Focal Tracer Uptake: A Potential Artifact in Contrast-Enhanced Dual-Modality PET/CT Scans

MATERIALS AND METHODS

Thirty whole-body scans covering the head, neck, thorax, abdomen, and pelvis of oncology patients were obtained on a dual-modality PET/CT system (biograph; Siemens Medical Solutions, Hoffman Estates, IL). For optimal CT image quality (vascular and parenchymal delineation comparable with a separately acquired CT scan), 140 mL of an iodinated contrast agent containing 300 mg of iodine per milliliter (Xenetix 300; Guerbet GmbH, Sulzbach, Germany) were administered intravenously with an automated injector (XD 5500; Ulrich Medical Systems, Ulm, Germany). To ensure sufficient vascular and parenchymal contrast, the flow rate was set to 3 mL/s for the first 80 mL. The remaining 60 mL were administered at 2 mL/s for continuous vascular enhancement during the acquisition of the CT sections. CT parameters were set to 130 mA, 120 kV, a slice width of 5 mm, and a table feed of 8 mm per rotation with a rotation time of 800 ms. After acquisition of the CT data, the patient table was automatically moved farther into the imaging system and PET images were collected starting with the upper thigh and pelvis.

PET images were collected 1 h after injection of 350 MBq of ¹⁸F-FDG within a scan window of 5 min per bed position (6–8 bed positions per patient). In-plane spatial resolution amounted to 4.6 mm (for a point source in air), and the axial field of view for 1 bed position was 15.5 cm, with a bed overlap of 15 planes (plane spacing, 2.425 mm). PET images were corrected for attenuation on the basis of the CT data (2). Iterative algorithms (nonlinear Fourier rebinning and nonlinear attenuation-weighted ordered-subsets expectation maximization) with 2 iterations and 8 subsets were used for image reconstruction. Data were filtered (3.0 mm in full width at half maximum) and corrected for scatter.

Imaging studies were evaluated qualitatively for the artifact, defined as a region of high count rate on attenuation-corrected images in accurate coregistration with a contrast-enhanced blood vessel. Each patient study was assessed by 2 radiologists and 2 nuclear medicine physicians, and evaluation was performed in consensus. In patients with the PET artifact (group 1), regions of interest were applied to the center of the underlying vessel (minimum size, 20 pixels) to quantify the degree of contrast enhancement. In patients without the PET artifact (group 2), density was measured in the subclavian vein of the arm into which the contrast agents had been administered. Mean density values and SDs were calculated for each group.

Each patient’s body surface area was calculated as described by Mosteller (3). Mean values of the body surface areas were compared between the 2 groups.

Statistical analysis was through determination of mean values and SDs. Significance was checked by the χ² test. P < 0.05 denoted statistical significance.

RESULTS

The contrast-related artifact mimicking focal tracer uptake was identified in 4 (13%) of the 30 evaluated examinations. All artifacts projected onto a highly enhanced thoracic vein. Three of these 4 patients presented with the artifact in projection onto the right subclavian vein (Fig. 1). In all 3 patients, the contrast agent had been injected through the right antecubital vein. In the fourth patient, the artificial focal tracer accumulation projected over the superior vena cava just distal to the tip of a central venous catheter that had been used for administering the contrast agent (Fig. 2).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Attenuation-corrected PET image (A), PET/CT fusion image (B), CT image (C), and PET image uncorrected for attenuation (D) of 43-y-old woman with centrally located non–small cell cancer of right lung and mediastinal lymph node metastasis. PET attenuation artifact is due to highly concentrated contrast agent in right subclavian vein (arrow in A), which is not detectable on PET image uncorrected for attenuation (D). On attenuation-corrected PET, hot spot in right axillary region could be misinterpreted as metastasis.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Attenuation-corrected PET image (A), PET/CT fusion image (B), CT image (C), and PET image uncorrected for attenuation (D) of 34-y-old man with hepatic metastases from colorectal cancer. Artifact caused by contrast agent leads to error in CT-based attenuation correction of PET images and simulates focal tracer uptake. Without knowledge of fusion images, interpreter is likely to misinterpret finding as mediastinal lymph node metastasis. When PET images are uncorrected for attenuation, artifact is not expressed (D).

Quantitative region-of-interest analysis of the intravenous density in the artifact areas revealed a mean density (±SD) of 2,262 (±304) Hounsfield units (HU) (range, 1,878–2,603 HU) on CT images. Measurements of CT HU in the subclavian veins of patients exhibiting artifact-free PET images revealed a mean density of 1,058 ± 209 HU and a range of 634–1,368 HU. CT density differences between veins exhibiting the PET artifact and veins that did not were statistically highly significant (P < 0.001).

