Multisection CT Protocols: Sex- and Age-specific Conversion Factors Used to Determine Effective Dose from Dose-Length Product

Introduction

Computed tomography (CT) is the most substantial contributor to the collective effective dose for all radiographic procedures. Surveys (1–5) have shown that the contribution of CT to the total collective dose can be as high as 67% in the United States and 40% in Europe, although CT examinations represent only about 11% and 4%, respectively, of all radiologic examinations. Therefore, the dose levels delivered in CT examinations should be known and available to patients and their physicians. The effective dose is primarily used to compare the stochastic risk associated with the exposure to ionizing radiation. According to the International Commission on Radiological Protection (ICRP) (6,7), effective dose represents a weighted sum of the equivalent doses in all tissues and organs of the body, where the equivalent dose for an organ represents the sum of the absorbed dose averaged over a tissue or organ weighted by the radiation weighting factor (ie, 1 in case of photons and electrons).

Generally, the effective dose is computed by using Monte Carlo dose simulation tools for reference human phantoms (8). The practical method that is most often used in clinical routine is to estimate the effective dose from dose-length product (DLP) measurements. In 1999, the European Commission published a set of conversion factors for different regions of the body for adults (9). These conversion factors are often cited in the literature when the effective dose is computed from the DLP. However, published conversion factors do not consider differences in scanning voltages, specific sex issues, age, or variation in body size and shape. Moreover, according to the ICRP publication 103 recommendations, new weighting factors and organs at risk have been defined (Table 1). It is worth mentioning that the conversion factors used in practice today were defined on the basis of Monte Carlo simulations for single-section CT scanners. Consequently, conversion factors should be revisited and possibly updated.

Currently, there is a lack of published data on conversion factors considering the new generation of CT scanners (eg, 64-section scanners) for adult patients and even less data for pediatric patients (10,11). Huda et al (10) reported conversion factors from DLP to effective dose for adult phantoms after the use of several dosimetry tools and 16-section CT scanners from various manufacturers. However, their published data do not contain references for pediatric patients, and the conversion factors were computed on the basis of ICRP publication 60 recommendations with respect to effective dose. Shrimpton (11) extended the conversion factor computation to pediatric patients by using mathematical phantoms that represented newborns and children of 1, 5, 10, and 15 years. The Monte Carlo computations were based on three models of scanners representing the first generation of CT scanners, and the conversion factors were computed on the basis of the old ICRP recommendations (6).

The aim of our study was to determine conversion factors for the new ICRP publication 103 recommendations for adult and pediatric patients and to compare the effective doses derived from Monte Carlo calculations with those derived from DLP for different body regions and scanning protocols.

Materials and Methods

No industry support was received for this study. One author (W.A.K.) is a consultant of Siemens Healthcare (Forchheim, Germany). Authors P.D.D. and Y.S. had control of inclusion of any data or information that might present a conflict of interest for the consultant author.

CT Scanner

All dose simulations were performed for scanner geometry, spectra, and filtration equivalent to those of a Somatom Sensation 64 CT scanner (Siemens Healthcare).

Phantoms

The phantoms used for our study were developed in our institution and were based on the Oak Ridge National Laboratory (ORNL) phantom series described by Cristy and Eckerman (12) that used a format created previously by Snyder et al (13) and Cristy (14). The series consists of mathematically defined phantoms that mimic newborns; 1-, 5-, and 10-year-old children; and adults.

For all phantoms, the organs that contribute to the effective dose as defined in ICRP publication 60 (6) were modeled. However, ICRP publication 103 (7) introduced organs and tissues that were not previously considered (eg, the salivary glands, extrathoracic tissue, lymph nodes, prostate, oral mucosa, and the nasal vestibule). Therefore, the original phantoms were modified to include the new organs and tissues. The organs were modeled by using a set of surface equations, such as those for ellipsoids, cylinders, and spheres, that were also used in the original ORNL phantoms (Fig 1a). The masses of the above-mentioned organs were determined by using ICRP publication 89 (15) and International Commission on Radiation Units and Measurements report 48 (16). Further details regarding the description of the organs are presented in Appendix E1 (online).

