Background The assessment of disease activity of Takayasu arteritis (TA) is difficult if symptoms and serum inflammatory marker were not detected. Even in those conditions, relapses were frequently observed during the dose reduction of corticosteroid and immunosuppressant. There is accumulating evidence that 18F-fluolodeoxyglucose-positron emission tomography (FDG-PET) and PET/ computed tomography (PET/CT) is useful for monitoring patients with TA when TA was clinically active. However, it is not clear the significance of FDG accumulations when TA was inactive.
Objectives To investigate a quantitative predictor in FDG-PET or PET/CT scans for the relapse of TA.
Methods We retrospectively investigated 76 FDG-PET or PET/CT scans and extracted 37 scans which were performed in inactive status. These scans were divided in two groups according to relapse of TA for 5years. The relapse was defined the increase of CRP and steroid dose or addition of immunosuppressant. FDG accumulations in aortic wall lesions of TA was evaluated by semi-quantitative index; the standardized uptake value (SUV). In addition to SUVmax in the aortic wall, we also calculated SUV ratio of maximum aortic wall uptake to mean lung uptake (ratio Lu), SUV ratio of maximum aortic wall uptake to mean liver uptake (ratio Li), and SUV ratio of maximum aortic wall uptake to mean aortic blood pool uptake (ratio BP). We compared groups using these parameters. We also determined the cutoff levels, sensitivity, and specificity of 4 sets of SUVs (SUVmax, ratio Lu, ratio Li, and ratio BP) for the prediction of relapse using Receiver Operating Characteristic (ROC) analysis. Moreover, Kaplan-Mayer analysis for the long-term relapse-free survival was performed to assess the reliability of these cutoff levels.
ResultsIn 37 total PET and PET/CT scan examinations, non-relapse and relapse groups included 17 and 20 scans, respectively. Relapse group had more immunosuppressant users than non-relapse group. Although CRP level and SUVmax were equivalent, ratio of SUV, especially ratio BP of relapse group was higher than that of non-relapse group (p=0.09) (Figure top panel). The cut-off level of these parameters was calculated as follows; SUVmax 1.4, ratio Lu 5.31, ratio Li 1.01, and ratio BP 1.41, respectively. Using these cut-off level, relapse rates of below and over cut-off level were as follows; SUVmax 50% vs 54%, ratio Lu 43% vs 62%, ratio Li 43% vs 69%, and ratio BP 31% vs 67%, respectively (Figure middle panel). Using Kaplan-Mayer analysis, relapse rate of these two groups divided by ratio BP was not significantly different (p=0.268) though these two curves looked different (Figure bottom panel).
Conclusions Our data suggest that ratio BP at stable condition, which represents relative FDG accumulation of the aortic wall lesions to aortic blood pool, could be a promising predictor to assess the relapse after successful treatment.
Disclosure of Interest None declared