Is it necessary to define new diagnostic reference levels during pandemics like the Covid19-?

Abstract

Introduction

This study intended to assess the dose length product (DLP), effective cumulative radiation dose (E.D.), and additional cancer risk (ACR) due to a chest CT scan to detect or follow up the Covid-19 disease in four university-affiliated hospitals that used different imaging protocols. Indeed, this study aimed to examine the differences in decision-making between different imaging centers in choosing chest CT imaging protocols during the pandemic, and to assess whether a new diagnostic reference level (DRL) is needed in pandemic situations.

Methods

This retrospective study assessed the E.D. of all chest imagings for Covid-19 for six months in four different hospitals in our country. Imaging parameters and DLP (mGy.cm) were recorded. The E.D.s and ACRs from chest CT scans were calculated using an online calculator.

Results

Thousand-six hundred patients were included in the study. The mean cumulative dose due to chest CT was 3.97 mSv which might cause 2.59 × 10⁻² ACR. The mean cumulative E.D. in different hospitals was in the range of 1.96–9.51 mSv.

Conclusions

The variety of mean E.D.s shows that different hospitals used different imaging protocols. Since there is no defined DRL in the pandemic, some centers use routine protocols, and others try to reduce the dose but insufficiently.

In pandemics such as Covid-19, when CT scan is used for screening or follow-up, DLPs can be significantly lower than in normal situations. Therefore, international regularized organizations such as the international atomic energy agency (IAEA) or the international commission on radiological protection (IRCP) should provide new DRL ranges.

1. Introduction

Computed tomography (CT) is a valuable tool in diagnosing lung diseases, especially those invisible in plain chest radiography (Hanna et al., 2014). During the Covid-19 pandemic, CT scan had a significant role in detecting this disease to the extent that it was suggested as a screening tool (Ai et al., 2020; Xie et al., 2020). Studies have shown that CT scan has a higher sensitivity but lower specificity than the reverse transcriptase-polymerase chain reaction (RT-PCR) test, which is the gold standard of Covid-19 detection. However, the superiority of CT scan compared to RT-PCR is that its results can be prepared immediately; therefore, it is used widely for diagnosis and triage of Covid-19 (Kim et al., 2020; Mair et al., 2021). In addition to the Covid-19 initial detection, some studies suggest applying this modality for follow-up, assessing the disease progress, and even predicting the patient's prognosis (Pan et al., 2020b; Yan et al., 2020). This led to performing repetitive CT scans.

Despite the proven effectiveness of CT scan imaging in the early phase and follow-up of Covid-19 patients, this issue is very controversial. Some studies suggest not performing repeated CT scans to follow up on the patients, especially those with improving symptoms (Chen et al., 2020; Gao et al., 2021; Zhou et al., 2020). American College of Radiology also recommends against applying CT scans for initial Covid-19 diagnosis (American college of radiology, 2020). The primary basis of these objections is mostly for preventing unnecessary radiation to patients.

According to report 74 of the international commission on radiation units and measurements (ICRU) (Zoetelief et al., 2005), dose limits do not apply to medical exposure; however, Diagnostic reference levels (DRL) can be used to identify the medical radiation facilities that deliver a high amount of radiation to patients and required optimization. The international commission on radiological protection (ICRP) (Vañó et al., 2017) suggests that the quantity which wants to be used as DRL should be easily measurable, and DRL should define by the 75th percentile of that quantity. In CT scan, dose length product (DLP) is a proper quantity to be used as DRL; therefore, in CT scan, DRL is 75th percentile of DRL. On this basis, the European Commission (Comission, 2014) defined a DRL value of 400 mGy cm for chest CT scans for DLP in the range of 270–700 mGy cm.

