Radiation Exposure of Patients Undergoing Whole-Body Dual-Modality ¹⁸F-FDG PET/CT Examinations

MATERIALS AND METHODS

We reviewed whole-body PET/CT acquisition protocols used between September 2003 and May 2004 in 4 German university hospitals. Table 1 summarizes the main technical details of the 4 different PET/CT models installed in these hospitals. ¹⁸F-FDG PET scans were characterized by the administered activity and the scan time; CT scans were characterized by the tube potential U, electrical current-time product Q_el, volume CT dose index CTDI_vol, scan length L, slice collimation h_col, and pitch factor p.

RESULTS

Table 2 gives an overview of the routine acquisition protocols used for whole-body ¹⁸F-FDG PET/CT examinations at the 4 university hospitals (designated H1–H4). For each protocol, the type and sequence of the various scans performed as well as the effective doses per scan and examination are listed. In hospital H3, a high-quality protocol is used in the majority of cases. However, in cases in which a recent diagnostic CT scan exists, the high-quality diagnostic CT scan (D-CT) is replaced by a low-dose scan acquired without intravenous contrast medium (LD-CT). At sites H2 and H4, no high-quality diagnostic CT scan is performed as part of the combined PET/CT examination in such cases. The effective dose values for the 4 high-quality PET/CT protocols (Table 2) were nearly identical. The uterine dose, which is often used to estimate exposure to an embryo in the early stage of pregnancy, was between 20.9 and 23.2 mGy.

Average ¹⁸F-FDG activities of 300 MBq (H2) and of 370 MBq (H1, H3, and H4) were administered, which resulted in estimated effective doses of 5.7 and 7.0 mSv, respectively. The acquisition time for the whole-body ¹⁸F-FDG PET scans was <45 min at all sites. The scan parameters used for the different CT scans are specified in Table 3. Because the symphysis was defined as the lower limit of the CT scan range, the testes were not in the imaged body region. Nevertheless, they were exposed by scattered radiation and due to the overranging effect to a varying amount (0.7–7.2 mGy). At the upper side, the thyroid was within the scan region in all cases. For a more detailed assessment of the dose distribution within the human body, dose coefficients for the relevant organs are listed in Table 4 for both ¹⁸F-FDG PET and CT examinations. Estimated CT dose coefficients for some representative organs are plotted in Figure 1 along with the corresponding mean values.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

CT dose coefficients Γ_T^CT for some representative organs. Symbols give dose coefficients determined according to Equation 2 for each of the 7 CT scans listed in Table 3, whereas horizontal lines indicate corresponding mean values. RBM = red bone marrow.

For the 7 CT protocols used (Table 3), the effective dose was calculated on the basis of Equation 2 using the mean dose coefficient of Γ_E^CT = 1.47 ± 0.02 mSv/mGy given in Table 4. The resulting dose values are plotted versus the corresponding values determined from the TLD measurements on the Alderson phantom in Figure 2. Linear regression analysis (SigmaPlot, version 7.101; SPSS Inc.) yielded a slope of 1.03.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Statistical relation between calculated and measured effective doses for 7 CT scans listed in Table 3. Solid line gives result of linear regression analysis through origin with a slope of 1.03 and dashed curves indicate 95% confidence interval. Error bars indicate the uncertainty of dose estimates.

DISCUSSION

The effective dose for patients undergoing high-quality whole-body ¹⁸F-FDG PET/CT examinations at the 4 university hospitals participating in this study was about 25 mSv. Despite the similarity of the effective dose values (23.7–26.4 mSv), there were some noticeable differences between the 4 PET/CT acquisition protocols, which are representative of the imaging scenarios reported in the literature. Mainly, the 4 clinical sites (H1–H4) had a different approach to the clinical implications from the CT scan of the combined PET/CT examination.

