Vascular Inflammation Imaging with ¹⁸F-FDG PET/CT: When to Image?

MATERIALS AND METHODS

Image Analysis

Coregistration and image analysis were performed using a Xeleris (GE Healthcare) workstation. PET/CT images were reviewed by a combined radiologist and nuclear medicine physician and a senior technologist in consensus. The area of most intense aortic wall ¹⁸F-FDG uptake was identified, and regions of interest (ROIs) were drawn over the abdominal aortic wall and lumen (Fig. 1). The maximum activity concentration for each region was recorded in each 2-min frame. For each period of dynamic imaging (four 2-min frames), the mean maximum activity concentration (corrected for decay) was recorded and converted to maximum standardized uptake value (SUV_max) normalized to body weight, calculated using the following formula:and the associated SE (SD/√4) was derived. This calculation was repeated in each of the 17 patients, and mean SUV_max for the aortic wall and lumen and their target-to-background ratio (TBR) were derived.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1. 

Fused axial PET/CT of ROIs applied to aortic aneurysmal wall and lumen at mid-point of dynamic acquisitions at 45 (A), 60 (B), 120 (C), and 180 min (D) after injection of ¹⁸F-FDG.

RESULTS

The mean aortic wall SUV_max at 45 min was 2.08, at 60 min it was 2.15, at 120 min it was 1.62, and at 180 min it was 1.99. The mean aortic lumen SUV _max at 45 min was 2.30, at 60 min it was 2.40, at 120 min it was 1.74, and at 180 min it was 1.70. The mean wall-to-lumen ratio (TBR) at 45 min was 0.91, at 60 min it was 0.91, at 120 min it was 0.96, and at 190 min it was 1.01 (Table 3).

The SDs and SEs for aortic wall SUV_max and lumen SUV_max and TBRs at 45, 60, 120, and 180 min after injection of ¹⁸F-FDG are shown in Table 3.

A significant difference in aortic wall SUV_max and lumen SUV_max with time (repeated measures ANOVA, P = 0.02 and P < 0.001, respectively) and no significant difference in TBR with time (repeated measures ANOVA, P = 0.206) were observed.

At the 2 specific time points of interest, there was no significant difference between SUV_max at 60 and 180 min in the aortic wall (paired t test, P = 0.367). There was a significant difference in SUV_max in the lumen at 60 and 180 min (paired t test, P = 0.001). There was no significant difference in TBR SUV_max between 60 and 180 min (paired t test, P = 0.131).

DISCUSSION

We conducted a study in a series of 17 patients with abdominal aortic aneurysms, which are strongly associated with both local and systemic atherosclerosis (11), to determine the optimal time to image vascular inflammation using ¹⁸F-FDG PET/CT. Multiple–time point PET of arterial ¹⁸F-FDG uptake showed that delayed imaging at 3 h, compared with earlier imaging, does not significantly change the TBR. There was also no difference in aortic wall SUV_max at 1 h, compared with at 3 h. Blood-pool activity as measured by lumen SUV_max was significantly different at 1 and 3 h, but this did not affect the TBR.

Given that there is increasing use of ¹⁸F-FDG uptake to assess arterial inflammation and atheroma vulnerability and to monitor the effects of pharmacologic therapies (1–9), these findings are likely to be relevant to this emerging technique. In addition, showing that there is no significant advantage of delayed imaging could favorably impact clinical practice. The prior recommended imaging time of 3 h (1) that has been implemented by others (2) could limit the feasibility in some centers of performing future screening or drug monitoring using ¹⁸F-FDG PET. By imaging at 1 h, this should allow a better workflow for imaging departments and make the ¹⁸F-FDG PET examination more acceptable to the patient, which is an important factor for any test.

We have also shown that the measurement of ¹⁸F-FDG arterial wall uptake becomes more variable with delayed imaging, most probably because of the resulting decay of the tracer and the subsequent increase in image noise. Therefore, the possible benefits of delayed imaging are counterbalanced by the effects of image noise and more variable SUV_max measurements. Given the relatively subtle changes involved in the imaging of atherosclerotic lesions with ¹⁸F-FDG PET, such variability may negatively affect the findings if imaging is delayed.

Imaging at 2 h showed a decline in ¹⁸F-FDG arterial wall uptake, a finding that was replicated in all but 6 of our patients. Although it is hard to explain such an uptake pattern, this finding could raise concern about the recommended use of imaging at 90 min (8).

Although PET has exquisite sensitivity, it does have limited spatial resolution. Therefore, using this technique to examine the arterial wall is a challenge, and it can be difficult to accurately localize the site of uptake from PET. This is one of the reasons for choosing to examine the aorta in this study, rather than the smaller carotid arteries. The addition of CT when using a hybrid PET/CT camera for image acquisition enables the exploitation of the superior spatial resolution of CT and thus improves the anatomic certainty of the site of ¹⁸F-FDG uptake. In particular, the use of a hybrid system helps differentiate arterial wall uptake from uptake in adjacent structures such as the lumen, which should help in the placement of ROIs and reduction of partial-volume effects. The CT used for image fusion was unenhanced; therefore, the thrombus within the sac of the aneurysm may make identification of the lumen more difficult with unenhanced CT, compared with contrast-enhanced CT. The original arterial wall uptake data suggesting that 3 h was the best imaging time (1) were acquired from the carotid arteries using a standalone PET camera, and the findings of this study may have been influenced by these factors of poor spatial resolution, partial-volume effects, and inaccurate ROI localization.

There has been an inconsistent use of methodology when performing PET vascular studies (Table 1). Different investigators have used a variety of strategies in expressing arterial ¹⁸F-FDG uptake, including visual assessment, absolute SUV_max, SUV thresholds, and TBR. Likewise, there has been a variety of imaging times used, ranging from 45 min (9,12) to 3 h. These inconsistencies probably reflect multiple factors, including habit and the demand for camera availability (13). Therefore, as is being sought (8), there is a need for the uniformity of methodology for arterial ¹⁸F-FDG studies.

One recent vascular ¹⁸F-FDG PET study did obtain dual–time point imaging data (12). In this study, it was noted that the delayed images provided better lesion-to-background contrast during visual assessment. However, the pattern and location of ¹⁸F-FDG uptake could be identified on both early and delayed images in all patients with significant carotid stenosis. Moreover, because the ¹⁸F-FDG uptake on delayed images greatly decreased in the control subjects, making it difficult for ROI placement along the arterial wall, the 45-min images were used for SUV comparison.

In the current study, we performed dynamic imaging during certain time points; however, it would have been ideal to have obtained continuous dynamic images over 3 h. In practice, with elderly patients, continuous dynamic imaging may be difficult to achieve; moreover, the images would be at an increased risk of movement-induced artifacts. It would have been advantageous to have performed studies on more than the 17 patients studied. Nonetheless, such studies are time-intensive, and most of the present ¹⁸F-FDG PET arterial studies have used similar-sized or smaller study populations (Table 1).

CONCLUSION

Our prospective aortic wall PET data from 17 patients showed that there was no significant advantage in imaging at 3 h over 1 h after ¹⁸F-FDG injection. Given the increasing use of vascular ¹⁸F-FDG PET studies for risk stratification and treatment monitoring, this finding has implications for patient throughput and acceptability.