NUK - logo
E-viri
Recenzirano Odprti dostop
  • Imaging in diagnosis, outco...
    Duftner, Christina; Dejaco, Christian; Sepriano, Alexandre; Falzon, Louise; Schmidt, Wolfgang Andreas; Ramiro, Sofia

    Rheumatic & musculoskeletal diseases open, 02/2018, Letnik: 4, Številka: 1
    Journal Article

    ObjectivesTo perform a systematic literature review on imaging techniques for diagnosis, outcome prediction and disease monitoring in large vessel vasculitis (LVV) informing the European League Against Rheumatism recommendations for imaging in LVV.MethodsSystematic literature review (until 10 March 2017) of diagnostic and prognostic studies enrolling >20 patients and investigating ultrasound, MRI, CT or positron emission tomography (PET) in patients with suspected and/or established primary LVV. Meta-analyses were conducted, whenever possible, obtaining pooled estimates for sensitivity and specificity by fitting random effects models.ResultsForty-three studies were included (39 on giant cell arteritis (GCA), 4 on Takayasu arteritis (TAK)). Ultrasound (‘halo’ sign) at temporal arteries (8 studies, 605 patients) and MRI of cranial arteries (6 studies, 509 patients) yielded pooled sensitivities of 77% (95% CI 62% to 87%) and 73% (95% CI 57% to 85%), respectively, compared with a clinical diagnosis of GCA. Corresponding specificities were 96% (95% CI 85% to 99%) and 88% (95% CI 81% to 92%). Two studies (93 patients) investigating PET for GCA diagnosis reported sensitivities of 67%–77% and specificities of 66%–100% as compared with clinical diagnosis or temporal artery biopsy. In TAK, one study each evaluated the role of magnetic resonance angiography and CT angiography for diagnostic purposes revealing both a sensitivity and specificity of 100%. Studies on outcome prediction and monitoring disease activity/damage were limited and mainly descriptive.ConclusionsUltrasound and MRI provide a high diagnostic value for cranial GCA. More data on the role of imaging for diagnosis of extracranial large vessel GCA and TAK, as well as for outcome prediction and monitoring in LVV are warranted.