Case 20-1

PET/CT Artifact: LV–Lung Mismatch ( Figure 20-1 )

A 53-year-old woman presented to the ED with new-onset atypical angina. Her risk factors included hyperlipidemia, hypertension, and family history of CAD. She was first ruled out for MI by ECG and serial troponin I. Subsequently she was referred for rubidium myocardial perfusion PET. The physical findings were remarkable for obesity (weight 230 lb, BMI 35 kg/m ² ) and hypertension. She underwent vasodilator stress testing with regadenoson. The ECG showed sinus rhythm at baseline. Stress ECG showed an isolated PVC and no ischemic changes. The images were obtained using a PET/CT camera operating in the 3D mode.

PET/CT artifact: LV–lung mismatch. A, Tomographic slices display with the resting results interleaved between the stress slices. Top four rows display the LV SA from apex (top left) to base (bottom right) . Next two rows display the stress and rest VLA slices from the septum to the lateral wall and the last two rows show the stress and rest HLA slices from the inferior to the anterior wall. Note there is a large reversible defect of moderate severity in the anterolateral wall. B, Polar maps representation of perfusion distribution. The three panels in the top row correspond to the stress, rest, and reversibility (normalized rest/stress) LV distributions. The bottom row is the defect extent polar maps display where areas that are abnormal in comparison to the normal database are highlighted in black . Note there is a blackout region in the anterolateral wall at stress indicating an abnormal study by quantitative analysis. The defect extent reversibility polar map (bottom right) shows a whiteout area in the anterolateral wall suggesting ischemic changes. C, Fused transaxial, coronal, and sagittal stress images. There is misregistration of the stress emission (color images) and the CT image (B/W images) . The apical, anterior, and lateral walls on the stress emission image overlay lung tissue on the CT image (red arrows) . The rest emission and the CT image are well-registered (images not shown). D, Fused transaxial, coronal, and sagittal stress images following emission-driven registration. The software artificially replaced lung tissue with soft tissue attenuation in the misregistered voxels (blue arrows) . E, the tomographic PET images and the polar maps (F) after registration and reconstruction no longer demonstrate perfusion defects. F, see previous page for legend.

Comments

Hybrid PET/CT cameras are rapidly replacing stand-alone PET cameras in cardiac imaging. While on stand-alone cameras, rotating sources (germanium or cesium) are used to obtain the attenuation image, on hybrid cameras, the CT image is used for PET attenuation correction. This reduces the attenuation scan time from minutes (3 to 10 minutes) to seconds (5 to 20 seconds). The PET image is averaged over a few minutes (about 5 minutes for rubidium). Because the CT scan time is short, sometimes the image does not match the LV–lung interface of the PET image. This increases the emission-transmission misalignment rate in cardiac PET. If the images are not properly aligned prior to reconstruction, perfusion defects may develop. The artificially low attenuation value of the LV wall on the CT image (lung tissue instead of soft tissue) decreases the expected photon attenuation and scatter from that region. When the images are reconstructed, the count density of the misregistered LV wall is artifactually decreased.

The original reconstructed PET images show a large reversible defect of moderate severity in the anterolateral wall. The stress perfusion defect is secondary to misalignment of the PET and CT images in the LV–lung interface. There is lung tissue on CT images overlaying LV activity in the anterolateral wall on PET images. Registration of the stress images was subsequently performed using an emission-driven method. This method is proprietary and not commercially available. It modifies the heart outline of the CT image based on the PET data. The program expects to find soft tissue attenuation on the CT image if the corresponding voxel on the PET image contains LV wall activity. If there is an inconsistency and lung tissue on the CT image overlays LV wall activity on the PET image, the attenuation value of the voxel is modified to match that of soft tissue. Following registration and reconstruction of the images, the PET images are normal.

Different CT protocols have been proposed in an attempt to minimize the emission-transmission misregistration rate on PET/CT imaging. At our institution, the “solution” to this problem has been to coach the patients to shallow breathe during the acquisition and to average the CT data over a longer time by slowing the rotational speed of the gantry and increasing the pitch. Despite these efforts, the CT attenuation scan time on multislice CT scanners is still typically under 20 seconds. We have tested other CT protocols including ECG-gated (with breath-hold), ungated (with and without breath-hold), “slow,” “fast,” and cine CT. The misregistration rate was highest with ECG-gated breath-hold CT (about 60%) and lowest with cine CT and “slow” ungated CT (about 20% each). Regardless of the CT protocol used for attenuation correction, it is important that the reader remembers to review the quality control images. The fused images should be available at the time of interpretation to ensure proper registration was performed prior to image reconstruction.