Figure 1a:

Download as PowerPointOpen in Image Viewer

Figure 1b:

Download as PowerPointOpen in Image Viewer

Dosimetry Tool and Simulation Setup

The Monte Carlo calculations were performed by using a validated dosimetry tool (ImpactMC; CT Imaging, Erlangen, Germany) for clinical CT scanners. Details regarding implementation and tool validation have been reported elsewhere (17). The dose simulations were performed according to the data in Tables 2 and 3 with respect to scanning protocol and scanning length, respectively. Thus, for each phantom, five scanned regions were defined according to a head, neck, chest, abdomen, and pelvis examination. Table 3 summarizes the irradiated lengths as a function of age, sex, and body region, whereas in Figure 1b, the anatomic landmarks of the regions are depicted by colors for the adult subjects. Thus, the regions were defined as follows: (a) the head (the rostral part, comprising the brain, eyes, ears, nose, and mouth), (b) the neck (from the end of the head region to the beginning of the torso), (c) the chest (from the thoracic inlet to the thoracic diaphragm), (d) the abdomen (from the thoracic diaphragm to the pelvic inlet), and (e) the pelvis (from the pelvic inlet to the pelvic diaphragm). The same landmarks were also used for pediatric subjects. For each region, four spiral scanning protocols (pitch, 1.0; collimation, 19.2 mm) were simulated for voltages of 80, 100, 120, and 140 kV. All other scanning parameters were kept constant during simulations. A total of 200 simulations were performed with a precision of 1% or greater.

After the simulations, the three-dimensional dose distributions were used together with organ index files to compute the organ doses as the average values over the voxels having the same index. The organ dose values were tabulated and used to calculate the effective doses and to estimate appropriate conversion factors.

Conversion Factor Estimation

For the calculations of effective dose, the old (ICRP publication 60 [6]) and the new (ICRP publication 103 [7]) published organ and tissue weighting factors were used.

The conversion factors (CFs) were determined according to the following equation:

where E is the effective dose, R denotes the scanned region, and G and A stand for sex and age, respectively. The DLPs were determined as the product of the volumetric CT dose index (CTDI_vol) and the irradiated length of region R (L_R), as follows:

and

where CTDI_w is the weighted CTDI measured in standard acrylic head and body phantoms of 16- and 32-cm diameters and p is pitch (9). For our study, with spiral examinations with a pitch factor of 1.0, the CTDI_vol was equal to CTDI_w. Table 4 gives the values of normalized measured CTDI_w as a function of voltage in head and body phantoms, whereas summarizes the irradiated lengths. For the head and neck regions, the DLP was computed from the CTDI_w measured in the 16-cm head phantom, whereas the DLPs for the chest, abdominal, and pelvic regions were determined from the CTDI_w measured in the 32-cm body phantom for both adult and pediatric individuals.

The conversion factors for each phantom and scanned region obtained from our Monte Carlo calculations were compared with the conversion factors published by the European Commission (9).

Results

Conversion factors calculated by using ICRP publication 60 as a reference led to larger differences with respect to the European Commission factors than with respect to the ICRP publication 103 factors for all ages (Table 5). For pediatric subjects, the conversion factors were much larger than the European Commission values for all scanned regions. The relative difference increased with the decrease in age. Values up to 4.7 times higher were computed for the abdominal region in the newborn phantom when the scanning and exposure parameters were kept constant. The largest differences were observed in the abdominal region for the newborn, 1-year-old, and 5-year-old phantoms; an exception was the value for a 5-year-old subject computed by using ICRP publication 60 recommendations. For 10-year-old subjects, the largest difference was found when scanning the neck region.

For adult subjects, for the head, chest, and pelvis, respectively, the simulated conversion factors were 33%, 20%, and 12% lower than the European Commission conversion factors for ICRP publication 60 and 18%, 14%, and 32% lower for ICRP publication 103. For the neck and abdomen, the conversion factors were slightly higher (by up to 6%) than the European Commission conversion factors, except for the neck when ICRP publication 103 recommendations were used (Fig 2a).

Figure 2a:

Download as PowerPointOpen in Image Viewer

Figure 2b:

Download as PowerPointOpen in Image Viewer

As expected, conversion factors decreased with increases in age and voltage (Fig 2b). For pediatric subjects, the difference between conversion factors when the voltage was varied from 140 to 80 kV decreased with an increase in age. For example, for newborns, the conversion factor was approximately 19% larger at 80 than at 140 kV. For adult subjects, no major difference was observed between conversion factors when the tube voltage was varied.

