The distribution of patients in groups during CT examination on COVID-19.
Accounting for the effective dose (ED, mSv) and calculating the radiation risk during CT is necessary to predict the long-term consequences of radiation exposure on the population. We analyzed the results of 1003 CT examinations of the chest in patients with suspected COVID-19 in the city diagnostic center. The average ED and confidence intervals (p ≤ 0.05) for patients with a single CT scan were: children (12–14 years) 2.59 ± 0.19 mSv, adolescents (15–19 years) 3.23 ± 0.17 mSv, adults (20–64 years), 3.43 ± 0.08 mSv, older persons (65 years and older) 3.28 ± 0.19 mSv. The maximum radiation risk values were 31.2*10–5 in women children and 29.3*10–5 in women adolescents, which exceeds the risk values for men in these age groups by 2.3 and 1.9 times, respectively. For the group of adult patients the risk was 11.2*10–5 in men and 17.4*10–5 in women, which is 1.6 times higher than in men. All these risk values are in the range of 10*10–5–100*10–5, which corresponds to the level LOW. For the group of older age patients, the radiation risk was 2.6*10–5, which corresponds to the level of 1*10–5–10*10–5, VERY LOW. Our materials shows in detail the technique to evaluate effective radiation doses for chest CT and calculate the radiation risk of the carcinogenic effects of this exposure.
- computed tomography
- chest CT diagnostics
- effective dose
- radiation risks levels
- the dependence of the radiation risk levels of sex and age
In the coming years, due to the introduction of methods of medical diagnostics and treatment using ionizing radiation, the growth of medical exposure of the Russian population expected to continue, especially due to computed tomography (CT). Therefore, it is important to evaluate radiation dose levels and population radiation risks in the form of a possible oncological pathology among the population in the long term after exposure [1, 2, 3, 4, 5, 6, 7, 8].
Estimating the stochastic effects on the basis of a linear non-threshold model,
Due to the widespread use of CT of the chest organs for the diagnosis of COVID-19, including during repeated examinations, this issue is of particular relevance.
The aim of the study was to assess effective radiation doses for chest CT for the diagnosis of Covid-19 and calculate the radiation risk of the effects of this exposure.
2. Material and methods
2.1 General characteristics of patients
We analyzed the results of 1003 CT examinations of the chest performed in patients with suspected COVID-19 during one week in October 2020 in the city diagnostic center. Among these patients were 6.2% children in the ages of 12–14 years old, 15.3% adolescents in the ages of 15–19 years old, 60.1% adults in the ages of 20–64 years old, and 18.4% older persons of ages 65 years and older. The average ages and confidence intervals (p ≤ 0,05) were 13.8 ± 0.20 years old in group 1 (children), 17.1 ± 0.41 years old in group 2 (adolescents); 45.8 ± 1.47 years old in group 3 (adults) (of which 41.8% are of ages 20–45 years old and 58.2% are of ages 46–64 years old); 69.4 ± 1.79 years old in group 4 (older persons). The percentage number of male (female) persons in the groups are 51.6% (48.4%) in group 1, 52.3% (47.7%) in group 2, 46.3% (53.7%) in group 3, 47% (53%) in group 4. The proportion of patients with CT signs of pneumonia and without pathological signs amounted to a total of 54,6% and 45,4%, respectively, for each of the four age groups. The distribution of the patients into groups during CT examination is given in Table 1.
|Groups||Age (ears)||Subgroups by patient sex||Number of patients||Proportion of patients with CT signs pneumonia, %||Proportion of patients without CT signs of pathology, %|
|1||Children (12–14)||1.1. Men||32||15.6||84.4|
|2||Adolescents (15–19)||2.1. Men||80||26.3||73.7|
|3||Adults (20–64)||3.1. Men||279||55.6||44.4|
|4||Older people (65 and older)||4.1. Men||87||77.0||23.0|
2.2 Description of computed tomography technique, calculation of effective dose and radiation risk
CT studies of the chest were performed on a Siemens Somatom Emotion 16 scanner (16-slice) using a standard algorithm. The voltage on the tube was 130 kV with automatic modulation of the amperage; the slice thickness was 0.8 mm (pitch 1.4) or 1.5 mm (pitch 1.2). Of each patient, the values of the parameters determining the radiation load were entered into the database CTDIvol (mGy), DLP (mGy*cm) and ED, mSv.
The calculation of the effective dose (ED, mSv) for a single phase CT scan was performed according to the following equation.
