Background Extracranial internal carotid artery (ICA) occlusion can be overestimated on emergent single phase CT angiography (CTA) of stroke patients with isolated intracranial ICA occlusion. We aimed to measure the ability of identifying the extracranial site of presumed tandem ICA occlusions on pre-procedural CTA relative to catheter angiography during acute endovascular stroke therapy.
Methods Retrospective study of patients with intracranial ICA occlusion, with or without extracranial ICA occlusion, who underwent single phase CTA before acute endovascular treatment. Two neuroradiologists reviewed CTA images for the presence or absence of extracranial ICA occlusion, blinded to the catheter angiography results. The sensitivity, specificity, and predictive values of presumed extracranial ICA occlusions on CTA were calculated in reference to catheter angiography.
Results 91 stroke patients with acute intracranial ICA occlusion met the inclusion criteria for the study. 24% of patients (22/91) had tandem ICA occlusion confirmed on catheter angiography. Single phase CTA had a sensitivity of 95.5% (95% CI 77.2 to 99.9%) and a specificity of 69.6% (95% CI 57.3 to 80.1%) for concomitant extracranial ICA occlusion (false positive rate 30.4%). The positive and negative predictive values of single phase CTA for extracranial ICA occlusion in the presence of a distal ICA occlusion were 50% (95% CI 34.2 to 65.8%) and 98% (95% CI 89.1 to 100%), respectively.
Conclusions Emergency single phase CTA is highly sensitive but has reduced specificity to identify extracranial ICA occlusion in patients with intracranial ICA occlusion, which may confound planning for acute endovascular stroke therapy and cause over exclusion of patients with isolated ICA terminus occlusion from clinical trials.
- ct angiography
Statistics from Altmetric.com
Acute tandem internal carotid artery (ICA) occlusion, defined as simultaneous occlusion of the extracranial ICA and intracranial ICA or middle cerebral artery (MCA), carries a poor prognosis if not treated with endovascular therapy due to low rates of arterial recanalization with or without intravenous tissue plasminogen activator (tPA).1 Some authors advocate concomitant angioplasty or stenting of the extracranial ICA before or after thrombectomy of the intracranial artery for improved revascularization, but the best approach has not yet been established.2–4 Moreover, patients with tandem ICA occlusions were typically excluded from some clinical trials of endovascular therapy, such as Solitaire With the Intention For Thrombectomy as Primary Endovascular Treatment (SWIFT PRIME) and Interventional Management of Stroke 3 (IMS-3), due to concerns that concomitant intracranial and extracranial intervention would confound the assessment of benefit of intracranial embolectomy.5 6 As a consequence, there is a need for accurate non-invasive imaging to differentiate between tandem and intracranial ICA occlusions in potential candidates for endovascular therapy.
Although multi-modal CT imaging is available in comprehensive stroke centers, single phase CT angiography (CTA) remains widely used in the emergency setting for rapid diagnosis of acute ICA occlusion, planning for endovascular therapy, and clinical trial screening.3 CTA has been shown to be highly sensitive and specific for detection of isolated extracranial ICA occlusions.7 However, when the intracranial ICA is acutely thrombosed, single phase CTA can lead to misdiagnosis of total occlusion of the extracranial ICA, or pseudo-occlusion, because rapid image acquisition outruns the sluggish inflow of contrast due to distal ICA blockage.8–10 As a result, there can be overestimation of presumed tandem ICA occlusion on emergent CTA of patients with isolated intracranial ICA occlusion.9 11 This can confound planning for acute endovascular treatment and cause misclassification of patients with intracranial ICA occlusion during clinical trial screening.3
The aim of this study was to determine the ability to identify extracranial ICA occlusion on single phase CTA relative to catheter based angiography of patients with acute intracranial ICA occlusion who were treated with endovascular therapy.
Materials and methods
Patients were retrospectively selected from a prospective database of acute endovascular stroke therapy maintained at our center, which was approved by the institutional review board. Inclusion criteria consisted of adult patients presenting with acute stroke who underwent emergent CTA of the head and neck before endovascular intervention and had a diagnosis of intracranial ICA occlusion on catheter based angiography. Patients were excluded if source CTA imaging was missing or of poor quality. Two board certified neuroradiologists who were unaware of the clinical and catheter angiography data identified intracranial ICA occlusions with or without complete extracranial occlusions on source CTA images. A final diagnosis of isolated intracranial ICA occlusion versus tandem ICA occlusion on catheter based angiography was confirmed separately by an experienced neurointerventionalist (reference standard).