Figure 3 presents the mean conversion factors, averaged over sex and voltage, for pediatric subjects computed with ICRP publication 60 (Fig 3a) and ICRP publication 103 (Fig 3b), together with the European Commission conversion factors. Generally, higher conversion factors compared with published values were found for the chest, abdomen, and pelvis.

Figure 3a:

Download as PowerPointOpen in Image Viewer

Figure 3b:

Download as PowerPointOpen in Image Viewer

For the chest, abdomen, and pelvis, the conversion factors were higher for the adult female phantom than for the adult male phantom for both ICRP publications (Fig 4). The largest differences were observed in the pelvis (approximately 97%) for ICRP publication 60 and in the chest (approximately 76%) for ICRP publication 103.

Figure 4:

Download as PowerPointOpen in Image Viewer

Discussion

Although the dose levels in CT examinations are generally well below the threshold dose for inducing deterministic effects, they may have an influence on the stimulation of gene mutations and carcinogenesis (6). Because of the cumulative effects of radiation dose, special attention has to be given, especially when repeated examinations are conducted (18,19). The quantity that tries to relate dose to the risk associated with radiation exposure and thus the correlation with stochastic effects is the effective dose (20,21). The effective dose cannot be directly determined by measurements. It is generally computed by applying Monte Carlo methods to data gathered in reference human phantoms and requires dedicated dosimetry tools and detailed knowledge of photon transport mechanisms. Therefore, in practice, alternative approaches such as the correlation between the DLP and effective dose are used.

We have studied the influence of sex, age, tube voltage, and ICRP recommended tissue weighting factors on conversion factors used to determine the effective dose from the DLP. The results were compared with the published conversion factors that are commonly used in clinical practice (22,23). Although the findings are interesting and, in some cases, reveal substantial differences from the published conversion factors presently in use, our primary intention was to give insights regarding the dependence of conversion factors on voltage, age, or sex rather than claiming the replacement of existing reference values.

Our computations indicate that, for adult subjects, the conversion factors published in 1999 (9) systematically estimate higher effective doses for the head, chest, and pelvis, independent of the organ weighting factors used in the calculation of the effective dose. For example, differences of up to 33% and 32% were found for the head and pelvic examinations, respectively. This may be related to the fact that the published conversion factors were based on calculations performed for CT scanners that were in use in the early 1990s. Our results are in good agreement with those reported by Shrimpton (11) for adult subjects for the neck, chest, abdomen, and pelvis using ICRP publication 60 recommendations. For the head, Shrimpton reported a value of 0.0021 mSv/mGy, compared with the value of 0.0016 mSv/mGy found in our study. Slight discrepancies can be explained by differences in the phantoms and scanned regions used.

With respect to pediatric subjects, the computed conversion factors are much larger than those published by the European Commission for all scanned regions and voltages. Our expectations were confirmed that the effective doses for children and, consequently, the conversion factors are higher than in adults when the exposure is kept constant. This is to be expected because in children, the organs responsible for the main contribution to the effective dose (lungs, liver, and gonads) receive higher doses at constant exposure factors than the organs of adults because of their smaller cross sections. In the case of ICRP publication 60, the highest conversion factors were found for the pelvic region, whereas for ICRP publication 103, the highest values were obtained for the abdominal region. This is related to the differences in organ weighting factors and effective dose calculation between ICRP publication 60 and ICRP publication 103. For example, in ICRP publication 103, the new weighting factor for the gonads is 0.08, whereas in ICRP publication 60, the weighting factor was 0.20.

Higher conversion factors, of course, do not necessarily mean higher effective doses, as the DLP has to be kept appropriately low in pediatric CT practice, which reduces the effective dose and offsets the effects of higher conversion factors. Compared with the values of Shrimpton (11), our computed conversion factors for children were lower on average by approximately 30% for the head and higher by 40%, 70%, and 83% for the neck, chest, abdomen, and pelvis, respectively. This can be explained by differences in DLPs (ie, different scan lengths) and phantom models. For our study, the DLPs were computed on the basis of CTDI_vol values measured in the 32-cm CTDI phantom for the chest, abdomen, and pelvis for both pediatric and adult subjects, whereas Shrimpton used the CTDI_vol measured in a 16-cm CTDI phantom for pediatric subjects.

