Accuracy of Distal Internal Carotid Artery Contrast Ratio to Infer Proximal Carotid Disease in the Mobile Stroke Unit
James L Barker1,2, Oshi Swarup3, Hannah T Johns1,2
1Department of Neurology, Melbourne Brain Centre (J.L.B., H.T.J., L.C., A.H.B., C.W., B.Y., A.B., F.L., H.Z., G.A.D., S.M.D., B.C.V.C.), The Royal Melbourne Hospital, The University of Melbourne, Parkville, Victoria, Australia.
Background:
Extracranial internal carotid artery (ICA) stenosis is a common cause of ischemic stroke and transient ischemic attack. Cervical computed tomography (CT) angiography can rapidly diagnose carotid pathology but is not possible with current-generation CT scanners used in mobile stroke units. We assessed the diagnostic accuracy of the intracranial ICA contrast density ratio (CDR) to identify extracranial ICA stenosis.
Methods:
We retrospectively reviewed consecutive patients at 3 comprehensive stroke centers in Melbourne, Victoria, who received CT angiography on the mobile stroke unit CereTom CT scanner, with reference standard in-hospital CT angiography or digital subtraction angiography in the Melbourne mobile stroke unit database. The mean Hounsfield units at 4 segments of the intracranial ICA, both ipsilateral and contralateral to the symptomatic hemisphere, were measured. The CDR was calculated using the formula: CDR=contralateral mean Hounsfield units/ipsilateral Hounsfield units. Statistical analysis was completed with univariate analysis of CDR, and leave-one-out cross-validation with receiver operating characteristic area under the curve to determine CDR thresholds associated with stenosis or occlusion, to determine sensitivity and specificity of the test.
Results:
Of 151 patients included, 31 of 151 (20.5%) had extracranial ICA stenosis >50%, and the average CDR was higher in patients with >70% stenosis with threshold ratio of >1.24 (sensitivity, 80%; specificity, 88%; positive predictive value, 57%; and negative predictive value, 96%) or extracranial occlusion (sensitivity, 100; specificity, 82%; positive predictive value, 29%; and negative predictive value, 100%). The threshold of 1.18 for 50% to 99% and carotid occlusion had lower sensitivity (sensitivity, 78%; specificity, 95%; positive predictive value, 88%; and negative predictive value, 91%).
Conclusions:
Assessment of the intracranial ICA CDR may assist in ruling out high-grade extracranial stenosis in patients who are being considered for direct transfer for endovascular thrombectomy after head-only CT angiography.
