Age-based diagnostic reference levels and achievable doses for paediatric CT: a survey in Shanghai, China

Computed tomography (CT) is extensively utilised in medical diagnostics due to its notable radiographic superiority. However, the cancer risk associated with CT examinations, particularly in children, is of significant concern. The assessment of cancer risk relies on the radiation dose to examinees. Diagnostic reference levels (DRLs) and achievable doses (ADs) were used to assess the level of radiation dose in CT examinations widely. Although the national DRLs of paediatric CT have been explored in China, few local DRLs at the city level have been assessed. To set up the local DRLs and ADs of paediatric CT, we investigated the radiation dose level for paediatric CT in Shanghai. In this survey, a total of 3061 paediatric CT examinations underwent in Shanghai in 2022 were selected by stratified sampling, and the dose levels in terms of volume CT dose index (CTDIvol) and the dose-length product (DLP) were analysed by 4 age groups. The DRLs and ADs were set at the 75th and 50th percentile of the distribution and compared with the previous studies at home and abroad. The survey results revealed that, for head scan, the DRLs of CTDIvol were from 25 to 46 mGy, and the levels of DLP were from 340 to 663 mGy·cm. For chest, the DRLs of CTDIvol were from 2.2 to 8.3 mGy, and the levels of DLP were from 42 to 223 mGy·cm. For abdomen, the DRLs of CTDIvol were from 6.3 to 16 mGy, and the levels of DLP were from 181 to 557 mGy·cm. The ADs were about 60% lower than their corresponding DRLs. The levels of radiation doses in children-based hospitals were higher than those in other medical institutions (P < 0.001). In conclusion, there was still potential for reducing radiation dose of paediatric CT, emphasising the urgent need for optimising paediatric CT dose in Shanghai.


Introduction
Computed tomography (CT) has been a powerful diagnostic imaging tool, widely used in medical diagnosis and treatment simulation positioning [1].UNSCEAR 2020/2021 Report estimated that about 400 million CT examinations were performed annually in the world, an 82% increase from that in the 2018 report.The global average annual frequency of CT examinations increased from 34 per 1000 population to 55 per 1000 population in the past 20 years [2].
Although CT can bring many medical benefits to patients facing many disorders, the potential radiation health risk should be considered because of its wide use and relatively high radiation dose.Of which, children exposed to higher cancer risk compared with adults because children are known to be more sensitive to radiation [3].Several epidemiological studies and related surveys have illustrated that the significant relationship between the radiation dose and relevant cancer risk to examinees (hereafter patients) [4,5].Moreover, the cancer risk is higher in infants and young children compared with the older [6].Therefore, the issue of radiation protection for paediatric CT is necessary for the society to avoid the possible risk.
The assessment of cancer risk is usually based on the radiation dose to examinees.The investigation of CT radiation dose is of importance for public in radiation protection.Diagnostic reference level (DRL) was proposed to assess the level of radiation dose in CT examinations by the International Commission on Radiological Protection (ICRP) in 1990 for the first time [7].The DRLs were benchmark and guidance for radiation protection by region, country, or even a city.DRLs are often taken as the third quartile value (75th) of the volume CT dose index (CTDI vol ) and dose-length product (DLP) [8].Generally, ICRP recommends utilising the DRLs to evaluate radiation doses for individual examinations to reduce unnecessary radiation exposure as possible [9].Hence, it is necessary for each country to establish itself DRLs in terms of different demographic characteristics in paediatric CT.Recently, detailed DRLs for paediatric CT examinations have been set up by the European Commission in 2021.Many other countries also established the DRLs for paediatric CT based on respective national demographic conditions [10][11][12][13][14].
Achievable dose (AD) was also proposed to find more typical value for radiation dose as an auxiliary mean in every CT examination by the United Kingdom National Radiation Protection Board, being set as the median or 50th percentile of CT dose distribution.ADs, along with DRLs, were used to conducted in establishment of guidelines for paediatric CT examinations in several studies [10,15].
Although DRLs in China for paediatric CT have been explored in the past several years [16], the local DRLs and ADs for paediatric CT in Shanghai have not been set up.The purpose of this study is to use the CT dose index to develop the DRLs and ADs for paediatric CT examinations in Shanghai, China.

