RESUMO
Aims: The goal of the investigation was to evaluate whether a semi-quantitative method reflecting myocardial 2-[18F]fluoro-2-deoxy-D-glucose (FDG) uptake heterogeneity has added value in addition to visual analysis in the diagnosis of cardiac sarcoidosis (CS). Methods and results: This retrospective analysis included 271 consecutive patients suspected of CS attending cardiac positron emission tomography combined with computed tomography (PET-CT) at our institution between 2007 and 2013. Visual analysis of PET-CT and semi-quantitative analysis of heterogeneity [coefficient of variation (CoV)] of myocardial FDG uptake were performed. The presence of CS and initial symptoms were verified from patient data. The criteria for CS included histological verification from the myocardium or from an extracardiac site. Thirty cancer patients without cardiac disease were included as controls. CS was diagnosed in 48/231 (20.8%) of analysed patients. Of these, 13 (27.1%) had no extracardial signs of the disease and 30 (62.5%) had FDG positive mediastinal lymph nodes. Visual analysis of PET-CT identified 48.9% of the CS patients. We found a cut-off value of 0.184 for CoV to have the best accuracy to detect CS from a patient population with suspected CS (75.0% sensitivity and 51.4% specificity). Compared to controls, CoV identified CS patients with a good accuracy (68.8% sensitivity and 93.3% specificity). CS patients with FDG positive mediastinal lymph nodes had higher CoV than CS patients without lymph node involvement (0.282 vs. 0.208, P = 0.016). CS patients with more severe initial symptoms had a higher CoV than patients with more benign symptoms (0.283 vs. 0.195, P = 0.01). Conclusion: CoV provides a good addition to visual analysis of cardiac FDG PET-CT in diagnosis of CS. As a semi-quantitative measure, it reduces intra-observer variability. It also seems to indicate more severe disease, but to confirm this, prospective studies are needed.
