Diagnostic Accuracy of Positron Emission Tomography/Computed Tomography of the Head, Neck, and Chest for Giant Cell Arteritis: A Prospective, Double‐Blind, Cross‐Sectional Study. Issue 8 (12th June 2019)

• Record Type: Journal Article
• Authors: Sammel, Anthony M.; Hsiao, Edward; Schembri, Geoffrey; Nguyen, Katherine; Brewer, Janice; Schrieber, Leslie; Janssen, Beatrice; Youssef, Peter; Fraser, Clare L.; Bailey, Elizabeth; Bailey, Dale L.; Roach, Paul; Laurent, Rodger
• Is Part Of: Arthritis & rheumatology. Volume 71:Issue 8(2019)
• Page Start: 1319
• Page End: 1328
• Publication Date: 2019-06-12
• Subjects: Arthritis -- Periodicals; Rheumatism -- Periodicals; 616.72
• DOI: 10.1002/art.40864
• Languages: English
• ISSNs: 2326-5191

Abstract

Abstract : Objective: Positron emission tomography/computed tomography (PET/CT) has not been well studied as a first‐line test for giant cell arteritis (GCA), due, in part, to historical limitations in visualizing the cranial arteries. The Giant Cell Arteritis and PET Scan (GAPS) study was therefore carried out to assess the accuracy of a newer generation PET/CT of the head, neck, and chest for determining a diagnosis of GCA. Methods: In the GAPS study cohort, 64 patients with newly suspected GCA underwent time‐of‐flight PET/CT (1‐mm slice thickness from the vertex to diaphragm) within 72 hours of starting glucocorticoids and before undergoing temporal artery biopsy (TAB). Two physicians with experience in PET reviewed the patients' scans in a blinded manner and reported the scans as globally positive or negative for GCA. Tracer uptake was graded across 18 artery segments. The clinical diagnosis was confirmed at 6 months' follow‐up. Results: In total, 58 of 64 patients underwent TAB, and 12 (21%) of the biopsies were considered positive for GCA. Twenty‐one patients had a clinical diagnosis of GCA. Compared to TAB, the sensitivity of PET/CT for a diagnosis of GCA was 92% (95% confidence interval [95% CI] 62–100%) and specificity was 85% (95% CI 71–94%). The negative predictive value (NPV) was 98% (95% CI 87–100%). Compared to clinical diagnosis, PET/CT had a sensitivity of 71% (95% CI 48–89%) and specificity of 91% (95% CI 78–97%). Interobserver reliability was moderate (κ = 0.65). Among the enrolled patients, 20% had a clinically relevant incidental finding, including 7 with an infection and 5 with a malignancy. Furthermore, 5 (42%) of 12 TAB‐positive GCA patients had moderate or marked aortitis. Conclusion: The high diagnostic accuracy of this PET/CT protocol would support its use as a first‐line test for GCA. The NPV of 98% indicates the particular utility of this test in ruling out the condition in patients considered to be at lower risk of GCA. PET/CT had benefit over TAB in detecting vasculitis mimics and aortitis.