Large Vessel Occlusion Score: A Screening Tool to Detect Large Vessel Occlusion in the Acute Stroke Setting

BACKGROUND

The results of recent trials of mechanical thrombectomy for acute ischemic stroke have increased the demand for identification of patients with large vessel occlusion (LVO) at the primary stroke center, where a prompt detection may expedite transfer to a comprehensive stroke center for endovascular treatment. However, in developing countries, a noncontrast computed tomography (NCCT) may be the only neuroimaging modality available at the primary stroke center scenario, what calls for a screening strategy accurate enough to avoid unnecessary transfers of noneligible patients for endovascular therapy. Algorithms based on National Institute of Health Stroke Scale (NIHSS) and NCCT findings can be used to screen for LVO in patients with anterior circulation stroke (ACS).

OBJECTIVE

To test the accuracy of a score based on NIHSS and NCCT to detect LVO in patients with ACS.

METHODS

We evaluated 178 patients from a prospective stroke registry of patients admitted to an academic tertiary emergency unit. NIHSS and vessel attenuation values of the middle cerebral artery on NCCT absolute vessel attenuation (VA) were collected by 2 investigators that were blind to CT angiography (CTA) findings. We used receiver operating characteristics curve analysis and C-statistics to predict LVO on CTA.

RESULTS

NIHSS and vessel attenuation were highly associated with LVO with an area under the curve (AUC) of .86 and .77. The LVO score, built by logistic regression coefficients of the NIHSS and VA, showed the highest accuracy for the presence of LVO on CTA (AUC of .91).

CONCLUSION

The LVO score may be a useful screening approach to identify LVO in patients with ACS.

Abstract TP213: Accuracy of Transcranial Color-coded Sonography for the Detection and Measurement of Intracranial Arterial Stenosis When Compared to CT Angiography in Patients With Acute Ischemic Stroke and Transient Ischemic Attack

Introduction: Transcranial Color-Coded Sonography (TCCS) including B-mode imaging and color-coded Duplex is non-invasive, bedside and inexpensive tool for the assessment of the intracranial circulation. TCCS can be used to assess intracranial stenosis but its accuracy has not been systematically compared to CT angiography in acute stroke patients.

Hypothesis: TCCS has high accuracy when compared to CT angiography for the detection and measurement of intracranial arterial stenosis in patients with acute ischemic stroke.

Methods: In this prospective study 43 patients (67% male; mean age: 59.8 ± 11 years) with acute ischemic stroke or Transient Ischemic Accident (TIA) underwent examinations of the intracranial arteries by TCCS and CT angiography. Different examiners that were blinded to test results performed the exams. TCCS and CT angiography were performed during the hospital stay. All stenotic segments of intracranial vessels were classified into 4 groups: normal or mild (<50%), moderate (50%-70%), severe (≥70%) and subocclusion/occlusion.

Results: 580 of 645 arterial segments in 43 patients were evaluated. The sensitivity, specificity, positive predictive value, and negative predictive value of TCCD in comparison with CT angiography for the detection of moderate to severe intracranial stenosis (≥50%) were 81.5%, 100%, 100%, and 76% respectively. The Cohen’s kappa between TCCS and CT angiography was 0.76 (p<0.001).

Conclusions: TCCS is a bedside non-invasive tool that can be used to evaluate intracranial arteries with high accuracy for moderate to severe stenosis when compared to CT angiography. Moderate agreement was found between TCCS and CT angiography for the evaluation of intracranial stenosis in patients with acute ischemic stroke or TIA.