We collected patient data on demographics, National Institutes of Health Stroke Scale, Alberta Stroke Program Early CT Score (ASPECTS), time of stroke onset to angiography, and angiographic characteristics, such as intracranial ICA occlusion subtypes, to determine the salient clinical and radiographic profile of the study cohort.
CTA was obtained with 64 slice CT scanners using helical acquisition of 1.25 mm section thickness from the aortic arch through the circle of Willis after injection of 70–100 mL of contrast medium. CTA image acquisition was manually triggered by a technologist, targeting peak enhancement over a region of interest placed in the aortic arch. In general, the extracranial ICA was considered occluded if there was complete lack of contrast opacification over an arterial segment on CTA source images, as determined by the interpreting neuroradiologist.
Catheter based DSA was performed using the transfemoral approach. Selective common and internal carotid angiograms were obtained in the anterior–posterior, lateral, and oblique projections. In most cases, ipsilateral and contralateral common carotid angiograms performed into the late venous phase were used to capture delayed arterial flow. Extracranial ICA occlusion was typically determined with microcatheter exploration and selective injection if delayed common carotid sequences had failed to capture distal ICA opacification. The patency of the proximal MCA, anterior cerebral artery (ACA), and anterior communicating artery were noted. Terminus ICA occlusions were categorized as ICA I, L, or T occlusions, based on the extent of collateral perfusion downstream of the proximal ACA and/or MCA occlusions, as previously described.12
Inter-observer agreement for the presence or absence of complete extracranial ICA occlusion on single phase CTA was determined using kappa statistics. Differences in interpretation were resolved by consensus. Two by two contingency table analysis was used to determine the sensitivity, specificity, and predictive values of presumed extracranial ICA occlusion on CTA in reference to catheter angiography. Values were expressed with their 95% CI. Fisher’s exact test was used to test differences in categorical proportions from contingency table analysis. Statistical analysis was performed with GraphPad Prism 5.02.
We identified a total of 191 patients with an angiographic diagnosis of acute occlusion of the intracranial ICA between January 2002 and August 2014: 100 patients were excluded due to unavailability of source CTA images, typically because of outside transfers. The final analysis included 91 patients who had a median age of 66 years, median National Institutes of Health Stroke Scale score of 17 and ASPECTS of 8 (table 1). Median time of stroke onset to catheter angiography was 5.8 hours (IQR 3.8 to 10.3). Intracranial ICA occlusions were located in the ICA terminus of 87 patients (95.6%) and in the petrous or cavernous ICA segments of 4 patients (4.4%). There were 8/87 (9.2%) ICA I occlusions, 55/87 (63.2%) ICA L occlusions, and 24/87 (27.6%) ICA T occlusions. The prevalence of tandem ICA occlusions on catheter angiography was 22/91 (24%).
Overall inter-rater agreement was good for the presumed occlusion status of the extracranial ICA on pre-procedural CTA (kappa=0.80, 95% CI 0.68 to 0.92). Of the 22 patients with concomitant extracranial and intracranial ICA occlusions on catheter angiography, extracranial ICA occlusion was correctly identified on CTA of 21 patients (true positives) for a sensitivity of 95.5% (95% CI 77.2 to 99.9%) (table 2). The majority of these patients had focal thrombosis of the extracranial ICA origin with collapsed but patent downstream vessel segment to the level of the intracranial ICA terminus occlusion on reference angiography (figure 1). Of the 69 patients with isolated intracranial ICA occlusion on catheter angiography, extracranial ICA occlusion was correctly ruled out on the CTA of 48 patients (true negatives) for a specificity of 69.6% (95% CI 57.3 to 80.1%) (table 2). The false positive rate for extracranial ICA occlusion (pseudo-occlusion) on the CTA of patients with isolated intracranial ICA occlusion was 30.4%. Patients with proximal ICA pseudo-occlusion on emergent CTA often had sluggish flow but patent extracranial ICA on subsequent angiography (figure 2), except in 2/21 (9.5%) cases which had high grade stenosis and underwent acute endovascular stenting. The positive and negative predictive values for presumed extracranial ICA occlusion on single phase CTA of patients with isolated ICA occlusion on catheter angiography were 50% (95% CI 34.2 to 65.8%) and 98% (95% CI 89.1 to 100%), respectively (table 2).
While CTA has been shown to be highly sensitive and specific for isolated extracranial ICA occlusion alone,7 overestimation of cervical ICA occlusion on emergent CTA of patients with isolated intracranial ICA occlusion can be common.9–11 Grossberg et al reported 21 cases of extracranial ICA pseudo-occlusion on pre-procedural CTA of 46 patients who had intracranial ICA occlusion during acute endovascular intervention.10 Our data from a larger cohort supports the latter study and provides quantification of sensitivity (95%) and specificity (69%) of single phase CTA for estimation of extracranial ICA patency in patients with acute intracranial ICA occlusions.