The CTDIvol values were entered into the database from the CT scanner console. Then, the DLP was calculated by the formula:
The procedure for registering the indicator “irradiated length (cm)” was as follows. Previously, the X-ray technician performed an X-ray (tomogram) of the chest. Then the region of interest (ROI) was installed on the CT scanner console in accordance with the Recommendations of EUR16262, 1999 : Volume of investigation (routine chest) - from lung apex to the base of the lungs. The length of this area (irradiated length) was measured individually in each patient. In this area, a CT scan was subsequently performed and, accordingly, the patient was irradiated. DLP was calculated for this zone.
In our study, for the chest the “irradiated length” (Median, 25th and 75th percentile) was (cm): 31.3 (30.1–33.4) - in children, 34.7 (32.6–36.6) - in adolescents, 36.6 (34.9–38.7) - in adults, 33.3 (31.6–36.8) – for persons 65 years and older.
Then, using the DLP, the effective doses was estimated according to the formula :
To calculate the effective dose (ED, mSv) the chest KED DLP conversion factor (mSv*mGy−1 *cm−1) used was KED DLP = 0.012 for both the children group (12–14 years old) and the adolescent group (15–19 years old) and KED DLP = 0.016 for the subjects older than 19 years [12, 13].
The method of calculating the risk of radiation consequences is based on the analysis of the frequency of leukemia and other oncological diseases, hereditary disorders in subsequent generations in the population after irradiation of people during the atomic explosions in Hiroshima and Nagasaki, the Marshall Islands, after gamma irradiation of patients with cancer and after incidents and accidents at nuclear reactors. Several hundred publications with this information were summarized in ICRP Publication 103, 2007 , and the risks of these consequences in persons of different genders and ages were calculated depending on the radiation dose received. In our study, calculations of radiation risk are carried out according to the National Methodological Recommendations  as follows:
R is the radiation risk per 100,000 population at an exposure dose of ED, mSv;
ED - effective dose, mSv;
r - risk indicator for exposure of 1 mSv (mSv−1).
A risk indicator for exposure of 1 mSv used, lifetime cancer risk of radiation is 5.5 * 10−5 mSv−1 for the entire population regardless of age and sex. However, in this study r (risk indicator) were used, taking into account the age and sex of patients (Table 2) in accordance with the National Methodological Recommendations . These values were calculated for the Russian population (mortality and morbidity data for 2008) using risk models and ICRP calculation methods [11, 15]. When calculating Radiation risk level, the scales listed in Table 3 were used.
|Age, years||Man||Woman||Age, years||Man||Woman|
|Radiation risk levels||Radiation risk|
|Values||Values per 100,000 people|
|NEGLIGIBLE||<10−6 (less than 1 case per 1,000,000 people)||< 0.1|
|MINIMUM||10−6–10−5 (1 to 10 cases per 1,000,000 people)||0.1–1|
|VERY LOW||10−5–10−4 (1 to 10 cases per 100,000 people)||1–10|
|LOW||10−4–10−3 (1 to 10 cases per 10,000 people)||10–100|
|MODERATE||10−3–3*10−3 (1 to 3 cases per 1,000 people)||100–300|
The mean and median values of effective doses in the formed groups were close, the assessment of the data according to Kolmogorov–Smirnov test for normality and Shapiro–Wilk’s W test showed that the nature of their distribution is close to normal. The measured data were expressed as the average ± confidence interval (X ± CI) at p ≤ 0.05, as well as median (Me, 25th and 75th percentile). The significance of differences between the groups according to Student t-criterion, P value<0.05 was considered for statistical significance. STATISTICA statistical software (version 10.0; Stat Soft. Inc., United States) was used for analysis.
3. Results and discussion
The average effective doses to patients with a single CT scan in the formed groups as illustrated in Table 4 and Figure 1A were 2.59 ± 0.19 mSv in group 1 (children 12–14 years old), 3.23 ± 0.17 mSv in group 2 (adolescents 15–19 years old), 3.43 ± 0.08 mSv in group 3 (adults 19–64 years old) and 3.28 ± 0.19 mSv in group 4 (older persons – 65 years and older).
