Key points
- •
Correct energy window position of the pulse height analyzer should be verified for each detector prior to SPECT acquisition.
- •
Maximal myocardial counts should be identified (usually by the computer) and the image display normalized to that value. In order to avoid normalization errors, appropriate time should be allowed for the radiotracer to be excreted from the liver. To avoid duodenogastric reflux of radiotracer, the patient should be instructed to drink water (at least 8 ounces) following radiotracer injection.
- •
The Ramp filter artifact occurs commonly with filtered backprojection tomographic reconstruction. Iterative reconstruction (ordered subset expectation maximization, [OSEM]) is useful to reduce this artifact. In order to avoid the Ramp filter artifact appropriate time should be allowed for the radiotracer to be excreted from the liver. This is particularly important for pharmacologic stress studies.
- •
Findings most consistent with photon attenuation by the left hemidiaphragm are an elevated left hemidiaphragm noted in planar projection images, a fixed inferior defect that appears to taper from the mid to basal portion of the ventricle, and normal wall motion and wall thickening.
- •
Findings most consistent with photon attenuation by the breasts are: a dense anterolateral breast “shadow” noted in planar projection images, a fixed anterolateral defect somewhat more severe in resting images, and normal wall motion and wall thickening. However, breast attenuation artifacts can vary depending on size and shape of the breast.
- •
In women with breast implants, which are more dense than soft tissue, associated attenuation artifacts extent tend to be smaller, more discrete, and more marked than breast attenuation artifacts caused by actual breasts.
- •
Attenuation correction, gated imaging, and prone imaging are helpful in interpreting studies with breast and/or diaphragmatic tissue attenuation.
- •
In patients with LVH, although there is generalized hypertrophy of the left ventricle, septal hypertrophy may be more marked. Therefore SPECT images demonstrate a relative increase in tracer concentration in the septum. Because tomograms are normalized to the region of myocardium with the highest count density, the septum appears relatively normal, and the remainder of the myocardium demonstrates relatively decreased count density. This is especially true in patients with end-stage renal disease who are on dialysis.
- •
Partial dose infiltration is one of several causes of suboptimal myocardial count density. To avoid “equivocal” interpretations of low count density scans, repeat imaging is appropriate.
- •
Incorrect ECG gating can occur for a number of reasons including: loose ECG leads, arrhythmias, or any cause of variability in the detected R-R time window.
Artifacts and normal variants are a significant source of false-positive interpretations of myocardial perfusion SPECT. By anticipating and recognizing such findings, the astute technologist and interpreting physician can increase test specificity in the diagnosis of coronary artery disease and avoid unnecessary catheterization of normal patients.
Case 3-1
Off-Peak Detector ( Figure 3-1 )
A 60-year-old man had atypical chest pain and no prior cardiac history. Technetium-99m sestamibi myocardial perfusion SPECT was performed using a single-day rest (9 mCi)/stress (32 mCi) protocol. The resting ECG was normal. The patient experienced no chest pain and there were no electrocardiographic changes during treadmill exercise. SPECT images were acquired using a 90-degree–angled dual-head camera and a 180-degree imaging arc from the 45-degree LPO projection to the 45-degree RAO projection. Therefore images from the 45-degree LPO projection to the 45-degree LAO projection, acquired with detector No. 1, are of good spatial resolution, whereas those images from the 45-degree LAO projection to the 45-degree RAO projection, acquired with detector No. 2, are of poor spatial resolution.
Comments
The fixed lateral wall defect is likely secondary to SPECT reconstruction, which includes projection images of poor spatial resolution during one half of the acquisition. Considering the patient’s history and normal regional LV function, lateral wall myocardial scarring is very unlikely. Prior to SPECT acquisition, correct energy window position of the pulse height analyzer should be verified. A 10%-15% symmetrical window should be centered on the 140-keV technetium-99m photopeak. For most cameras the selected energy window is automatically applied simultaneously to both detectors. In this case, however, there was a camera software error that failed to apply the selected energy window to detector No. 2.
Case 3-2
Normalization Error ( Figure 3-2 )
A 56-year-old woman with risk factors for CAD and atypical chest pain underwent a single-day, low-dose (9 mCi)/high-dose (32 mCi) technetium-99m sestamibi scan. During treadmill exercise the patient experienced no chest pain and there were no ECG abnormalities.
