Abstract TP65: Perfusion Artifacts May Overestimate Thrombectomy Eligibility in Patients With Acute Large Vessel Occlusion

Background: Automated computed tomography perfusion (CTP) is effective at identifying salvageable penumbra in patients with acute ischemic stroke due to large vessel occlusion (LVO). Perfusion imaging is now recommended for the selection of thrombectomy candidates in the extended window (>6 hours) which may incorporate the penumbra mismatch ratio (T max >6s:rCBF<30%) >1.7. However automated perfusion output may include artifactual findings that are discordant with a patient’s symptoms and vessel imaging, overestimating the tissue at risk. Hypothesis: Artifactual findings on automated CTP penumbral assessment (T max >6s) will misclassify thrombectomy eligibility in some patients. Methods: We reviewed a retrospective multi-site consecutive cohort of patients undergoing CTP for suspected acute ischemic stroke ≤24 hours from onset (6/2017-12/2017). The primary outcome was the volume of any discordant T max >6s abnormality on RAPID automated perfusion CT imaging, calculated by two independent readers using manual assessment (ImageJ, NIH). The discordant penumbral volume was compared to the automated output and corrected mismatch ratios were generated. Results: Of 410 consecutive patients who underwent CTP, 60 (14.6%) had acute LVO of the anterior circulation and were included. The median age was 78 years (IQR 64-84), 36 (60.0%) were female, and 25 (41.7%) were White. Fifteen patients (26.3%) had T max >6s abnormalities that were discordant with the clinical symptoms or anticipated vascular distribution, with strong inter-rater agreement (r 2 =0.927). The median difference between manual and automated T max >6s volumes was 18.4cc (IQR 10.80-35.77) accounting for 11.3% of the automated volume (IQR 7.8-46.7%). Following manual recalculation of the T max >6s volume, 1 patient was reclassified as having an “unfavorable” mismatch ratio (1.7%, 95%CI 0.2-11.5%). Conclusion: About one quarter of patients had discordant penumbral imaging on automated CTP, which may lead to misclassification of thrombectomy eligibility. While artifactual findings are reliably identified by trained raters, our results emphasize the need to evaluate CTP results with knowledge of the patient’s clinical symptoms and vascular imaging.