Considering radiation hazards, during Covid-19, some centers change their routine imaging protocols to reduce the radiation dose of patients who might need repetitive imaging. For instance, Tabatabaei et al. reported that they reduced the mean effective dose of the patients who underwent chest CT scan from 6.60 ± 1.47 to 1.80 ± 0.42 mSv by reducing mAs from 150 to 30. The other imaging parameters were the same as their standard protocol. They declared image quality was a little sacrificed; however, the diagnostic accuracy was the same (Tabatabaei et al., 2020).

Meanwhile, some studies assessed the effectiveness of low-dose and ultra-low-dose CT scans in detecting and evaluating Covid-19, and most of them confirmed that low or ultra-low-dose CT can reduce a patient's radiation dose without reducing its efficiency (Agostini et al., 2020; Azadbakht et al., 2021; Fan and Liu, 2020; Greffier et al., 2021; Jalli et al., 2021). For instance, Samir et al. declared that at the beginning of Covid-19 they changed their CT scan protocol, according to the American Association of Physicists in Medicine (AAPM) suggestion for lung cancer screening. Even in the second wave of Covid-19 pandemic, they tried to reduce the patient's radiation dose even more by applying ultra-low dose CT scan. They reported that they could reduce the mean effective dose from 0.85 (low-dose protocol) to 0.59 mSv/mGy cm (ultra-low-dose protocol) by reducing mAs from 45 to 22. They also stated that ultra-low-dose CT did not affect the diagnostic efficiency (Samir et al., 2021b).

This shows that decision-making for reducing patient's radiation dose is critical and can significantly affect the cumulative dose of patients and reduce the stochastic effect of radiation, including radiation-induced cancer or genetic effects on future generations (Chang et al., 2021).

Despite the abundant knowledge about radiation-induced hazards and the introduction of low-dose and ultra-low-dose CT scans, there is no defined standard protocol or dedicated DRL value for CT imaging during the pandemic to be followed by all centers. Contrary to the centers that decide to reduce the radiation dose of their patients, some centers use the same routine protocols, although they know the Covid-19 patient might need more chest CT images for follow-up or subsequent infection with Covid-19.

This study intended to assess the DLP, effective cumulative radiation dose, and additional cancer risk due to a chest CT scan to detect or follow up the Covid-19 disease in four university-affiliated hospitals that used different imaging protocols. Indeed, this study aimed to examine the differences in decision-making between different imaging centers in choosing chest CT imaging protocols during the pandemic, and to assess whether new DRL is needed in pandemic situations.

2. Materials and methods

This retrospective study assessed the radiation dose of all chest CT scans requested to detect or follow up on Covid-19 patients. Patient's data were gathered from February 20th to August 21st, 2020. It should be mentioned that February 19th was the date of the first official announcement of the Covid-19 appearance in our country (Head of the Public Relations of the Ministry of Health, 2020); therefore, we considered the mentioned date to ensure that there was no imaging related to Covid-19 before that date. The inclusion criteria were, having chest CT imaging related to Covid-19. The patients who had not recorded dose reports, imaging parameters, or explicit imaging requests that showed the relationship of imaging to Covid-19 were excluded from the study.

All the chest imaging data related to Covid-19 were gathered for each patient in addition to patients' demographic information, imaging parameters, and DLP (mGy.cm) from the picture archiving and communication system (PACS, INFINITT). It is also worth mentioning that the accuracy of DLP and CTdose index of all scanners had been verified as a part of regular quality control. The radiation effective doses and risks from chest CT scans were calculated using an online X-ray radiation risk calculator (Hanley et al., 2018) available at: https://www.xrayrisk.com/calculator/calculator.php by inserting the patient's sex, age, and DLP of CT scans.

MedCalc version 19.5.1 was used for the statistical analysis of data. The Kolmogorov–Smirnov test was used to test the data normality. Kruskal- Wallis test was applied to compare the data of different hospitals. The p-value of less than 0.05 was considered a significant level.