The false positive rate of 30.4% for presumed extracranial ICA occlusions on CTA in our study possibly explains the difference in prevalence of isolated intracranial ICA occlusion across the recent clinical trials of endovascular stroke therapy because at least 1 in 3 of these cases may have been erroneously excluded as having tandem ICA occlusion during enrollment. In SWIFT PRIME, where occlusion of the extracranial ICA on CTA was an exclusion criterion, only 18% of patients treated with endovascular therapy had occlusion of the intracranial ICA.6 In comparison, the prevalence of intracranial ICA occlusion was respectively 25%, 28%, 27%, and 31% in Multicenter Randomized Clinical trial of Endovascular treatment for Acute ischemic stroke in the Netherlands (MR CLEAN), Endovascular Revascularization With Solitaire Device Versus Best Medical Therapy in Anterior Circulation Stroke Within 8 Hours (REVASCAT), Endovascular Treatment for Small Core and Proximal Occlusion Ischemic Stroke (ESCAPE), and Extending the Time for Thrombolysis in Emergency Neurological Deficits-Intra-Arterial (EXTEND-IA) which enrolled patients who had extracranial ICA occlusion on CTA.13–16 Because single phase CTA is also used for acute endovascular procedure planning, the low positive predictive value of 50% for extracranial ICA occlusion in our study suggests that patients could be at risk for unnecessary pretreatment with antiplatelet agents in anticipation of emergent stenting for presumed tandem ICA occlusions. Whether this could influence the rate of symptomatic hemorrhagic conversion in patients receiving both thrombolytic and endovascular therapies is uncertain. Future studies are needed to determine predictors of extracranial ICA pseudo-occlusion in patients with terminal ICA occlusion who are emergently evaluated with CTA and its potential impact on clinical outcomes.
One possible solution to increase CTA specificity for the extracranial site of presumed tandem ICA occlusions may be to use delayed phase sequences or dynamic CTA to extend detection time of sluggish contrast inflow in the ICA, as has been previously suggested.9 17 Potential problems with these approaches would be additional radiation exposure18 and further delay from diagnostic scan to endovascular recanalization. Future studies are needed to test the net benefit of these advanced CTA protocols for pre-procedural endovascular planning and patient selection in future trials of stroke due to intracranial ICA occlusion in the emergent setting.
The present study has some limitations, including potential selection bias associated with its retrospective design and our single center experience. Most patients had occlusion of the ICA terminus that could theoretically be associated with higher clot burden and greater occurrence of extracranial ICA occlusion. However, the prevalence of tandem ICA occlusions on catheter angiography in our cohort was 24%, which is within the prevalence range of 18.6–32.2% reported in the recent randomized clinical trials of endovascular therapy.13 14 Moreover, our findings are based on a cohort of patients with acute intracranial ICA occlusion who had similar demographics and stroke severity to those included in the recent clinical trials of endovascular stroke therapy.
This study measured the ability of identifying extracranial ICA occlusion on single phase CTA relative to catheter angiography in patients with acute stroke due to intracranial ICA occlusion. These data indicate that interpretation of single phase CTA has high sensitivity and relatively lower specificity for differentiating patients with isolated intracranial ICA occlusion from tandem ICA occlusions in the emergent setting. This has implications for interpreting the results of clinical trials of endovascular stroke therapy that may have over excluded patients with isolated terminal ICA occlusion from enrollment based on single phase CTA. From a clinical perspective, this information also needs to be taken into consideration for acute endovascular procedure planning. Prospective studies are needed to test the safety and feasibility of multi-modal CTA protocols in the emergent setting that can be more specific for identification of tandem ICA occlusions in acute stroke patients at greater risk for poor clinical outcomes.
Contributors Conception and design: MR, WTD, and TGJ. Data acquisition: all authors. Drafting the article: MR and TGJ. Critically revising the article: all authors. Final approval of the version to be published: all authors. Agreement of accountability for all aspects of the work: all authors.
Competing interests TGJ reports grant, non-financial, and other support from Neuravi (steering committee-modest), Codman Neurovascular (DSMB-modest), Stryker Neurovascular (PI DAWN-unpaid), and Fundacio Ictus (PI REVASCAT unpaid). Stock: Anaconda, Silk Road, Blockade Medical, Route 92, and FreeOx (modest).
Ethics approval The study was approved by the institutional review board.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.