LAUSR

OBJECTIVES To determine the impact of imaging acquisition time on interpretation of disease activity on 18F-fluorodeoxyglucose positron emission tomography (PET) in large-vessel vasculitis (LVV) and assess the relationship between clinical features and image acquisition time. METHODS Patients with giant cell arteritis (GCA) and Takayasu's arteritis (TAK) were recruited into a prospective, observational cohort. After a single injection of FDG, all patients underwent two sequential PET scans at one and two-hour time points. Images were interpreted for active vasculitis by subjective assessment, qualitative assessment, and semi-quantitative assessment. Agreement was assessed by percent agreement, Cohen's kappa, and McNemar's test. Multivariable logistic regression identified associations between PET activity and clinical variables. RESULTS 79 patients (GCA  =  44, TAK  =  35) contributed 168 paired one and two-hour PET studies. A total of 94 out of 168 scans (56%) were interpreted as active at the one-hour time point, and 129 scans (77%) were interpreted as active at the two-hour time point (p < 0.01). Associations between clinical variables and PET activity categories (dual inactive, delayed active, dual active) were evaluated. Using multivariable nominal regression, clinically active disease was significantly more common in patients in the delayed active group (Odds Ratio 1.94, 95%CI 1.13-3.53; p  =  0.02) and the dual active group (Odds Ratio 1.71, 95%CI 1.06-2.93; p  =  0.04) compared to the dual inactive group. CONCLUSIONS Imaging protocol significantly influences interpretation of PET activity in LVV. A substantial proportion of patients with LVV have PET activity only detected by delayed imaging. These patients were significantly more likely to have concomitant clinically-determined active disease.