|Groups||Age (years)||Subgroups by sex||ED, Me (25th; 75th percentile), mSv*||ED, X ± CI, mSv*||Radiation risk|
|Cases, per 100,000 people||Level|
|1||Children (12–14)||Men||2.65 (2.45; 2.96)||2.92 ± 0.30||ED*4.6||13.4||10–100||LOW|
|Women||2.17 (1.88; 2.46)||2.23 ± 0.16||ED*14.0*||31.2||10–100||LOW|
|Total||2.46 (2.13; 2.70)||2.59 ± 0.19||ED*9.3*||24.1||10–100||LOW|
|2||Adolescents (15–19)||Men||3.38 (3.01; 3.88)||3.50 ± 0.23||ED*4.4||15.4||10–100||LOW|
|Women||2.51 (2.37; 3.15)||2.93 ± 0.23||ED*10.0||29.3||10–100||LOW|
|Total||3.13 (2.37; 3.60)||3.23 ± 0.17||ED*7.2||23.3||10–100||LOW|
|3||Adults (20–64)||Men||3.69 (3.05; 4.10)||3.61 ± 0.08||ED*3.1||11.2||10–100||LOW|
|Women||3.27 (2.42; 3.77)||3.28 ± 0.13||ED*5.3||17.4||10–100||LOW|
|Total||3.39 (2.72; 3.93)||3.43 ± 0.081,2||ED*4.2||14.4||10–100||LOW|
|4||Older people (65 and older)||Men||3.15 (2.57; 3.90)||3.30 ± 0.23||ED*0.7||2.3||1–10||VERY LOW|
|Women||3.26 (2.05; 4.20)||3.26 ± 0.30||ED*0.9||2.9||1–10||VERY LOW|
|Total||3.21 (2.51; 3.90)||3.28 ± 0.193,4||ED*0.8||2.6||1–10||VERY LOW|
These doses are comparable with the ED values shown in the report  on the evaluation of DRLs for adult CT in European countries and in studies of other authors [17, 18, 19]. The DRLs for CT of adult chest organs in European countries were: 4.7–6.31 mSv in Netherlands [11, 14], 5.1–5.95 mSv in Germany [16, 18], 6.8 mSv in Austria , 5.95–10.4 mSv in Great Britain [16, 19], 7.31 mSv in Finland , 8.5–10.5 in Denmark , as well as 7.65 mSv in Australia .
In our earlier study , with standard protocols on different CT scanners, the values of the ED were in the ranges of 2.4–6.04 mSv and 8.4–15.33 mSv, for a single-phase and multiphase with contrast CTs, respectively. The use of low-dose protocols (tube voltage from 80 to 100 kV with automatic modulation of current) made it possible to reduce the ED to 1.6 mSv, when applying the iterative reconstruction algorithm MBIR for single-phase CTs and to 4.41 mSv when applying the iterative reconstruction algorithm ASIR for multiphase CTs .
Based on the risk indicator value for exposure of 1 mSv with age and sex (Table 3) were calculations radiation risk values and radiation risk levels after chest CT radiation per 100,000 exposed persons (Table 4). The maximum radiation risk values for a single CT were observed (Figure 1B) in groups of children (24.1*10−5) and adolescents (23.3*10−5). As can be seen in Figure 2B, the radiation risk values for a single CT were 31.2*10−5 in women children (12–14 years old) and 29.3*10−5 in women adolescents (15–19 years old), which exceeds the risk values for men in these groups by 2.3 and 1.9 times, respectively. For the group of adult patients the average risk was 14.4*10−5, (11.2*10−5) in men and 17.4*10−5 in women, which is 1.6 times higher than in men. Nevertheless, all these risk values are in the range of 10*10−5–100*10−5, which corresponds to the level LOW. For the group of older age patients, the radiation risk was 2.6*10−5, which corresponds to the leval rang of 1*10−5–10*10−5, VERY LOW.
We have compared the calculations with estimates of radiation risks in other studies.
For example, when planning the limits of exposure of astronauts , the risk of oncological diseases and genetic effects are rather low: 0.2*10−6 for leukemia, 0.2*10−6 for other types of malignant neoplasms, and 0.05*10−6 for genetic effects, per year per dose of additional irradiation of 1 mSv. Spontaneous incidence are 50*10−6 for leukemia, 1000–2000*10−6 for other types of malignant neoplasms and 8000*10−6 for genetic effects per year.
In publication 103 of the ICRP , new views of the ICRP on the principles and approaches to ensuring radiation safety, are formulated in comparison with the previous document - Publication 60 of the ICRP . Epidemiological data obtained since the publication of Publication 60 of the ICRP served as a reason for revising the values of the nominal risk factors per unit dose for radiogenic cancers and hereditary effects (Table 5).
|Irradiated population||Malignant neoplasms||Hereditary effects||Total|
|Publ. 103||Publ. 60||Publ. 103||Publ. 60||Publ. 103||Publ. 60|
As we can see, the new risk values in Publication 103 are slightly lower as those specified in Publication 60. But, at the same time, for children compared with adults, they were increased in terms of Malignant neoplasms from 1.5 to 1.68, for hereditary defects from 2.25 to 3.0, and in the total number of negative effects from 1.61 to 2.0. Our results are comparable to these guidelines.
