To determine the interrater agreement for measurements of vertebral artery origin stenosis using NASCET, ECST, WASID, VOTE, and Qureshi methods.

Large artery atherosclerosis (LAA) in the posterior circulation remains understudied.  Approximately 20-25% of posterior circulation stroke (PCS) is attributed to LAA, primarily occurring at the vertebral artery origin (VAO).  The optimal methodology for measuring VAO stenosis, however, is undetermined.

We retrospectively reviewed digital subtraction angiography images for 20 consecutive patients from the University of Minnesota Posterior Circulation Stroke Database. Two raters measured the degree of VAO stenosis using NASCET, ECST, WASID, VOTE, and Qureshi methods. A Bland-Altman analysis was performed to quantify the reproducibility coefficient (RPC), bias, limit of agreement (LoA), and coefficient of variation (CV%) between the two raters.  To further assess agreement between methods, the Cohen’s Kappa (k) was calculated using a 4-point discrete scale (no stenosis, mild [1-49%], moderate [50-69%], and severe [70-99%]).

A total of 27 measurements of VAO stenosis were performed (7 patients had bilateral measurements).  Bland-Altman analysis demonstrated highest interrater agreement for NASCET and VOTE with a CV% of 20% and 22%, and LoA between 12 to -7.8 and 13 to -8.6, respectively. Additionally, NASCET (RPC=10) and VOTE (RPC=11) demonstrated the strongest reproducibility. Cohen’s Kappa (k) demonstrated “substantial” agreement for NASCET (k=.72), VOTE (k=.66), WASID (k=0.72), and Qureshi (k=0.73), and “moderate” agreement for ECST (k=0.47).

Our study reveals that NASCET and VOTE methodologies have the best CV%, RPC, and LoA for measuring VAO stenosis.  Other than ECST, all methodologies had substantial agreement using Cohen’s Kappa (k).  Our findings are applicable to PCS epidemiology studies in addition to future medical and interventional trials for secondary stroke prevention in LAA-mediated PCS.