Life Time Attributable Cancer Risk Estimated Using Scanner Reported Dose Length Product during Chest Computed Tomography Imaging in Young Children ()
1. Introduction
Children are generally considered to be at higher risks of developing radiation-induced tumors because of the young age of exposure and increased tissue radio sensitivity in some of the organs [1] . Children are clearly more likely to develop one of a quarter of these types, including leukemia, brain, breast, skin, and thyroid cancer [2] .
For a range of x-ray examinations, including radiography and CT, average lifetime risks of cancer incidence per [Sv] may be around two or three times higher for exposures at age 0 - 9 yr than at age 30 - 39 yr. For patients exposed in their 60 s, the estimated lifetime risks are about half those for patients in their 30 s, falling to about one-tenth for those in their 80 s [3] .
Monitoring radiation doses from medical exposures is highly recommended by international radiation protection standards [4] .
The stochastic risk is the estimation of the potential increase in the number of cancers resulting from the CT diagnostic examinations of the chest in young children. The risk is proportional to the calculated Effective dose (ED) of the chest CT and the dose length product (DLP), the age and the gender of the patient. The ED will depend on the radiation received by the irradiated organs during the CT scan and the radio-sensitivity of the organs.
the weighted sum of the tissue-specific dose equivalents, called the effective dose, should be proportional to the total estimated detriment from the exposure, whatever the distribution of equivalent dose within the body [5] .
The LAR estimates are based on the same epidemiological data as ICRP uses for the risk coefficients related to effective dose, and differentiate the cancer risk into age and gender specific subgroups and have also a clearly defined detriment in the form of either the excess risk of receiving a cancer or the excess risk to die from the received cancer.
In this study we examine the stochastic risk form radiation exposure of chest computed tomography (CT) examinations routinely performed on pediatric patients in a tertiary care medical institution.
2. Methods
Data were collected using the Radiology information system (RIS) hospital information system (HIS) and the DICOM dose report files for each chest CT examination included in this study. The data collected were for both genders and for ages from less than a year to fifteen years of age. Technical parameters such as tube kilo-voltage (kV) and volumetric computed tomography dose index (CTDIvol) are included in the analysis along with the gender and the age group. The study included 125 patients in total. The data analyzed are from two scanners.
To generate effective dose per unit dose length product (ED/DLP) conversion factors incorporating ICRP Publication 103 tissue weighting factors must be used.
Calculating organ doses using software for each patient is laborious and time consuming; as an alternative, scanner-derived exposure outputs, volume CT dose index (CTDIvol), and dose-length product (DLP) that are recorded for each examination have been used to approximate effective dose according to DLP, using DLP-to–effective dose (ED) conversion coefficients (k factors) [6] [7] .
ED in [mSv] = DLP in [mGy∙cm] * k in [mSv∙mGy−1∙cm−1] (1)
LAR = ED in [Sv] * risk coefficient in [Sv−1] (2)
ICRP nominal risk coefficient for detriment-adjusted cancer is 5.5% per Sv [5] . The k factors are taken from Table 1. The ED and LAR calculations in this work were conducted using Equations (1) and (2) above.
The statistical analysis and plots in this work was conducted using Matlab version R2016b, statistical and machine learning toolbox.
3. Results
The results for each one of the two CT scanners used in this study are included in Table 2 and Table 3.
The LAR values can be compared with the Values reported in Table 4. Themost significant parameters in the form of distributions are included in Figure 1 in the form of histogram. Figure 1(a) showing the DLP distribution for our data in the form of histogram. Figure 1(b) showing the ED values distribution. And Figure 1(c) contains the kV values used in chest CT examinations with 100 kV as the most common value used in our institution followed by 80 kV and finally few exams are done with 120 kV. The LAR values distribution are in Figure 1(d).
The summary of our results can be expressed in Figure 2 which illustrates an incresing value of the DLP with age. And Figure 3 showing that smaller values of CTDIvol and DLP corresponds to lower LAR value. Figure 4 showing the LAR values obtaine for the 4 age groups, the difrence is quite small among the groups. Figure 5 clearly demostrates that an increase in the kV value will increase the LAR estimate. Therfore it is important to specifiy the used kV value in order to compare published reserch results reporting ED of LAR.
4. Discussion
Effective dose calculation is more complex and prone to error in children compared with adults because of the use of various correction factors to convert adult to pediatric dose [10] .
The k coefficient method usually under estimate the real ED value and the need to improve the LAR assessment reside in the fact that ED must be calculated using MC based software to calculate organ doses and the corresponding ED for the chest CT exam for example. The coefficients will decrease in magnitude with increasing age.
Adopting a retrospective study in a large database, the authors’ findings demonstrated consistency that organ doses are the best quantity to assess radiation risks, emphasizing that risk estimations must be interpreted as average estimations applicable to patient groups and not for individuals [11] . In all cases, this outlines the importance of patient-based radiation dose monitoring for
![]()
Table 1. Normalized effective dose (ED) in mSv per dose-length product (DLP) in mGy−1∙cm−1 for pediatric patients of various ages for chest CT examinations published in [7] .
![]()
Table 2. Data are for scanner.1 (Siemens, N = 49) the reported values are the calculated average ± (standard deviation).
![]()
Table 3. Data are for scanner.2 (GE, N = 76) the reported values are the calculated average ± (standard deviation).
![]()
Table 4. Lifetime attributable risk of cancer incidence (LARCI) per 100,000 persons exposed to a single dose of 0.1 Gy [8] .
*values reported in BEIR VII, 2006 report [9] .
![]()
Figure 2. Boxplot of the DLP values per age group 1 (from 0 to 1) years, group 2 (more than 1 and less than 5), group 3 (more than 5 and less than 10), group 4 (more than 10 and less than 15.
![]()
Figure 3. The combined effect of CTDIvol and DLP on the LAR values grouped as more than 1% and less than 1%.
![]()
Figure 4. Boxplot of the obtained LAR value per age group.
![]()
Figure 5. Boxplot of the LAR values in (%) as a function of the kV.
patient risk assessment, in particular for CT exams [12] . However, the k factors are estimated with a sexless, ageless, uniform phantom and do not represent the varied attributes and organ distributions seen in patient populations [13] .
The obtained results are comparable with other published studies and contribute to the existing literature about radiation dose estimates and analysis of chest CT examination in young children [14] [15] [16] [17] .
Among the examined scanning parameters only the kV used had a significant effect on the estimated LAR as seen in Figure 5.
Multiplying the universally reported DLP by scalar coefficient offers a simple, rapid method for clinicians to estimate ED, which can be included in the medical record and tracked over time [18] .
5. Conclusion
The DLP based method used to estimate the excess cancer risk from radiation exposure occurring during medical chest CT examinations is straightforward and provide practical approach to risk estimation and a method to monitor radiation dose levels applied to pediatric patients. In order to obtain a better ED and LAR estimations organ doses must be calculated using commercially available softwares based on Monte-Carlo simulation considered the gold standard in clinical dosimetry; we recommend the use of such approach in another research work similar to this one. Also analyzing a larger number of CT dose data will improve the accuracy of the estimation.