Comments
The various commercially available software programs used to process myocardial perfusion SPECT normalize reconstructed tomograms by different means. Preferably, maximal myocardial counts should be identified and the images normalized to that value. If instead the image is normalized to the entire frame or reconstructed volume, the maximal pixel counts in the entire frame will be chosen for normalization. If that pixel is outside of the myocardium (i.e., the stomach, bowel, or liver), the myocardium will be assigned a lower intensity value, as in the present case example. To correctly view and interpret reconstructed tomograms under these circumstances, the interpreting physician must arbitrarily increase the intensity of the images in which the normalization error occurred. This must be done carefully to avoid obscuring (overintensifying) actual perfusion abnormalities. In order to avoid normalization errors, appropriate time should be allowed for the radiotracer to be excreted from the liver. To avoid duodenogastric reflux of radiotracer, the patient should be instructed to drink water (at least 8 ounces) after radiotracer injection.
Case 3-3
Compton Scatter Image Degradation ( Figure 3-3 )
A 60-year-old woman had multiple risk factors for CAD and atypical angina. The patient underwent dual-isotope myocardial perfusion SPECT, receiving 3.5 mCi of thallium-201 at rest and subsequently 30.0 mCi of technetium-99m sestamibi during dipyridamole pharmacologic stress. The stress tomographic images are slightly suboptimal for interpretation due to Compton scatter into the inferoapical region. The resting images are inadequate for interpretation due to marked Compton scatter into the inferior wall.
Comments
The phenomenon shown in Figure 3-3 is due to Compton scatter of photons emanating from subdiaphragmatic activity. In the process of Compton scattering, these scattered photons have not lost enough energy to be eliminated by the energy windows positioned over the 68- to 80-keV Hg-201 mercury x-ray of thallium-201 or the 140-keV photon of technetium-99m, respectively. Therefore, as the scattered photons pass through the parallel-hole collimator, they are accepted by the pulse height analyzer and superimposed upon the inferior wall of the LV.
Because reconstructed myocardial tomograms are normalized to the area of greatest count density, in this case they are normalized to the artifactually intense inferior wall, resulting in a relative decrease in count density in the contralateral anterior wall. In this particular case example, this artifact is more marked in the resting thallium-201 tomograms than in the stress technetium-99m sestamibi tomograms.
To avoid normalization errors and endocardial edge-tracking errors, as present in this case example, appropriate time should be allowed for the radiotracer to be excreted from the liver. To avoid duodenogastric reflux of radiotracer, the patient should be instructed to drink water (at least 8 ounces) after radiotracer injection.
Case 3-4
Ramp Filter Artifact ( Figure 3-4 )
A 62-year-old man had multiple risk factors for CAD and nonanginal chest pains. The patient had no history of prior MI, and his resting ECG was normal. The patient underwent a single-day, low-dose (9 mCi)/high-dose (32 mCi) technetium-99m sestamibi scan. During dipyridamole pharmacologic vasodilatation, the patient experienced no chest pain and there were no ECG abnormalities.
Comments
In this case example the fixed inferoapical defect is most likely secondary to a Ramp filter artifact. After the process of filtered backprojection used to reconstruct planar projection images into tomograms, the Ramp filter is routinely applied to minimize/eliminate the “star artifact” caused by reconstruction of a finite number of projection images. By this means the Ramp filter suppresses counts immediately adjacent to “hot” objects/structures. The Ramp filter operates primarily in the x -plane. Therefore counts in the inferoapical wall of the LV in the same x -plane as the intense activity in the dome of the right lobe of the liver are suppressed by the Ramp filter.
The Ramp filter artifact occurs commonly with filtered backprojection tomographic reconstruction. Iterative reconstruction (ordered subset expectation maximization, [OSEM]) is useful to minimize the Ramp filter artifact. However, although in the present case example images were reconstructed using OSEM, the Ramp filter artifact is still present.
To avoid the Ramp filter artifact, appropriate time should be allowed for the radiotracer to be excreted from the liver. This is particularly important for pharmacologic stress studies, for which poststress radiotracer concentration is more marked than with exercise. To avoid duodenogastric reflux of radiotracer into the stomach, which also may lie in the x -plane of the inferior wall, the patient should be instructed to drink water (at least 8 ounces) after radiotracer injection.
Case 3-5
Diaphagmatic Attenuation ( Figure 3-5 )
A 45-year-old man with atypical angina underwent a single-day, low-dose (9 mCi)/high-dose (32 mCi) technetium-99m sestamibi scan. During peak treadmill exercise, the patient experienced no chest pain and there were no ECG abnormalities.