Case 20-2

PET/CT Artifact: LV–Diaphragm Mismatch ( Figure 20-2 )

A 42-year-old man was referred for rubidium myocardial perfusion PET because of worsening shortness of breath on exertion. He had type 2 diabetes and hyperlipidemia. Calcium scoring CT performed 1 year prior showed a total Agatston score of 102. He underwent vasodilator stress testing with regadenoson. There were no ST changes on ECG. The images were obtained using a PET/CT camera operating in 3D mode.

PET/CT artifact: LV–diaphragm mismatch. A, Tomographic slices display with the resting results interleaved between the stress slices. Top four rows display the LV SA from apex (top left) to base (bottom right) . Next two rows display the stress and rest VLA slices from the septum to the lateral wall and the last two rows show the stress and rest HLA slices from the inferior to the anterior wall. Note that there is relatively increased uptake in the inferior wall, which is slightly more pronounced on the stress images. B, Polar maps representation of perfusion distribution. Note the absence of blackout regions in the defect extent polar maps, indicating a normal study by quantitative analysis. C, Fused transaxial, coronal, and sagittal stress images. There is no LV–lung mismatch between the stress emission (color images) and the CT image (B/W images) . However, the CT images show hypoinflated lungs, the diaphragm is elevated (red arrows) .

Comments

The PET images in this patient show relatively increased uptake in the inferior wall, which is slightly more pronounced on the stress images. Fused stress PET/CT attenuation images reveal no LV–lung mismatch. However, there is LV–diaphragm mismatch. The lungs on the CT image are hypoinflated relative to the averaged lung volumes of the PET image. The elevation of the diaphragm on the CT image increases the expected photon attenuation from tissues in the same axial plane of the inferior LV wall because lung attenuation is replaced by soft tissue attenuation from the subdiaphragmatic organs. As a result, the reconstructed PET images show an artifactually increased count density in the inferior LV wall.

The LV–diaphragm mismatch is less prevalent than the LV–lung mismatch; it occurs in about 5% of cases at our institution. When severe, the “hot spot” in the inferior wall can produce downscaling in the remainder of the LV and artifactual perfusion defects may develop in the LAD and LCX distributions. These artifactual defects are sometimes “reversible,” particularly when the patient hyperventilates during the stress PET acquisition.

Can we prevent “reversible” artifacts by performing a CT scan post-stress for stress PET attenuation correction? At our institution, we use the same rest CT image for rest and stress PET attenuation correction. We have not been able to demonstrate an improvement in registration rate of the stress images by adding a CT attenuation scan post-stress. This is because we use regadenoson for pharmacologic stress testing. By the time the stress PET acquisition is finished and a second CT attenuation scan is obtained, hyperventilation and tachycardia have usually subsided (the same applies for adenosine). As a result, the CT image poststress is essentially another rest CT attenuation scan that does not improve the stress PET image quality.

Case 20-3

Rubidium 3D PET Artifact: Prompt-Gammas ( Figure 20-3 )

A 63-year-old woman was referred for rubidium myocardial perfusion PET because of one episode of unexplained syncope. Her risk factors for CAD were hypertension and hyperlipidemia. She underwent vasodilator stress testing with adenosine. Baseline ECG showed sinus rhythm, no Q waves, and no ST-T abnormalities. Stress ECG was unremarkable. The images were obtained using a PET/CT camera operating in 3D mode.

Rubidium 3D PET artifact: prompt-gammas. A, Schematic representation of 2D and 3D PET imaging. Note the presence of septa in the 2D mode. The lack of septa in the 3D mode increases the count sensitivity to coincident, random, and scattered events. The blue arrows represent recorded coincident events. The red arrows represent unrecorded photons. B, The tomographic PET images show a large, moderate to severe, predominately-fixed defect involving the apical, septal, and anterior walls. In the apical wall, the perfusion defect is more severe at rest. Note the background activity is nearing zero. C, Polar maps representation and quantification of perfusion defects. The perfusion defect in the LAD territory involves 23% of the LV, of which 21% is reversible by quantitative analysis (in the mid and basal anterior wall). D, Upon reconstruction of the images applying prompt-gamma correction, there is resolution of the defects, both visually in the tomographic display. and by quantitative analysis (E).

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Use in Stable Patients With Known or Suspected Coronary Artery Disease
2. MPI of Special Patient Groups
3. Myocardial Perfusion SPECT/CT: The Added Value of CT Imaging
4. Extracardiac Incidental Findings

Stay updated, free articles. Join our Telegram channel

Join