Optimizing Transcranial Doppler Screening for Vertebro-Basilar Stenosis

Abstract

109 Background and Purpose: Intracranial vertebrobasilar (VB) stenosis carries a high risk of stroke. Transcranial Doppler (TCD) has lower sensitivity and specificity for VB lesions compared to anterior circulation. We performed a case-controlled study to develop additional criteria to optimize predictive value of TCD velocity criteria for VB stenosis. Methods: We performed TCD in consecutive patients with stroke or TIA symptoms using a standard insonation protocol and a priori established mean flow velocity (MFV) criteria for VB stenosis (≥50 and ≥60 cm/sec cut-offs). Patients who had abnormal VB TCD findings and angiography were selected. Age-matched patients with normal angiograms (DSA) who also had TCD were used for comparison. Results: Of 1289 patients screened, 115 had abnormal VB findings on TCD (9%). Of these, 50 patients had angiography including 24 DSAs. A total of 24 age-matched controls with TCD and DSA were identified among patients with normal VB results. When MFV criteria were applied alone, MFV cut-offs had sensitivity of 92%, specificity 44%, PPV 44%, NPV 92%. False-positive MFV elevation occurred mostly due to flow collateralization with carotid or VB occlusions (64% of false positive studies). The only two false negative TCD results were found with long irregular terminal VA plaques. When additional criteria for flow collateralization and focal MFV elevation (stenotic : pre- or post-stenotic VB MFV ratio ≥2 : 1) were used, TCD sensitivity was 82%, specificity 92%, PPV 82%, NPV 92%. Conclusions: Criteria for flow collateralization and velocity ratios can reduce the number of false positive TCD findings in patients with VB stenosis when velocity cut-offs are used alone. Although limited to a single center, our study suggests that the predictive value of TCD can be optimized to allow accurate noninvasive screening of patients for isolated VB stenoses.