3. Results

2128 patients were assessed from four University-affiliated hospitals. 1600 patients were included in the study, composed of 908 males (mean age of 49.06 ± 16.48 years) and 692 females (mean age of 50 ± 17.35 years) with a total mean age of 49.47 ± 16.68 years old. 528 patients were excluded from the study for not having recorded dose reports (43 cases) or explicit imaging requests (485 patients). CT scanner information and imaging parameters used in each hospital are presented in Table 1 . The hospitals were named H1 to H4.

Table 1.

Number of patients and CT scans in each hospital. CT scanner information and imaging parameters containing kiloVoltage (kV), milliampere-seconds (mAs), slice thickness (not reconstructed), and spacing used in each hospital before and during covid-19. The mAs in H2 and H4 differed for each patient, so they were presented as mean mAs±SD. NDR= Number of detector rows, RC = Recuostruction, IR= Iterative reconstruction, filtered back projection (FBP), # = Number of.

Hospital	#patients	#CT	CT scanner	Scan type	NDR	RC	Before Covid-19				During Covid-19
							Slice thickness (mm)	Pitch	kV	mAs	Slice thickness (mm)	Pitch	kV	mAs
H1	100	162	GE, Brightspeed	Helical	4	FBP	2.5	1.5	120	30	2.5	1.5	120	17
H2	300	535	Philips, Ingenuity CT	Spiral	128	IR	2	1.2	120	160	2	1.2	120	157.29 ± 41.85
H3	200	287	GE, Brightspeed	Helical	4	FBP	2	1.5	120	30	2.5	1.5	100	17
H4	1000	1479	Siemens, syngo VC40A	Spiral	128	IR	2	1.5	120	160	5	1.5	110	64.87 ± 12.87

Overall, on average, 2463 CT scans were performed for these 1600 patients, with almost 1.54 CT scans for each patient. The number of CT scans performed in each hospital is presented in Table 1. The Number of CT scans and the additional risks resulting from radiation dose are presented in Table 2 . The CT scan number for each patient was also evaluated, and the percentage of patients with more than two or more than four CT scan imaging is presented in Table 3 .

Table 2.

The Number of CT scans, effective radiation dose, and the additional risks of radiation doses. The additional cumulative risk is 5.89 × 10⁻¹.

Parameter	# CT	effective dose (mSv)	risk
Mean	1.540	3.969	2.59 × 10−2
SD	0.930	4.470	3.59 × 10−2
Minimum	1.000	0.155	5.09 × 10−4
Median	1.000	2.376	1.40 × 10−2
Maximum	8.000	47.142	5.13 × 10−1

Table 3.

Percentage of patients who had more than 2 and 4 CT imaging totally and in each hospital.

Hospital	>2 CT (%)	>4 CT (%)
H1	20.3	2.3
H2	11.5	2
H3	12	2
H4	9.8	1.4
Total	12.1	1.7

According to Kolmogorov–Smirnov test, the radiation dose data from different hospitals were not normally distributed (D > 0.25 and p-value<0.0001). Therefore, the difference between radiation doses was calculated using Kruskal–Wallis test. The CT scan DLP and cumulative effective radiation dose results and its descriptive statistics from each hospital are presented in Table 4 . According to Kruskal–Wallis and post-hoc test, there was a significant difference between DLP and effective doses of all hospitals (p-value<0.0001).

Table 4.

Descriptive statistics of CT scan dose length products and cumulative effective radiation dose (mSv) in each hospital. SD = standard deviation.

Factor	# patients	Dose length products (mGy.cm)							Cumulative effective radiation dose (mSv)
		Minimum	25th percentile	Median	75th percentile	Maximum	mean	SD	Minimum	25th percentile	Median	75th percentile	Maximum	mean	SD
H1	100	113.88	201.07	216.09	229.47	320.08	217.07	26.29	2.050	3.845	6.26	8.55	30.70	6.99	4.13
H2	300	152.10	287.50	344.80	437.80	873.00	379.71	237.21	2.833	5.432	7.21	10.28	47.14	9.51	7.04
H3	200	85.74	125.94	140.01	147.05	160.12	134.29	18.10	1.543	2.502	2.70	5.11	12.51	3.88	2.13
H4	1000	20.00	59.00	64.00	79.25	224.00	93.71	402.13	0.155	1.026	1.17	2.23	14.49	1.96	1.91

