Positron emission tomography–computed tomography (PET-CT) has been used as an imaging modality in workup of fever of unknown origin (FUO). The aim of our study is to evaluate the diagnostic utility of PET-CT in FUO workup in a resource-limited setting. We also looked at laboratory parameters as predictors of contributory PET-CT scans and propose an algorithm for evaluation of FUO in resource-limited tropical regions. This retrospective observational study included patients admitted for FUO workup under general medicine in a teaching hospital in South India from June 2013 to May 2016. PET-CT was done when the patient remained undiagnosed after a detailed clinical assessment and first- and second-tier investigations. Among 43 patients included in our study, a definite diagnosis was established in 74% (32). Noninfectious inflammatory diseases, infections, malignancies, and miscellaneous diseases were diagnosed in 37.2% (16/43), 23.3% (10/43), 9.3% (4/43), and 4.7% (2/43), respectively. Tuberculosis was the single most common disease seen in 20.9% (9/43). PET-CT scans were contributory toward establishment of final diagnosis in 90.7% (39/43). High C-reactive protein (CRP) and aspartate aminotransferase (AST) levels were associated with contributory PET-CT scans (
P
= 0.006 and 0.011, respectively). PET-CT delineating organ/tissue for diagnostic biopsy was associated with final diagnosis of infectious disease (
P
= 0.001). Sensitivity, specificity, and positive and negative predictive value of PET-CT scans were 76.9% (20/26), 33.3% (2/6), 83% (20/24), and 25% (2/8), respectively. High CRP and AST were predictors of contributory PET-CT scans. PET-CT scans have high sensitivity and positive predictive value when used in evaluation of FUO. Although it is a useful tool in FUO workup, especially in the diagnosis of tropical infections, PET-CT should be done after a comprehensive clinical assessment and basic investigations.