Reliability and accuracy of 3-mm and 2-mm maximum intensity projection CT angiography to detect intracranial large vessel occlusion in patients with acute anterior cerebral circulation stroke

Purpose To evaluate the reliability and accuracy of thick maximum intensity projection (MIP) CTA images to detect large-vessel occlusion (LVO) in the anterior circulation in patients with acute stroke. Methods A total of 140 acute stroke patients (41 with and 99 without LVO) were evaluated by two neuroradiologists for LVO using axial 3-mm and 2-mm MIPs. Results Interobserver agreement was substantial using 3-mm MIPs (x138; = 0.67) and almost perfect using 2-mm MIPs (x138; = 0.82). Using 3-mm MIPs, sensitivities were 80.5% and 68.3%, with specificities of 98.0% and 96.0%. Using 2-mm MIPs, sensitivities were 82.9% and 73.2%, with specificities of 98.0% and 99.0%. Sensitivity and specificity of 3 mm and 2 mm MIPs were not statistically significantly different (P >= 0.375). The majority of LVOs in the distal intracranial carotid artery, and/or M1-segment were correctly identified: 96.0% (observer 1, 3-mm MIPs), 88.0% (observer 2, 3-mm MIPs), 96.0% (observer 1, 2-mm MIPs), and 96.0% (observer 2, 2 mm MIPs). Using 3-mm MIP images, observers 1 and 2 missed 7/15 (46.7%) and 9/15 (60.0%) of isolated M2-segment occlusions, respectively. Using 2-mm MIP images, observers 1 and 2 missed 5/15 (33.3%) and 6/15 (40.0%) of isolated M2-segment occlusions, respectively. Conclusion Thick (2-3 mm) axial MIPs are not useful to detect proximal LVO in the anterior circulation.