I.A. Tsalafoutas, G.V. Koukourakis  emphasize that stochastic negative effects can be caused even by small doses of radiation, and give the following example of calculating the risk associated with radiation during CT. The assumption of a 5% probability of risk per 1 Sv (1,000 mSv) for the occurrence of cancer or hereditary effects means that the examination, which leads to patient exposure in ED = 10 mSv (typical for CT of the abdomen and pelvis), implies 0.05% chance of such risks. That is, for every 10,000 patients, who underwent CT with a dose of 10 mSv, five people can be expected, to develop cancer or hereditary effects as the result of radiation.
There was calculation individual of effective dose and risk of malignancy based on Monte Carlo simulations after whole body CT . The Excess Relative Risk (ERRMC), as a measurement of the exceeding risk of an exposed person compared to a non-exposed person, calculated using the solid cancer mortality in the United States as baseline (female: 17,500/100,000; male: 22,100/100,000).
There was calculation individual of Effective Dose and estimation of organ-specific additional Lifetime Attributable Risk (LAR) of cancer mortality after Whole Body Computed Tomography based on Monte Carlo simulations and report VII about Biologic Effects of Ionizing Radiation (BEIR VII). Considering the effective doses of 1.48 ± 0.15 mSv for the lungs, the LAR for mortality from lung cancer [n / 100,000] was 13.25 ± 4.24.
In our study, it was shown that with a single chest CT scan in patients with suspected COVID-19, additional (to a spontaneous level) cases of oncological pathology per 100,000 people may occur: 24.1 cases in children, 23.3 cases in adolescents, 14.4 cases in adults, 2.6 cases in older persons.
The average effective dose will increase in proportion to the increase in the number of CTs performed on the patient from 2.6–3.4 mSv with a single CT scan to the calculated values of 7.8–10.3 mSv with three times CTs. This will lead to a threefold increase in radiation risks to levels per 100,000 people may occur (Figure 1B): 72.3 cases in children, 69.8 cases in adolescents, 43.2 cases in adults, 7.9 cases in older persons. Due to the increased post-radiation risks in children; сcurrently, both the European and the American Society of Pediatric Radiology do not recommend the use of CT to diagnose COVID-19 pneumonia in children. CT is indicated only for severe, where concurrent pathology need to be excluded.
In men, the average radiation doses in the four age groups were slightly higher than in women (Figure 2A). However, with an increase in the number of CT scans from one (Figure 2B) to two (Figure 2C) and up to three (Figure 2D) in females, the increase in the calculated radiation risk compared to men is more significant, especially in women children (in 2.3 times) and among women adolescent (in 1.9 times). The radiation risk in men and women in all subgroups by age up to 65 years remains at the LOW level (10*10−5–100*10−5), and in the older subgroup at the WERY LOW level (1*10−5–10*10−5). However, with a three-fold CT scan in groups of children and adolescents, the radiation risk in women approaches the border of the MODERATE level (100*10−5–300*10−5), and in the old group to the border of the LOW level (10*10−5–100*10−5).
By evaluating the lung irradiation with the doses used in the ongoing clinical trials to treat COVID -19 patients, our data shows that a radiation dose 0.5 Gy provides an acceptable Risk Identification Checklist (RIC) estimate (LAR 1%), irrespective of sex and age at exposure . However, a promising direction is the use of modern CT scanners, which allow the use of low-dose algorithms for CT diagnostics , while significantly reducing the radiation exposure to patients.
Because the study established effective radiation doses for chest CTs of patients with the diagnosis of COVID-19, the radiation risks for a single, double and triple chest CTs in different age and sex of patients were calculated. It has been found, that the radiation risk due to a single, double and triple chest CTs for patients under 65 years old is LOW, and for 65 years old and older patents is VERY LOW. Taking into account the radiation risk during CT is necessary to reduce the long-term consequences of radiation exposure on the population.
The study was performed without external funding.
Conflict of interest
The authors declare no conflict of interest.
Conformity with the principles of ethics
The study was approved by the local ethics committee.
|ASIR||Adaptive Statistical Iterative Reconstruction|
|CTDIvol||Computed tomography dose index|
|DLP||Dose length product|
|DRLs||Diagnostic reference levels|
|ICRP||International Commission on Radiological Protection|
|LAR||Lifetime Attributable Risk|
|MBIR||Model-Based Iterative Reconstruction|