Evaluation of the dependence of conversion factor on tube voltage showed that for adult subjects, the variation was within 2.6% for all scanned regions and therefore can be neglected. For pediatric subjects, the variations were larger; they may go up by 19% when lower voltages are applied. Results of the sex comparison showed generally that higher conversion factors were obtained for adult women, especially for the chest, abdomen, and pelvis. This was expected, owing to the distribution of the radiosensitive organs and tissues. Differences of up to 76% were observed in, for example, the chest region with ICRP publication 103 because of the increased weighting factor for the breast, whereas for the pelvic region, differences of up to 97% were observed when the conversion factors were computed on the basis of ICRP publication 60. For the first case (ie, the chest region and ICRP publication 103), the weighting factor for breast tissue increased from 0.05 in ICRP publication 60 to 0.12 in ICRP publication 103, yielding a larger conversion factor for adult women. For the second case, (ie, the pelvic region and ICRP publication 60), this can be explained by the scan range and the position of the male gonads. The male gonads were partially irradiated, while the female gonads were fully irradiated. This raises the question of whether conversion factors should be specified separately for each sex instead of using an averaged value that will result in an overestimation of effective dose for male patients and an underestimation of effective dose for female patients.

The conversion factors in our study were determined for spiral examinations performed by using a pitch factor of 1.0. They are valid for other pitch values as long as other scanning parameters and the scan length are kept constant. For shorter scan lengths and higher pitch values, the conversion factors may vary slightly owing to the position of the organs and tissues with respect to the scanning path.

The findings of our study were based on the use of only one multidetector CT scanner because the primary aim was to investigate the influence of scanned region, voltage, and patient age and sex on conversion factors. Because no appreciable dependence on tube voltage (ie, on energy) was found, it can be expected that the results are valid for other scanners. Differences in filtration, for example, will shift the effective energy less than the changes in voltage investigated here and should therefore not cause noticeable differences.

Variations in conversion factors also have to be expected when using other types of phantoms—for example, those derived from segmented image sets in patients (24,25). In this respect, ICRP publication 103 recommends the computation of organ and effective doses on the basis of acquired scans developed by Zankl et al (26) as reference models. To our knowledge, phantoms representing pediatric patients were not made generally available. This is another reason why established mathematical models were used for our study. Although they are modeled with sets of surface equations that approximate the form and volume of human organs, the advantage of mathematical phantoms is that they are more easy to implement and more flexible in use (8). However, they should be compared with new phantoms when these are made available.

Another aspect that was not considered in our study is related to the angular and longitudinal tube current modulation that is provided by all modern multidetector CT scanners (27). For all simulations, the tube current was kept constant. As was shown in a recent study (28), a reduction of up to 8% can be expected in adult conversion factors when tube current modulation schemes are taken into account. For pediatric subjects, especially for newborns, this reduction is smaller: One- and 5-year-old children present an approximately circular cross section rather than an ellipse, resulting in less efficient modulation. Nevertheless, further work that takes into account all aspects aimed at reducing radiation dose, such as tube current modulation and scanning protocol optimizations (29,30), may have to be performed. No noticeable effect is expected with respect to total collimation, z-overscanning, and adaptive z-collimation, because these effects are taken into account by the DLP.

Last but not least, an important aspect that has to be considered when the effective dose is determined on the basis of the DLP indicated on the scanner console is represented by the limitations of the CTDI as an indicator of dose. In practice, the CTDI is measured by using 10-cm ionization chambers. As shown by Boone (31), the CTDI is a good indicator of dose for a total collimation of up to 40 mm. However, wider collimations lead to underestimations of the dose, with consequences that are not covered at present in the conversion factors.

In conclusion, the conversion factors used to compute the effective dose from the DLP may have to be updated with respect to modern cone-beam CT scanners. They have to reflect the new ICRP recommendations and should be specified separately for both sexes. With respect to pediatric CT, new conversion factors should be established that also may have to include the influence of tube voltage.

Author Contributions

Author contributions: Guarantors of integrity of entire study, P.D.D., W.A.K.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, all authors; clinical studies, P.D.D., W.A.K.; experimental studies, P.D.D., W.A.K.; statistical analysis, all authors; and manuscript editing, all authors