Direct comparison of the diagnostic accuracy of PET/CT, cranial MRI, ultrasound and temporal artery biopsy in giant cell arteritis.

Abstract

To evaluate the diagnostic accuracy of PET/CT, cranial MRI, ultrasound and temporal artery biopsy (TAB) in patients with suspected giant cell arteritis (GCA) in a direct comparison. Consecutive patients with a suspicion of GCA and at least 2 diagnostic tests ≤ 7 days after initiation of glucocorticoids between June 2021 and June 2024, were included retrospectively. The gold standard for the diagnosis of GCA was the judgment of experienced clinicians after a follow-up of ≥ 6 months. Examinations were compared within subgroups undergoing the same tests. Sixty-one GCA patients and 50 patients with an alternative diagnosis were included. Combined cranial and large vessel PET/CT had the highest sensitivity (89% [95%CI 77-96%]) and specificity (98% [95%CI 88-100%]). Cranial PET/CT and TAB yielded a better sensitivity compared to temporal artery ultrasound (83% [95%CI 64-94%], 77% [95%CI 59-90%] and 55% [95%CI 36-74%], respectively, p = 0.023) without difference in specificity (100% [95%CI 100 - 84%], 95% [95%CI 76-100%] and 81% [95%CI 58-95%], respectively, p = 0.136). Cranial MRI had a sensitivity of 56% (95%CI 21-86%) and specificity of 82% (95%CI 48-98%). Large vessel PET/CT resulted in a better sensitivity compared to axillary artery ultrasound (68% [95%CI 45-86%] vs. 18% [95%CI 5-40%], p = 0.001) without difference in specificity (100% [95% CI 82-100%] vs. 90% [95%CI 67-99%], p = 0.50). PET/CT had a better sensitivity than ultrasound and cranial MRI. TAB and cranial PET/CT had a similar diagnostic yield. Not applicable.