The use of PET scanning in patients with human immunodeficiency virus infection and fever of unknown origin, confusion and/or weight loss was investigated.
Methods: Eighty patients were examined using PET. Fifty-seven patients had half-body scans with [18F]fluorodeoxyglucose (FDG), and 23 patients had brain studies performed with FDG. Fourteen patients also had [11C]methionine studies (2 chest, 1 abdomen and 11 brain) performed.
Results: Thirteen patients with lymphoma had the extent of the disease clearly identified in both nodal and extranodal sites. Patients with a variety of infections (Cryptococcus neoformans, Pseudomonas aeruginosa, Mycobacterium tuberculosis and Mycobacterium avium intracellulare) had disease localized for appropriate biopsy or sampling procedures. A half-body FDG-PET scan had a sensitivity of 92% and a specificity of 94% for localization of focal pathology that needed treatment. High uptake of FDG (greater than liver) had a positive predictive value for pathology needing treatment of 95%. FDG brain studies showed that 16 patients with CD4 T-lymphocyte counts less than 200 cells/ml had reduced cortical uptake compared with that in basal ganglia. FDG scans were abnormal in all 19 patients with focal space occupying lesions identified by magnetic resonance scans. The standardized uptake values (SUVs) over cerebral lesions due to toxoplasma were in the range of 0.14-3.7 (13 patients) and due to lymphoma were in the range of 3.9-8.7 (6 patients). Three more patients with progressive multifocal leukoencephalopathy had SUVs in the range of 1.0-1.5 over the lesions. Another patient had a low-grade oligodendroglioma (SUV = 2.9). Carbon-11-methionine uptake also was high in patients with cerebral lymphoma but did not add to the discrimination between toxoplasmosis and lymphoma in these patients obtained with the FDG scan.
Conclusion: In hospitals with access to PET facilities, FDG scanning allows the rapid evaluation of the whole body, including the brain, of patients with human immunodeficiency virus infection, with a report potentially available within 4 hr of injection. Sites of infection and tumor were identified, and discrimination between cerebral pathologies was possible.