Can dedicated cardiac SPECT outperform traditional gamma imaging?

For patients who have just suffered myocardial infarction due to CAD and heart failure, gleaning information about left ventricular (LV) viability with gated blood-pool imaging is highly recommended. Dedicated cardiac SPECT cameras with state-of-the-art detectors and image processing offer a new alternative that may provide advantages compared with traditional SPECT systems.

The only caveat is the technology’s lack of capability to provide the dynamic planar imaging suggested for radionuclide angiography, but evidence suggests that iterative image reconstruction may suffice, according to a study published June 1 by the Journal of Nuclear Cardiology.

R. Glenn Wells, PhD, from the department of cardiology at the University of Ottawa, and colleagues tested traditional SPECT cameras against dedicated cardiac SPECT systems and image reconstruction to compare how the two measured parameters of LV function, including ejection fraction (EF), wall motion, peak ejection rate (PER) and peak filling rate (PFR).

“Dedicated cardiac SPECT cameras have recently been introduced, which offer many advantages over traditional SPECT systems,” wrote Wells et al. “The new cameras use multiple detectors that are focused on the heart, and the design provides a greatly increased sensitivity. Some of the systems also use cadmium-zinc-telluride (CZT) detectors that have greatly improved energy resolution.”

Since these systems are unable to acquire traditional planar imaging, clinicians are using image data reconstruction, or reprojection, to home in on these LV parameters.

“Reprojection of 24-frame gated blood-pool SPECT images is an effective means of obtaining LV functional measurements with a dedicated cardiac SPECT camera using standard 2D-planar analysis tools,” the authors wrote.

For this research, a total of 48 patients underwent gated blood-pool imaging using a traditional gamma camera, which was immediately followed by dedicated cardiac SPECT between January and August 2010 and March 2011 and January 2012. Standard equilibrium blood-pool imaging studies and Tc-99m labeled red blood cells were used to acquire 24-frame gated planar images. Additional 24-frame ECG-gated data were gathered during 8 minutes of dedicated cardiac SPECT imaging. The gated SPECT image volumes were reconstructed in post-processing, or iterative reconstruction, into a 24-frame gated planar format using specialized software to calculate EF, end diastolic volume (EDV), PER and PFR. Reconstructed images were not pristine, but they were useful and comparable to traditional methods, according to the authors.

“The reprojected images from the dedicated system have lower spatial resolution than those obtained directly as planar images with a traditional system,” the researchers wrote. “Though this did not impair the correlation of EF and EDV, the effect on other types of evaluations such as wall motion or phase analysis still requires further investigation. An advantage of the SPECT-reprojection approach is that the angle of reprojection can be adjusted after acquisition, allowing optimization of the view that best separates the LV.”

Average difference of EF measurements between the two imaging methods was approximately 0.4 percent.

This method of image data processing could potentially be used with dedicated cardiac SPECT to expand the technology, which provides higher resolutions in as little time as a few minutes, whereas standard SPECT acquisitions tend to reach a cardiac cycle ceiling of 8 or 16 gates, which hampers EF, PER and PFR calculations.

Researchers also found statistically significant contrast between observers of the dedicated cardiac SPECT technology.

“The value for EF matched well between the two systems, but the EDV and PER values were significantly different suggesting a need for the establishment of new normal limits for the dedicated camera,” they wrote. “Reprojection of the gated dedicated cardiac SPECT image provides an effective means of creating traditional planar blood-pool datasets.”