Background/Purpose:

EULAR recommendations suggest diagnostic imaging in all GCA suspects. Vascular ultrasound (US) is cheap, readily available and the recommended first line examination in cranial GCA (c-GCA). Hence, US is an attractive first line examination also in large vessel GCA (LV-GCA). However, lower incidence of LV involvement is reported in US studies than in PET studies indicating a lower diagnostic sensitivity of US.

In a prospective study of glucocorticoid-naïve patients suspected of new-onset GCA, we evaluated the diagnostic accuracy of axillary artery US in the diagnosis of LV-GCA using 18F-FDG PET/CT as reference standard.

Methods:

Patients suspected of GCA were consecutively considered for inclusion. Inclusion criteria were: 1) age ≥50 years; 2) CRP>15 mg/L or ESR>40 mm/h; 3) either a) cranial symptoms, b) new-onset claudication c) protracted constitutional symptoms d) polymyalgia rheumatica (PMR) symptoms. Main exclusion criteria were: 1) recent or ongoing glucocorticoid or DMARD treatment; 5) previous diagnosis of GCA or PMR ; 6) large vessel inflammation mimicking LV-GCA.

Clinical evaluation and imaging was performed before treatment initiation. The reference diagnosis for LV-GCA was a clinical diagnosis of GCA and a 18F-FDG PET/CT revealing aortic and/or subclavian/axillary artery FDG uptake > liver uptake. Patients not diagnosed with GCA were considered controls. US was performed by experienced sonographers, blinded to PET results. Axillary arteries were assessed for the presence or absence of the ‘halo sign’ and intima media thickness (IMT) was measured. Sensitivity and specificity of the halo sign in axillary arteries was evaluated. ROC curve analysis was performed to estimate axillary IMT cut off.

Results:

86 patients were included (97 screened). 45 were diagnosed with LV-GCA (with or without concomitant c-GCA), 10 with isolated c-GCA, 21 with PMR and 10 with other diseases. Baseline characteristics of LV-GCA and controls are shown in table 1.

None of the controls had a positive axillary US, whereas 36/45 LV-GCA patients were axillary US positive yielding a specificity of 100% (95%CI: 89-100%) and a sensitivity of 80% (95%CI: 65-90%)). Of the 73 PET positive axillary arteries in LV-GCA patients, 53 were axillary US positive (sensitivity 72% (95%CI: 61-83%)). Four PET negative axillary arteries were US positive (specificity 95% (95%CI: 87-99%)). An AUC of 0.86 (95%CI: 0.79-0.92) was obtained by ROC curve analysis of axillary IMT with axillary PET diagnosis as a reference. An IMT cut off value of 0.9mm revealed a sensitivity of 74% and a specificity of 92%.

Conclusion:

In the hands of experienced sonographers, axillary arteries US shows high sensitivity and specificity for the diagnosis of LV-GCA which clearly suggests US as a first line imaging test in LV-GCA suspected patients. Suggested IMT cut off confirms findings of previous studies using different reference standards.

« Back to 2018 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/ultrasound-diagnosis-of-large-vessel-inflammation-in-new-onset-treatment-naive-gca-patients-using-fluorine-18-fluorodeoxyglucose-pet-ct-as-the-reference-standard-a-prospective-study-of-86-pa/