Background/Purpose: Ultrasound
and Magnetic resonance angiography (MRA) yields a high sensitivity and
specificity regarding the diagnosis of cranial giant cell arteritis (GCA).
Ultrasound and MRA can also depict extracranial large vessel vasculitis (LVV)
in both GCA and Takayasu arteritis (TA) patients. Until now, no studies have
examined the interobserver agreement of the ultrasonographic and MRA findings
in LVV patients. Hence, the aim of this study was to examine the interobserver
agreement of the ultrasound and MRA examination of temporal arteries and large
vessels in LVV patients.

Methods: This study is a
part of the MUSES project (Magnetic resonance angiography vs ultrasonography in
Systemic large vessel vasculitis), a prospective cross-sectional study.
Patients who were diagnosed with LVV by ultrasound, MRA or CTA were identified
and included in the study at the Department of Rheumatology, Hospital of Southern Norway Trust in Kristiansand from January 2014 to January 2015. One
ultrasonographer experienced in vascular ultrasound (APD) examined and
recruited the LVV patients. MRA imaging was performed within one week after the
ultrasound evaluation and one radiologist evaluated the MRA images locally
(FL). The common temporal, temporal parietal branch, temporal frontal branch ,
carotid , subclavian , vertebral , axillary arteries, and thoracic aorta were
scanned in all patients. Cranial arteries were assessed in T1-weighted
high-resolution transverse slices post contrast, and the aorta and supraaortic
arteries based on coronal MRA images. Films and images of both ultrasound and
MRA evaluation of every artery were recorded, and two experts; one for vascular
ultrasound (WAS) and one for MRA (JG) surveyed the data. Both experts were
blinded to clinical and laboratory data. To calculate the inter-observer
agreement between ultrasonographers (WAS, APD) and radiologists (FL, JG) the
Cohen’s kappa test has been used.

Results: Twenty-seven patients were included in this
study [(12 males, 15 females, median
age 69 years (IQR 62-76)]. Twenty-five patients were diagnosed with GCA and 2
with TA. Eight patients had new onset LVV and 19 had long lasting disease
(median disease duration 2 years, IQR 0-5). Median CRP was 8 mg/l (IQR 3-24)
and ESR 24 mm/hr. (IQR 9-53). None
of the patients had affection of the vertebral artery and thoracic aorta on
MRA.

The inter-observer agreement for
the various arteries for both ultrasound and MRA are reported in table 1.

Conclusion: Ultrasonographic
findings of temporal, aorta and supraaortic arteries in patients with LVV
appears to be highly detectable and can be recorded in films for further
evaluation. The low interobserver reliability in
MRA could be partly explained by the sequences used for vascular imaging.

Table.
1