Abstract WP214: Transcranial Color Coded Sonography versus Transcranial Doppler for the Detection of Right-to-left Shunt and Patent Foramen Ovale in Young Patients With Stroke

Introduction: Patent Foramen Ovale (PFO) is the most common right-to-left shunt (RLS) and is often found in young patients with stroke related to paradoxical embolism. Contrast-Enhanced Transesophageal Echocardiography (TEE) is considered a gold standard to visualize PFO. Transcranial Doppler (TCD) with bubble test is often used to detect RLS with good correlation to TEE for the diagnostic of PFO. More recently, Transcranial Color Coded Sonography (TCCS) which included B-mode and color coded imaging has overcome TCD in many clinical applications but the accuracy of TCCS for detection of RLS and PFO has not been systematically evaluated.

Hypothesis: To determine if the TCCS is an accurate tool to identify both PFO and RLS.

Methods: We investigate 106 patients with ischemic stroke under 55 years-old admitted from 2012 to 2014 in a tertiary academic hospital. Patients were evaluated with TEE, TCD and TCCS, and all exams included a saline bubble test. The examiners were blinded for the other tests results. Kappa agreement was calculated inter-examiners for TCCS and TCD. Accuracy of TCCS was calculated in comparison to TEE.

Results: We detected a RLS in 54 (50.9%) patients (age mean 43.9 ± 8.2) with kappa agreement 0.92 (95%CI 0.78-1.0) when performed TCCS and TCD. TEE and TCSS were positive in 23/98 (23.4%) and TEE did not reveal contrast in 20/98 (20.4%). In 30 (28.3%) patients only TEE revealed a PFO. TCSS had a sensitivity of 88.4% (95%CI 0.68-0.97), specificity of 72.2% (95%CI 0.60-0.81) and positive likelihood ratio of 3.18 (2.14-4.73) of the diagnosis of PFO.

Conclusion: TCCS and TCD had an excellent agreement. TCCS has a good accuracy for the detection of PFO and RLS in young patients with stroke.

Abstract TP46: A Score to Detect Proximal Artery Occlusion in Patients With Acute Ischemic Stroke of the Anterior Circulation Based on NIHSS and Non-contrast Brain CT: the PAO Score

Introduction: After the results of the new endovascular trials that demonstrated a robust effect of endovascular treatment for acute ischemic stroke (AIS), early detection of proximal artery occlusion (PAO) has become a fundamental task during the initial assessment of acute stroke patients at the emergency department. Nevertheless, an accurate identification of PAO may be particularly challenging in smaller hospitals and in developing countries, areas with restricted assess to vascular neuroimaging modalities such as CTA and MRA.

Hypothesis: Algorithms based on NIHSS and non-contrasted CT (NCCT) findings can be accurately used to detect PAO in patients with AIS of the anterior circulation.

Method: We retrospectively evaluated 194 consecutive patients with AIS of the anterior circulation from a prospective stroke registry of patients admitted to an academic tertiary emergency unit in Brazil during 2014 that had a NCCT and a CTA at admission. NIHSS scores and attenuation of major intracranial arteries of the anterior circulation on NCCT were collected by two experienced investigators that were blind to the CT angiography findings. We used a ratio between two ROIs (rVA) that were drawn on NCCT blinded to CT angiography: (i) on the region of highest vessel attenuation ipsilateral to the involved hemisphere and (ii) mirror ROI on the corresponding vessel segment of the contralateral hemisphere. We used ROC curve analysis and C-statistics to predict CT angiography PAO.

Results: NIHSS and vessel attenuation values were highly associated with the PAO with an area under the curve (AUC) of 0.88 (p < 0,001) and 0.83 (p < 0,001), respectively. An NIHSS of 10 at admission had a sensitivity, and negative predictive value of 97% and 97%, respectively. The rVA ≥ 1.50 had a specificity and positive predictive value 96% and 85%, respectively. The POA score was then built by logistic regression from NIHSS and rVA and showed even higher accuracy for the presence of POA on CTA, with an AUC of 0.93 (p < 0,001).

Conclusion: The PAO score based on admission NIHSS and proximal vessel attenuation on NCCT can be accurately used to detect PAO in patients with AIS of the anterior circulation. Further studies are necessary to validate this score in a multicenter setting.