4. Discussion

The Covid-19 pandemic showed that medical imaging using ionizing radiation could be very helpful in disease detection and patient follow-up. Although many low-dose protocols have been introduced (Brain et al., 2017; Cheng et al., 2020; Konda et al., 2018; Pinsky, 2018) to reduce the radiation dose and, therefore, the risk of radiation-induced hazards, many imaging centers apply their routine imaging protocols, since there are no rules for applying lower dose protocols. This study assessed chest imaging of four important University-affiliated hospitals by evaluating CT imaging parameters, DLP, radiation doses and calculated additional risks.

The results of Table 1 showed how hospitals altered their routine protocols during Covid-19. According to Table 1, H2 used its routine protocols of CT imaging (kV = 120, mAs = 157.29 ± 41.85, slice thickness = 2 mm) without any consideration to reduce the dose; however, H1 (kV = 120, mAs = 17) reduced the radiation dose by decreasing mAs from 30 to 17; H3 (kV = 100, mAs = 17) tried to decrease the radiation dose using lower mAs (30 Vs 15) and a slight increase in slice thickness (2.5 Vs 2 mm); however, H4 (kv = 110, mAs = 64.87 ± 12.87) applied a lower dose CT scan for Covid-19 patients by increasing slice thickness (5 Vs 2mm), reducing kV (from 120 to 110) and mAs (from ∼160 to ∼65).

An optimization study by Niu et al. suggests kV = 100 and mAs = 40 for Covid-19 patient's chest CT scans (Niu et al., 2021). In this study, different hospitals used different kV and mAs. H1, H3, and H4 used low mAs. H3, and H4 used lower kVs of 100 and 110, respectively. A review article by Azadbakht et al. (2021) declared that many studies used kV of 120 and mAs between 30 and 250, which shows that not just in evaluated centers in this paper but almost in many centers around the world, Covid-19 CT imaging is performed with almost routine kV and mAs.

According to Table 2, on average, each patient receives a 3.97 mSv radiation dose that increases the excess cancer risk by almost 2.59 × 10⁻², i.e., this radiation dose to 100 people might cause radiation-induced cancer in 2.59 individuals. Although it seems trivial, in a large population, it will be significant. According to woldometer (available at https://www.worldometers.info/) till September 25th, 2021, 231920335 Covid-19 cases had recorded, 4751679 of which passed away, and 208530431 cases recovered. Considering excess cancer risk of 2.59 × 10⁻², there is a possibility of developing radiation-induced cancer in 5400938 cases. If we could deliver the minimum mean radiation dose (0.18 mSv) to the whole population, the excess cancer risk could be decreased to 5.09 × 10⁻⁴.

It should be added here that some researchers do not believe in additional radiation risk calculation for medical radiation. Since radiation risk calculations are mainly based on the Biological Effects of Ionizing Radiation (BEIR) VII report, and this report was the result of a study on the survivors of the Japanese atomic bombing that had different conditions compared to people who refer for medical radiation, so the risk estimation based on this may not be completely accurate (Hendee and O’Connor, 2012). However, it should be considered that (BEIR) VII report proved radiation exposure could cause some hazards that can be prevented by reducing the amount of radiation. Therefore, it is logical that we should not give up any efforts to reduce patients' radiation doses.

