The diagnosis of an infectious process may be easy, particularly when the site is visible or close to the surface in a febrile patient. Unfortunately, there are many circumstances when either the source or the extent of the infection is not obvious. Over the last 25 years, the diagnostic specialities of radiology and nuclear medicine have developed a large number of imaging techniques to aid the clinician in the evaluation of patients with suspected infections. Whereas 20-30 years ago such invasive methods as “exploratory laparotomy” were necessary, in 1998 non-invasive or minimally invasive imaging procedures can often lead to successful and accurate diagnosis and thence appropriate therapy.
In this issue, Dr. Elgazzar presents an excellent and comprehensive review of the diagnosis and localization of infection, with particular emphasis on the various nuclear medicine techniques available.1 He presents a series of potential diagnostic algorithms for the evaluation of both soft tissue and skeletal infections. Several other excellent reviews of the role of radioisotope techniques in various aspects of infection imaging have appeared in recent years.2–5 Clearly, this is a rapidly changing field with new developments still on the horizon. The amount of effort being expended in developing new physiologically based imaging methods is a reflection of not only the importance of this area to the management of patients with infections, but also the continuing need for easier, faster and more reliable techniques.
There is no doubt that nuclear medicine techniques have poorer resolution than any of the morphological techniques. The sub millimeter resolution of modern ultrasound (US), computerized tomography (CT) and magnetic resonance imaging (MRI), will always be better than any nuclear medicine study where resolutions of 4-5 millimeters for even superficial lesions are difficult to achieve. So why have nuclear medicine techniques continued to proliferate and play a major role in infection imaging and in some circumstances even become the gold standard by which other methods are evaluated? Morphologic techniques like US, CT and MRI rely on the development of structural changes that occur secondarily in infected tissues; the development of abscess cavities, inflammatory masses, tissue destruction, tissue swelling, and fluid shifts. On the other hand, nuclear medicine techniques have concentrated on attempting to directly depict the actual physiologic processes relative to infection; infiltration of white blood cells (WBC), accumulation of lactoferrin, binding of chemotactic factors, and increases in glucose metabolism.
Admittedly, some nuclear medicine techniques proposed for imaging infection (e.g., labeled liposomes and immunoglobulins) are non-specific and not directly related to the infectious process itself, but are more a reflection of secondary processes such as increases in vascular permeability and increased tissue fluid accumulation. These non-specific techniques are less attractive as potential imaging methods, although patient trials with labeled human polyclonal IgG have yielded some potentially useful indications, if one is able to wait the 12-24 hours necessary for localization.6 Other non-specific techniques such as labeled nanocolloids, initially thought to hold some promise, have now been largely abandoned as either ineffective or unable to equal the accuracy of other currently available techniques. Even several of the so-called specific techniques such as labeled antigranuloycte antibodies may owe much of their localization at sites of infection to non-specific mechanisms such as increased vascular permeability. Still other specific techniques (like labeled chemotactic factors) have had difficulty getting out of the laboratory and animal models and into actual patient trials, despite up to ten years of development work.7
Positron emission tomography (PET) using 18-Fluorine labeled flourodexyglucose (18-FDG) detects areas of increased glucose metabolism in the body, and is capable of higher resolution images than normally possible with standard nuclear medicine instruments. Glucose metabolism, especially anaerobic, is increased many fold at sites of infection and PET imaging has shown great promise in studies in patients with infection, with results being available within two hours. However, PET instruments are expensive and not widely available (there is only one available in the Arab world at King Faisal Specialist Hospital, Riyadh) and the short half-life of 18-FDG (2 hours) limits its usefulness outside of major centers. Finally, it must be remembered that no nuclear medicine imaging technique is very accurate at differentiating sterile inflammatory lesions from infectious lesions. Gallium-67, labeled WBCs, 18-FDG, labeled IgG, and labeled antigranulocyte antibodies all accumulate to varying degrees at sites of simple sterile inflammation without infection. Certainly, intense uptake of these agents at a site is more often associated with acute infectious processes but low grade infections and inflammation without infection often appear identical in these imaging studies.
