Abstract TP50: Non-contrast CT Maximum Intensity Projections (MIPs) for the Detection of Large Vessel Occlusion (LVO)

Abstract

Background: Identification of LVO is becoming increasingly important with recent guidelines supporting endovascular therapy in selected patients with acute ischemic stroke (AIS). Non-contrast CT scans are generally used to assess for stroke or hemorrhage in acute stroke code evaluations. CT angiograms (CTA) can be used to assess for LVO but require additional time and contrast administration. Some hospitals lack the resources to acutely obtain and interpret CTAs. Non-contrast CT MIPs may be an alternate modality to detect LVO. Methods: We retrospectively reviewed patients with AIS in our UCSD Stroke Registry, presenting between 6/2014-7/2016 that received a MIPs, generated from non-contrast CT scans, and a CTA. MIPs were evaluated by a group of stroke specialists (3 Faculty, 2 Fellows, 2 Acute Care Practitioners [ACPs]). No clinical information was provided. Gold standard comparison was to CTA. Results were stratified by subgroup based on level of training. Inter-rater agreement was assessed using Fleiss’ Kappa Coefficient. Results: 24 scans were reviewed (12 with LVO, 12 without LVO). Using MIPs for the detection of LVO, the Faculty subgroup had a sensitivity and specificity of 81% and 86% for any LVO, 95% and 92% for ICA/M1, 42% and 100% for M2 and 67% and 96% for basilar occlusions. The Fellows and ACPs subgroup had a sensitivity and specificity of 77% and 81% for any LVO, 61% and 94% for ICA/M1, 19% and 83% for M2 and 75% and 95% for basilar occlusions. Inter-rater agreement among Faculty readers was k=0.58 for the detection of any LVO, k=0.75 for ICA/M1, k=0.79 for M2 and k=0.14 for basilar occlusions. Among Fellows and ACPs, k=0.48 for any LVO, k=0.57 for ICA/M1, k=0.40 for M2, and k=0.27 for basilar occlusions. Conclusions: Non-contrast CT MIPs have high sensitivity and specificity for the detection of LVO when compared to CTA. Inter-rater agreement between readers of MIPs is good. Better results in the Faculty subgroup were likely due to more experience reviewing imaging. These results support the use of non-contrast CT MIPs for the detection of LVO to both save time and contrast exposure in patients with AIS. Further studies investigating a larger dataset, optimizing MIP parameters, and examining the utility/cost-effectiveness of this technique are being pursued.