Vascular Ultrasound in the Evaluation of Large-Vessel Vasculitis

Abstract

TITLE: Distribution patterns and correlations among affected vessels on vascular ultrasound for the evaluation of giant cell arteritis Purpose: Vascular ultrasound (VUS) is a sensitive and specific method for evaluating giant cell arteritis (GCA). Understanding patterns of temporal artery (TA) and large artery involvement may be important for aiding in diagnosis. The purpose of this work is to determine correlations among VUS abnormalities in patients with GCA and to categorize distinct patterns of disease. Methods: We performed a retrospective study among 139 GCA patients who underwent VUS at a large academic medical center, 2013-2017. Descriptive statistics were used to summarize baseline characteristics. The VUS protocol visualized the right and left common superficial TAs (STA) and their frontal (TA-F) and parietal branches (TA-P), as well as the common carotid (CCA), subclavian (SCA), and axillary arteries (AXA). The STA, TA-F, and TA-P were abnormal if the “halo sign” or hyperechoic wall thickening was present. The CCA, SCA and AXA were abnormal if halo sign, hyperechoic wall thickening, stenosis, and/or occlusion were present. VUS studies were interpreted as consistent with acute arteritis, no arteritis, or hyperechoic wall thickening without acute arteritis. The Phi coefficient (j) was used to determine correlations between abnormal vessels. Significance was determined using Fisher’s exact test with Bonferroni correction. A sensitivity analysis was performed to specifically examine correlations between vessels with a halo sign. Agglomerative hierarchical clustering was used to identify clusters of affected vessels with any abnormality and with the halo sign alone. Results: Among 139 GCA patients, 50.4% had a history of GCA or biopsy-proven aortitis prior to VUS with median disease duration 16.9 (IQR 5.0, 48.8) months. The remainder had no prior history of GCA and were diagnosed post-VUS. The most common signs/symptoms at time of evaluation were an elevated ESR or CRP level (82.0%), headache (49.6%), and fatigue (42.5%). Jaw claudication was present in 18.7%. 70.5% were taking glucocorticoids at time of VUS. Acute arteritis was present on VUS in 32.4% and hyperechogenicity without acute arteritis was present in 15.1%. The majority of studies showed no arteritis. The most commonly affected vessels were the left SCA (19.4%) and left AXA (17.3%). The halo sign was most frequent in the left STA (13.0%). Abnormalities in the STA, TA-F, and TA-P were moderately correlated ipsilaterally (j = 0.46-0.62, p<0.0001) and contralaterally (j = 0.51-0.58, p<0.0001). Abnormalities in the AXA and SCA were strongly correlated with the ipsilateral and contralateral AXA and SCA (j = 0.64-0.82, p<0.0001). Correlation between right and left TAs was moderate (j = 0.59, p<0.0001), while a stronger relationship was found between right and left large arteries (j = 0.77, p<0.0001). Abnormalities in TAs were not correlated with large arteries (j » 0 for all comparisons; not significant). Affected vessels with any abnormality fell into three distinct clusters (C). C1 contained the bilateral CCAs. C2 included the bilateral SCAs and AXAs. C3 was composed of the cranial vessels. Clustering vessels based on the halo sign alone yielded similar results with the main differences being the distance between C1 and C3, as well as the distances between vessels within individual clusters. Conclusion: We observed moderate correlations of TA branches and strong correlations of large artery branches on VUS among GCA patients. Abnormalities in the TAs and large arteries were not correlated. Agglomerative hierarchical clustering revealed three distinct patterns of distribution. This work underscores the importance of scanning both cranial and large vessels with VUS and has implications for future recommendations regarding the use of imaging in the evaluation of large vessel vasculitis (LVV).