PET will make SPECT irrelevant in 10 years

When Sir Godfrey Hounsfield introduced CT for brain imaging in 1973, my colleagues and I at the University of Pennsylvania were using a dedicated SPECT instrument, designed and assembled at our institution, to examine blood-brain barrier abnormalities in a variety of neurological disorders.

Upon the installation of our first CT unit, we initiated a prospective research project to compare CT with iodinated contrast agents to SPECT with technetium-99m pertechnetate in brain tumor, head injury, and stroke.

It soon became clear that CT's exquisite spatial resolution combined with iodinated contrast agents was substantially superior to SPECT and Tc-99m-labeled compounds for visualizing blood-brain barrier abnormalities. As a result, the desire to use SPECT for central nervous disorders almost completely disappeared over the ensuing years.

Efforts in our group to determine regional brain function with single-photon-emitting radionuclides yielded no substantial rewards. As a result, we were totally convinced that SPECT as an imaging modality suffered from significant limitations that could not be readily overcome in the foreseeable future. Three decades later, the same limitations still exist and threaten SPECT as a competitor to the existing powerful imaging techniques, especially PET. I base my assessment on the following observations.

Planar imaging with single-photon-emitting radionuclides is a suboptimal technique, and it cannot compete with modern tomographic techniques in the day-to-day practice of medicine. It suffers from a low sensitivity for detecting disease sites. We project that use of this approach will be completely limited to studies in which measurement of gross function is the main concern. As a result, the need for instruments with planar imaging capability alone will noticeably decrease in the next decade.

SPECT is obviously superior to planar imaging and substantially improves the sensitivity of the technique for detecting deeply seeded lesions. However, it also suffers from poor spatial resolution compared with PET and can easily miss small lesions with low contrast. In addition, because of its limited speed in screening the entire body, SPECT is usually confined to specific anatomic sites such as the brain, chest, or abdomen alone and not in combination. This limits the use of SPECT in patients with cancer and other systemic disorders for whom screening the entire body is the focus of the investigation.

Labeling single-photon-emitting radionuclides to biologically important compounds has posed a challenge to chemists throughout the history of nuclear medicine. This drawback particularly pertains to Tc-99m pertechnetate, which has restricted the expansion of the field to novel applications. While only a limited number of compounds labeled with single-photon-emitting radionuclides have been introduced in the past two decades, an overwhelming number of PET tracers have been synthesized and tested around the world during the same period. The proliferation of PET tracers is due to the feasibility of labeling positron-emitting elements such as carbon-11 or fluorine-18 and several metals to biologically relevant molecules. In addition, all the current available kits used for preparing Tc-99m radiopharmaceuticals can be adapted for labeling with Tc-94m, a positron emitter with a reasonable half-life, for routine clinical purposes.

Certain preparations such as fluorine-18 fluoride can be successfully used for bone imaging instead of Tc-99m-labeled phosphates. This tracer is readily extracted from the target and can be used with minimal quality-assurance efforts. The cost of producing F-18 fluoride for bone imaging is markedly lower than that of other PET tracers and can be quite competitive with conventional bone imaging agents. It is clear that conventional planar bone imaging with Tc-99m-labeled phosphates is suboptimal for detecting active sites in the spine and pelvis. Tomographic images with F-18 fluoride can overcome this major deficiency.

Because of its extraordinary sensitivity, PET can be more effective in acquiring images at short intervals, which obviously is not possible with SPECT techniques. This capability is important in quantitative models where temporal measurements of compound concentration are essential for calculating important physiological and metabolic parameters. This approach is of great interest in research applications but may be necessary in certain clinical settings.

In addition, PET is substantially more accurate than SPECT in quantitative assessment of regional concentration of radiotracers. Such accuracy is of considerable importance for both clinical and research applications.

For all these reasons, I believe that PET will become the standard of care in nuclear medicine, and the use of single-photon emitters will substantially decline over the next decade. Manufacturers of nuclear medicine instruments and the companies currently distributing single-photon-emitting radiopharmaceuticals should diversify their domains to include PET as the focus for future markets. Cyclotrons will eventually become the main source of generating radionuclides, and they will replace Tc-99m generators at institutions and commercial sites.

Dr. Alavi is chief of nuclear medicine at the Hospital of the University of Pennsylvania in Philadelphia.