The mean body surface area of all examined patients amounted to 1.96 ± 0.21 m². The mean surface area in patients expressing the artifact (1.67 ± 0.11 m²) was lower than that in patients without the artifact (2.01 ± 0.18 m²). This difference was also statistically significant (P = 0.035).

DISCUSSION

To exploit the diagnostic power associated with dual-modality PET/CT, knowledge of associated artifacts is crucial. This study highlights such an artifact, which is related to the passage of iodinated contrast agents during the acquisition of CT images.

In dual-modality PET/CT tomographs, PET images are corrected for attenuation on the basis of CT data (2), which must be collected in the presence of intravenous and oral contrast agents to permit vascular and intestinal delineation, respectively. Furthermore, intravenous enhancement is crucial for the evaluation of parenchymal organs to maximize both sensitivity and specificity (4). Contrast bolus passage and vascular and parenchymal contrast enhancement are, however, temporary effects. Although present during the acquisition of CT images, most of the contrast agent will already be eliminated from the vascular system and the parenchymal organs when the PET data are collected. CT x-rays are therefore attenuated by the contrast agent, whereas PET annihilation quanta are not. The resulting overestimation of the PET attenuation can result in an artifact, as illustrated in this study.

Furthermore, attenuation of x-rays by iodinated contrast agents is significantly higher at CT energies (up to 140 keV) than at the 511-keV PET photon energy (Fig. 3) (5). An overestimation of PET attenuation in the presence of intravascular iodinated contrast agents is therefore to be expected in all currently available PET/CT tomographs.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Attenuation per unit density at different energies for soft tissue, bone, and iodine. Iodine contrast agents cause higher attenuation at CT energies than at PET photon energy.

The PET artifact was limited to the thoracic veins containing undiluted iodinated contrast agent. No artifacts were present in other body regions. This finding suggests that the contrast-related PET artifact is dependent on threshold and requires very high contrast agent concentrations, which are achieved during passage of only the undiluted contrast bolus. Hence, the presence of such an artifact outside the thoracic veins is highly unlikely.

The artifact was present in 4 of 30 evaluated whole-body examinations. In all 4 patients, the inflowing veins contained highly concentrated contrast agent with a mean density of 2,262 HU. The lowest density resulting in an artifact amounted to 1,878 HU. In contrast to these 4 patients, the density of the intravenous contrast agent in the inflowing subclavian veins of the 26 patients without the artifact amounted to a mean of 1,058 HU. The density values did not overlap, resulting in a highly significant statistical difference (P < 0.001) between these 2 groups.

Compared with the body surface areas of patients in whom the PET artifact was not present (mean, 2.0 ± 0.18 m²), the body surface areas of all 4 patients expressing the artifact were low (mean, 1.67 ± 0.11 m²). This difference proved statistically significant (P = 0.035). Adaptation of the amount of contrast agent in patients with a lower body surface area may therefore be necessary. There are, of course, other factors influencing contrast agent kinetics, such as cardiovascular function, blood volume, or venous anatomy, making it difficult to predefine the correct amount of required contrast agent. Increasing the delay between contrast injection and CT acquisition may reduce the risk of artifact development but goes along with decreased scan quality due to acquisition of the scans in a suboptimal enhancement phase. The acquisition of an unenhanced CT scan as part of the combined PET approach would, of course, avoid the artifact due to contrast agent administration but would reduce the CT to merely a means for anatomic correlation without its own diagnostic potential. An artifact induced by the temporary presence of highly concentrated iodinated contrast agents within thoracic veins will likely continue to be a problem in contrast-enhanced dual-modality PET/CT examinations.

By coregistration of CT and PET images, a hot spot in the PET image can be directly attributed to the highly contrast-enhanced blood vessel, thus identifying this region of focal tracer uptake as an artifact and preventing misinterpretation. Pathologic changes close to the blood vessel causing the artifact may, however, be missed on film evaluation, as focal tracer uptake within a small lesion may be attributed to the artifact. Evaluation of PET images uncorrected for attenuation may be necessary in these cases. Close evaluation of CT and PET images in the area of the artifact will nevertheless be crucial to prevent diagnostic errors.

CONCLUSION

Contrast-enhanced dual-modality PET/CT examinations may result in a PET artifact of apparently increased tracer uptake due to the transient bolus passage of an undiluted intravenous contrast agent. The evaluating physician has to be aware of this artifact to prevent false-positive interpretations.