The usefulness of 18F-FDG PET-CT in the management of fever of unknown origin - Prospective multi-central study

1968 Objectives 18F-FDG-PET-CT (PET-CT) has been utilized in the management of fever of unknown origin (FUO). This study aims to demonstrate the diagnostic performance of PET-CT in patients with FUO. Methods 231 consecutive patients with FUO (187male, mean52+/-18) in 2 general hospitals were prospectively studied using PET-CT after non-conclusive conventional investigations. Final diagnosis was based on biopsy, microbiological tests and imaging follow-up. Results The cause of FUO was identified only in 129/231 (56%): 51 infection, 27 malignancy, 41 non-infectious inflammatory diseases and 10 other causes. The patients had PET-CT scan at day 14 on average (+/- 5days) of hospital admission. Results Table: True Positive in 98/231: 48 infection, 22 malignancy, 28 inflammation. False Positive in 18/231, due to reactive nodes confirmed with subsequent CT. True Negative in 84/231, clinically self-limiting conditions with full spontaneous recovery. Such spontaneous recovery took 22 days on average (+/- 18 days) after hospital admission. False Negative in 31/231: 3 infection, 5 malignancy, 13 inflammation, 10 others. PET-CT appears to miss connective tissue diseases, non-FDG-avid malignancy and poorly-FDG-avid malignancy or infection within the organs of normal physiological FDG-uptake such as brain and liver. The PPV, NPV and accuracy of PET-CT were 84%, 73% and 79% respectively. On multivariate analysis, none of the inflammatory markers (white cell counts, ESR and CRP) statistically satisfied as independent predictor of PET-positivity. Conclusions 18F-FDG PET-CT correctly demonstrated or excluded a cause in 79% of FUO patients. A negative PET-CT with no spontaneous recovery still requires further investigations in order to exclude sinister causes such as myeloma and small gastrointestinal/renal/pancreatic malignancies, although it could take up to a month to reach final diagnosis.