Material and methods
The ethical review committee of Shanghai Municipal Center for Disease Control and Prevention approved this retrospective study (No. 2021-35).The requirement for informed consent was waived.

Data collection
This survey was conducted based on the deployment of Institute of Radiation Protection and Nuclear Safety Medicine, Chinese Center for Disease Control and Prevention.To assess the radiation dose levels of paediatric CT examinations in Shanghai through the survey of paediatric CT examinees.
According to comprehensive information system of radiological health in Shanghai, there were a total of 381 medical institutions conducting CT examinations and 776 CT equipment in 16 districts in Shanghai in 2022.Of which, 4 children's hospitals owned 19 CT equipment.Considering the distribution of paediatric visits in Shanghai, we conducted the investigation for understanding the medical visits of children before conducted this survey for CT dose.The result revealed that paediatric visits in Shanghai are predominantly concentrated in the four children-based hospitals, with relatively few paediatric cases in other medical institutions.
Thus, we mainly focused on the data of paediatric CT examinations in the 4 children's hospitals.Moreover, data of paediatric CT examinations from other medical institutions (non-children-based hospitals) were collected based on stratified sampling method upon districts and the levels of hospitals.According to compliance of hospitals in each district, 2 hospitals with the level II and III were investigated by Centers for Disease Control & Prevention in 16 districts.
To ensure the adequate sample size, we increased it by 50% above the minimum required by national standards to enhance the representativeness of the samples.Thus, not less than 30 paediatric CT examinations were analysed in each age group and in each examination region in 4 children-based hospitals.Due to relatively few examinations for children in other medical institutions, not less than 10 paediatric CT examinations were analysed.This survey was conducted using the same questionnaire of CT radiation dose.The questionnaire was approved by a professor of radiology and radiation protection officers from 16 districts in Shanghai.
Information including the patient demographic data and technical acquisition parameters data were exported from the medical imaging system equipped with the CT scanners.The patient demographic data were included in this survey as follows: the patient ID, age, sex, data of birth, etc. Technical acquisition parameters (manufacturer and equipment number of CT scanners, kV, mAs, length scan, pitch, etc) and Radiation dose metrics (CTDI vol and DLP) were included were also included from the dose reports after the CT examinations.
In this survey, the data of contrast-enhanced CT examinations were excluded.The examinations with invalid data, including missing data of age and sex, unclear scanning parameters and anatomical region were also excluded.The image quality of CT examinations was assessed and scored according to a five-point scale by two radiologists with >3 years of medical experience who were blinded to all CT parameters and patient data (1 = very poor with non-diagnostic, 2 = low image quality that non-confidence to make diagnosis, 3 = moderate image quality with confidence to make diagnosis, 4 = good image quality, 5 = excellent image quality).The score of 4 or more of the examinations were selected in this survey.