At 2 hospitals (H2 and H4), separate low-dose CT scans were acquired for attenuation correction of emission data in addition to a contrast-enhanced CT scan. At the other 2 sites (H1 and H3), a single, contrast-enhanced CT scan was used both for a fully diagnostic evaluation and for CT-based attenuation correction. This may imply the question of whether the administration of an intravenous CT contrast agent leads to serious artifacts in the attenuation-corrected PET images, since structures with a strong enhancement in the CT scans may be considered as bone by the attenuation correction algorithm, thus resulting in an overestimation of regional attenuation coefficients (14). However, recent evidence indicates that these artifacts rarely cause a diagnostic challenge in the clinical setting (15) and that these artifacts can be avoided prospectively when using adapted contrast administration protocols (16).

Nevertheless, if a contrast-enhanced diagnostic CT scan has already been performed on a conventional CT system as part of the regular clinical work-up, it is in general acceptable to acquire only a low-dose CT scan as part of the combined PET/CT study (17). The image quality of this scan is certainly adequate for anatomic correlation and attenuation correction (18). In the present study, the effective dose determined for 3 low-dose scans was <5 mSv (Table 2).

The effective doses determined for the 4 high-quality CT scans listed in Table 3 varied between 14.1 and 18.6 mSv. These values are somewhat higher than the dose estimates (mean ± SD) of 14.5 ± 4.9 mSv from a recent survey on whole-body, multislice CT examinations (19), which is mainly due to the inclusion of the thyroid in the whole-body scan range covered in this study.

The dose coefficients listed in Table 4 make it possible to estimate organ doses and—using the corresponding tissue weighting factors—effective doses related to whole-body ¹⁸F-FDG PET and CT scans. All data presented are for a standard person with a body weight of about 70 kg and are generic rather than patient specific since the age, sex, and constitution of individual patients are not considered. Nevertheless, they provide a reasonably good indicator of the relative radiation risks to patients (12) resulting from nonuniform exposures related to whole-body PET and CT procedures and, thus, for protocol optimization.

PET/CT users should note that the CTDI_vol value displayed on the operator’s console is the principal descriptor to characterize patient exposure in CT on a local dose level. It represents an estimate of the average dose within an irradiated slice of a standardized CT dosimetry phantom and, thus, reflects not only the combined effect of the selected scan parameters but also of scanner-specific factors such as beam filtration, beam-shaping filter, geometry, and overbeaming. A detailed discussion of the various scan parameters and system features determining patient exposure in CT as well as strategies for dose reduction can be found elsewhere (19,20). Besides the CTDI_vol, the length of the scan region is the second parameter that determines the effective dose and, thus, the integrated detriment to patients related to a CT examination. Whenever clinically justifiable, the range of whole-body scans should be limited by the symphysis at the lower limit and should exclude the eye lenses from the cranial imaging range.

However, adaptation of the scan length to the individual body size may not be possible at current PET/CT systems because the axial CT range can be set up only in integer multiples of the fixed axial field of view of the PET system. This technical limitation can be overcome in the future, for example, by the implementation of continuous bed motion for PET measurements. In general, noncongruent imaging ranges of PET and CT scans, as well as multiple contiguous spirals with different CT scan parameters, should become available with the clinical PET/CT acquisition software. This flexibility would open the possibility of acquiring a high-quality CT scan for only part of the body and imaging the remaining axial ranges with a low-dose CT, or even without attenuation correction. Moreover, prospective measures that offer the potential for dose reduction in CT without a considerable loss in image quality—such as automatic tube current modulation or adaptive filtering—should be adopted for routine PET/CT.

CONCLUSION

The PET/CT acquisition protocols examined in this study reflect the range of whole-body PET/CT imaging scenarios reported in the literature today. We estimated an average effective patient dose from whole-body ¹⁸F-FDG PET/CT examinations of about 25 mSv independent of the acquisition protocol preferred. Considering the increased patient exposure compared with individual CT or PET examinations, a judicious medical justification has to be made with every PET/CT referral. The derived dose coefficients provide a valuable tool for estimating organ and effective doses for a diversity of whole-body CT scans and, in turn, for protocol optimization. Independently, prospective dose reduction measures from state-of-the-art CT practice should be adopted in PET/CT imaging, and modifications to the existing acquisition software should be considered.