Fever of unknown origin and inflammation of unknown origin are highly challenging diagnostic conditions. The current practice for evaluating patients is to conduct a positron emission tomography-computed tomography (PET-CT) scan as either a first- or a second-line modality. We aimed to assess the contributory effect of PET-CT to the diagnosis and compare it with the contributory effect of CT alone.

We performed a systematic review and meta-analysis. We included all cohorts that examined the contribution of PET-CT to the investigation of classical fever of unknown origin and inflammation of unknown origin. The primary outcome was the contribution of PET-CT to the final diagnosis. Secondary outcomes were sensitivity and specificity of PET-CT and CT scans, and contribution of a CT scan. We pooled the results of all studies and calculated the pooled contributory effect of PET-CT.

Thirty-six studies (3516 patients) were included in the systematic review. The pooled contribution of PET-CT was 75.4%. The compiled sensitivity and specificity values for all studies were 85.9% and 59.5%, respectively. Five studies (405 patients) compared between the PET-CT component and the total body CT component. The pooled contribution of a CT scan was 68%. The summed sensitivity and specificity values of a CT scan for all studies were 63.1% and 84.4%, respectively.

PET-CT has a contributory effect of 75% for the diagnosis of fever of unknown origin and inflammation of unknown origin. PET-CT had superior sensitivity and inferior specificity vs the CT scan.