We used a retrospective multicenter study of patients with anterior circulation large vessel occlusion, baseline MR or CT perfusion (MRP, CTP), and definitive LKN within 24 hours. We studied four MM profile definitions: 1) Mismatch ratio (MR)≥1.2, penumbral volume (Pen)≥10ml, and (if CTP) core volume defined by CBF<30%; 2) MR≥1.8, Pen≥15ml, CBF<30%; 3) MR≥1.2, Pen≥10ml, CBF<38%; 4) MR≥1.8, Pen≥15ml, CBF<38%. Two definitions of large core were evaluated (≥50ml or ≥70ml; defined as CBF<30% or CBF<38% on CTP and as ADC ≤620x10^-6mm²/s on MRP). We used the Cochran-Armitage Trend Test to evaluate for trends in MM presence over time and McNemar Test to compare MM profiles for differences in prevalence.

Results
Of 2012 patients included, 49% were male, median age was 75 (IQR 64-84), median baseline NIHSS was 15 (8-20), median time from LKN to imaging was 5.7 hours (1.8-8), 44% had CTP. In patients with cores ≥50ml and ≥70ml (Figures 1 & 2), MM profile defined by MR≥1.2 and Pen≥10ml was more prevalent than MM defined by MR≥1.8 and Pen≥15ml (P<0.001). MM prevalence decreased over time in three of four MM profiles in the ≥50ml core group but only one MM profile in the ≥70ml group (P<0.001 in all comparisons). In all core sizes (Figure 3), there was a significant trend toward decreased MM prevalence over time in all four profiles (P<0.0001) and between the MR≥1.2 and Pen≥10 vs MR≥1.8 and Pen≥15 profiles (P<0.001 in all comparisons). Similar significant differences between the MR≥1.2 and Pen≥10 vs MR≥18 and Pen≥15 existed in CTP only, MRP only, and witnessed onset only subgroup analyses (P<0.001 for all).

Conclusions
MM is more prevalent when defined as MR≥1.2 and Pen≥10ml than when defined as MR≥1.8 and Pen≥15ml, especially in large cores. MM prevalence decreases over time in all core sizes, though trends are less clear with increasing infarct size. Understanding real-world prevalence of MM in large core strokes may help contextualize results of recent trials, while clarifying the optimal definition and prevalence of MM in the large core population is of import to future prospective study design. This study is limited by potential bias in favor of higher MM presence due to one site’s patient transfer acceptances contingent on favorable CTP profiles.