Statistical analysis
The number, minimum, maximum, median, 25th and 75th percentiles of CTDI vol and DLP were presented in this study in terms of different CT examinations and age groups.The distributions of dose indexes of CTDI vol and DLP were non-normal distribution tested by Shapiro-Wilk Normality Test.Thus, CTDI vol and DLP between different groups and the difference between children-based hospitals and other medical institutions were compared using Kruskal-Wallis test.A P value < 0.05 was considered as statistically significant.
The details of CT scanning protocols between the children-based hospitals and other medical institutions are listed in table 4. Higher kV values and more length scan are in other medical institutions, and the pitch factors in children-based hospitals are slightly higher.
The comparison of CTDI vol and DLP for head, chest and abdomen examination regions between children-based hospitals and other medical institutions is listed in tables 5 and 6.For each examination region, the average CTDI vol and DLP were significantly lower in children-based hospitals compared to other medical institutions considering the factor of age (P < 0.001).
Head, chest, and abdomen, as the commonly performed anatomical regions, were separately analysed in terms of the dose value based on CTDI vol and DLP.The frequency distribution of age based CTDI vol and DLP in 3 examination regions (head, chest, and abdomen) by box plot were shown (figures 1 and 2).
The comparisons of age based DRLs of the CTDI vol and DLP among different studies are shown in table 8. Compared to other countries, the DRLs of head, chest and abdomen in this survey were lower than those in Japan, but higher than those in America and France in terms of both CTDI vol and DLP.Compared to nationwide survey in China, lower value in CTDI vol and DLP were shown in this survey in the region of head, chest, and abdomen.
The comparisons of ADs for paediatric CT examinations between different studies are presented in table 9.The ADs of head and chest regions in this survey were close to the ADs in America and France in terms of CTDI vol and DLP, higher in abdomen.The ADs of head, chest and abdomen regions were lower than Japan and the previous nationwide study in China.
The number and distribution of the CTDI vol values and DLP are listed in tables 2 and 3.For each examination region, there were significant differences between the 4 age groups based CTDI vol and DLP from CT (P < 0.05).
The ADs and DRLs based CTDI vol and DLP in age groups are shown in table 7. DRLs of paediatric CT as well as ADs generally increased with the age.

Discussion
This survey conducted an analysis of 3061 paediatric CT examinations performed on 39 CT scanners in Shanghai to develop the DRLs and ADs.Chest, head, and abdomen CT were the most performed examinations, consistent with previous research findings [10,13].Moreover, the older age group exhibited the highest number of CT examinations.The results indicated an increasing trend in DRLs, particularly in terms of DLP, with increasing age for paediatric CT examinations of the head, chest, and abdomen.
Besides, the lower values of CTDI vol and DLP from children-based hospitals were observed in this survey in the three frequently performed examination regions.From the details of scanning protocols, the lower of kV values and length scan and higher pitch factors were observed in the four paediatric centres.Besides, more differences in head protocol between the children-based hospitals and other medical institutions, which could be attributed to the much higher mAs value in other medical institutions.Thus, the CT dose   index parameters might be set up for paediatric patients and the measures of parameter optimisation have been implemented in children-based hospitals.In contrast, most of CT dose index parameters were likely set only for adults in other medical institutions, potentially resulting in higher radiation doses compared to those in children-based hospitals.Moreover, from the DLP/CTDI vol ratios, the higher DLP in head region was mainly due to the higher CTDI vol in head region incorporating both box plots.For chest and abdominal region, the DLP/CTDI vol ratios stands out as a reflection of length scan.
There were certain differences between China and other countries in terms of the DRLs based on CTDI vol and DLP.The DRLs of head, chest and abdomen in Shanghai have been close to the level of other countries,  even lower than Japan.Compared to America and France, the DRLs in this survey were still higher.In general, the body sizes of children in China might be smaller, suggesting that there is still potential for reducing radiation dose in CT examinations in Shanghai.
Additionally, the DRLs of paediatric CT among head, chest and abdomen were lower in Shanghai from this survey compared to the sub-nationwide study in China, only slightly higher DRLs in abdomen in 10-<15 age group in terms of CTDI vol and DLP [17].It may indicate that the difference between CT parameter setting and medical practice among different provinces in China.
ADs for paediatric CT were also put forward in this survey.Unlike DRLs which is to avoid the patients from unnecessary exposure, ADs can serve as a benchmark for dose optimisation given that 50% of the doses can be lower than ADs in medical practice [18].ADs can provide greater optimisation potential for CT doses, and a 60% reduction in average dose was achievable in this survey if ADs were followed.
The strength of this survey was that the data was collected from hospitals in all 16 districts in Shanghai.The all 4 children-based hospitals and hospitals of level II and III were both analysed in this survey, which  could assess the level of radiation dose for paediatric CT in Shanghai.Besides, the differences of CTDI vol and DLP between children-based hospitals and other medical institutions were shown in this investigation for improvement of dose optimisation in practice.The limitations of this survey were that these DRLs and ADs were obtained from only 39 CT machines and 3061 paediatric patients in Shanghai.There were more than 400 CT scanners in Shanghai, and the CTs analysed in this survey were only less than 10%.More CTs and radiation information of children patients included should be collected in the further survey.Only 4 CT examination regions were analysed in this survey, and the DRLs and ADs of paediatric CT in more examination regions may be set up in the further study.What is more, the consideration of patients' size has not been shown in this study.In addition to the CTDI vol and DLP based on DRLs, the concept of size-specific dose estimates (SSDEs) [19] which consider the size of paediatric patients has been used to estimate the dose of CT in many countries, as a new method to summarise the DRLs of paediatric CT [10][11][12].Recently, some studies in Asia have also indicated that patient size as a factor may better reflect dose levels [18,[20][21][22][23][24][25][26][27].The method of SSDE may can be introduced into the further study.Besides, the relationship between the patient's age and CT dose needs to be considered in medical practice.However, many countries have further set up the subgroups from paediatric patients aged 0-5 years [10,[28][29][30].It considered that the rapid growth and development changes in children at this age.The age groups of paediatric CT examinations in line with their radiation dose should take into more careful consideration in the future work.
Overall, the local DRLs and ADs for paediatric CT in Shanghai were calculated from CTDI vol and DLP from the investigation of CT dose in all 4 children's hospitals and other hospitals, which could guide individual hospital in optimisation.It suggested that setting of child-specific CT scan parameters and scan length are key to optimising CT dose for children in this region.Besides, considering the ease and rigor of collecting samples from hospitals in this investigation, the radiological information system needs to be improved and put into serve in radiation dose survey building on existing system in Shanghai [31].
The DRLs, as an important measure for CT dose optimisation, are helpful for hospitals and radiologists to keep the radiation dose to patients as low as reasonably achievable facing the problem of medical radiation protection [17].The DRLs have been used for medical radiation protection in many countries in the world [12,15,28,[32][33][34].In light of the latest European guidelines on DRLs for paediatric imaging, radiology practitioners, experts and scholars have made efforts to explore the establishment of Chinese DRLs using the Chinese data [16].Recently, CT radiation dose to children patients were investigated in China in several studies [17,35].The national and local DRLs for paediatric CT still need to be set up and update in China.

