A prospective study of computed tomography, ultrasound, and gallium imaging in patients with fever.

Computed tomography, ultrasound, and 67Ga-citrate imaging were analyzed prospectively in patients thought to have a focal source of sepsis. They were divided into three groups: (a) postoperative, fever greater than or equal to 38.3 degrees C; (b) fever greater than or equal to 38.3 degrees C for less than four weeks, unrelated to surgery; and (c) any fever present for more than four weeks. ROC curves showed no significant difference in the ability of the three modalities to differentiate focal from nonfocal sources of sepsis. If any two examinations were used and either study was abnormal, the sensitivity increased from about 60% to nearly 90% while the false-positive rate increased from about 15% to 25%. When focal disease was diagnosed only if two examinations were abnormal, the false-positive ratio dropped to nearly zero but the sensitivity fell to below 40%. The authors concluded tha all three modalities have a similar ability to detect sepsis and that sensitivity can be increased by using any two of them.