Overall, gallium and WBC scans are the only two nuclear medicine procedures used with any frequency in patients with suspected infections. While there is a tendency for gallium scanning to be more accurate in chronic infections and WBC scanning in acute infections, either technique is very useful in most clinical settings when interpreted by a skilled physician who is knowledgeable about the pitfalls inherent in each technique. Tc-99m WBC scans usually are preferred to gallium-67 scans because of greater availability, higher resolution and the earlier diagnosis of infection (6 hours in comparison to 24 hours or more for gallium). WBC scans are performed after labeling WBCs in vitro using a technetium-99m (Tc-99m) compound and a process that requires 1½ to 2 hours. The technique of separating WBCs from a sample of the patient's blood and labeling with Tc-99m is laborious and requires meticulous attention to detail, but should be available in all but the smallest nuclear medicine facilities. Gallium-67 must be purchased commercially and may not be routinely available in all nuclear medicine departments. Injection is directly intravenous and no handling and processing of the patient's blood is necessary. A medium energy collimator must be available for the resolution of the nuclear medicine camera used is considerably inferior to that obtained with Tc-99m imaging. The two major indications for gallium scanning are granulocytopenic patients (white blood cell count less than 4×109/L) and imaging for osteomyelitis in the spine and pelvis where normal marrow uptake of Tc-99m WBCs makes WBC scanning relatively insensitive.
In the work up of patients with suspected infections, the choice of imaging procedure may be more determined by the availability of local equipment and personnel than a scientific evaluation of the most appropriate technique in any particular clinical setting. When the local performance or interpretation of US, CT or nuclear medicine studies is of a poor quality, the use of alternative less effective procedures may be justified, if performance is of a higher quality. For example, a superb ultrasonographer with new equipment will consistently out perform a poor nuclear medicine study on outdated equipment interpreted by an inexperienced physician. Of course, the reverse may be the case in a neighboring institution. Furthermore, the diagnostic imaging specialist who makes himself available to the referring clinicians, examines the patients sent for study and reviews the findings with his colleagues, will be the most useful in helping to arrive at the correct diagnosis.
Thus, while the diagnostic algorithms proposed by Dr. Elgazzar in this issue1 are very helpful in picking the right test from the maze of possible investigations available, the most astute clinicians may need to modify these guidelines to suit their own particular local situation. In this regard, careful follow-up of the results of infection imaging procedures performed on their patients requires not only bacteriologic and/or pathologic confirmation, but also good clinical follow-up with an emphasis on clinical outcome. Only in this way will the clinician responsible for the patient be able to determine the reliability of the various techniques available in their institution and thus be able to modify the algorithms presented, if necessary. In analyzing clinical outcome, the impact of any diagnostic procedure can at best be only as good as the subsequent treatment and management of the patient. The results of our diagnostic studies must be acted upon in an appropriate and effective manner.
In summary, while there is a large number of possible procedures available to image infectious processes, careful consideration of the algorithms presented should greatly aid the attending physician to choose the best possible procedure that will allow him to arrive at the correct diagnosis in the shortest time and in a cost effective manner. As newer techniques become available, we may need to further modify these algorithms and indeed, in the future, it is most likely that the study selected for each patient will be individualized according to the advantages of each agent in that particular patient's clinical setting.
1. Elgazzar AH. Imaging of infection: a correlative and algorithmic approach J of Family and Community Medicine. 1997;2:21–32
2. Datz FL. Infection imaging Semin Nucl Med. 1994;24(2):87–179
3. Datz FLFreeman LM. Musculoskeletal Infection Nuclear Medicine Annual. 1993 New York Raven Press:77–122
4. Palestro CJFreeman LM. Musculoskeletal Infection Nuclear Medicine Annual. 1994 Raven Press New York:91–120
5. Kipper SLFreeman LM. Radiolabeled leukocyte imaging of the abdomen Nuclear Medicine Annual. 1995 New York Raven Press:81–128
6. Rubin RH, Fischman AJ. The use of radiolabeled non-specific immunoglobulin in the detection of focal inflammation Semin Nucl Med. 1994;24:169–79
7. Vallabhajosula S. Technetium 99m labeled chemotactic peptides: specific for imaging infection? J Nucl Med. 1997;38:1322–6