Conclusion
The DRLs and ADs for paediatric CT based on age were developed in Shanghai.There was still room for radiation dose reduction of paediatric CT in Shanghai compared to the other countries.The radiation dose of paediatric CT in other medical institutions was higher than that in children-based hospitals.There is an urgent need to optimise paediatric CT dose in Shanghai.

Figure 1 .
Figure 1.The box plot of frequency distribution of age based CTDI vol in head, chest, and abdomen CT examinations.

Figure 2 .
Figure 2. The box plot of frequency distribution of age based DLP in head, chest, and abdomen CT examinations.

Table 1 .
Number and distribution of paediatric CT by age and examination region.
a Data in parentheses are the percentages of all age groups.

Table 2 .
The minimum, maximum, median, 25th and 75th percentiles of the CTDI vol values in the 4 age groups.

Table 3 .
The number, minimum, maximum, median, 25th and 75th percentile values of DLP from CT examinations in 4 paediatric age groups.

Table 4 .
Details of CT scanning protocols between the children-based hospitals and other medical institutions.

Table 5 .
Comparison of CTDI vol (mGy) between the children-based hospitals and other medical institutions.

Table 6 .
Comparison of DLP (mGy•cm) between the children-based hospitals and other medical institutions.

Table 7 .
Age-based ADs and DRLs of paediatric CT examinations in Shanghai.
a DLP/CTDI vol ratios are calculated based on DRLs.

Table 8 .
Comparisons of DRLs for paediatric CT examinations in different studies.

Table 9 .
Comparisons of ADs for paediatric CT examinations in different studies.
a The values were rounded off.