The number of CT scans was in the range of 1–8 in this study. However, Table 3 showed that only 12.1% and 1.7% of 1600 patients had more than two and four CT imaging, respectively. Therefore, 87.9% of patients had less than or equal to two CT scans, which is acceptable. It should be noted that repetitive CT imaging was used to assess patient's condition during treatment or follow-up in this study, and there was no repeated imaging for technical errors or patient movement amongst the included population. Other studies also reported more than one CT scan for Covid-19 patients. For instance, Zhou et al. reported an average of 4 ± 2 CT scans for each patient with the median cumulative effective dose of 17.34 mSv between the range of 2.05–53.39 mSv (Zhou et al., 2021). In another study by Cristofaro et al., the median total cumulative dose was 9.04 mSv (Cristofaro et al., 2021). Other studies also reported repetitive CT imaging for Covid-19 patients between three to eight in a short period (Bernheim et al., 2020; Huang et al., 2020; Kang et al., 2020; Pan et al., 2020a). The average number of CT scans in this study was 1.5 (Table 2), with the median cumulative effective dose of 2.376 mSv, in the range of 0.155–47.142 mSv, which is significantly lower than the mentioned studies.

The mean cumulative effective CT radiation dose (3.97 mSv) was less than the mean dose of a standard CT scan (7 mSv) (Larke et al., 2011; Mettler et al., 2008), which means that in a general view, our hospitals were slightly successful in reducing the CT radiation doses. However, by analyzing the median radiation doses separately for each hospital (Table 4), it revealed that H2 delivered almost standard CT radiation dose to the patients, H1 was slightly successful in dose reduction, H3 and H4 could reduce the radiation dose significantly by a factor of 2.6 and 6 respectively. Considering the 75th percentile, just H3 and H4 could reduce the overall CT scan doses to less than a standard chest CT scan for most patients.

Widening the slice thickness and slice spacing in conjunction with a mild kV and mAs reduction was an effective way to reduce the chest CT radiation dose to evaluate lung engagement to Covid-19. Sergiacomi et al. also declared that applying a multi-slice CT scan with a 5 mm slice thickness provided a diagnosable image with a better dose reduction than a high-resolution CT scan. They also reported that there were not any significant differences between the detection of ground-glass opacities in either method (Sergiacomi et al., 2010). This result was consistent with our results. The images of H4, which used a 5 mm slice thickness with pitch = 1.5, were diagnosable. The data that might miss in this protocol is negligible compared to the benefits of a considerable reduction in radiation dose. Here, it should be pointed out that the reports of all images included in this study have revealed that the images of all three hospitals that altered their imaging protocol during Covid-19 were suitable for diagnosis, and a slight decrease in image quality had not disturbed the diagnosis, and none of them needed to be replaced.

Regarding the pitch number, some studies suggest a pitch of less than 1 to have an image with enough quality to detect the ground-glass nodules (Niu et al., 2021). However, other studies used pitch numbers of 1.4 and 1.5 and slice thickness of 5 mm, satisfying the results (Samir et al., 2021a; Song et al., 2020), similar to the results of H4 in this study.

A brief review of the results revealed that H1 and H3 had CT equipment with the same features. During Covid-19, H3 decreased the kV from 120 to 100, reducing the mean cumulative effective dose from 6.99 mSv to 3.88 mSv (44.49% dose reduction). Another study by Van Der Wall et al. assessed the effect of reducing kV from 120 to 100. They also reduced the effective radiation dose from 15.4 mSv to 8.5 mSv (44.8% dose reduction). This study confirms the results of Van Der Wall et al. (Van Der Wall et al., 2011).

Many studies have shown that the reconstruction method is essential in radiation dose reduction. Since by using iterative reconstruction, good-quality images can be provided via lower kV and mAs (Klink et al., 2014; Solomon et al., 2017). Other studies also show the effect of imaging protocols and reconstruction on radiation dose (Dimitroukas et al., 2022; Metaxas et al., 2022). In this study, H2 and H4 were equipped with CT scanners with almost the same features that were eqqipped with iterative reconstruction. H4 could reduce the mean effective cumulative dose to 1.96 mSv; however, H2 did not use this facility to reduce patients' dose, and the mean effective cumulative dose in this center was 9.5 mSv.

According to the radiation protection principle, the applied radiation dose should be as low as reasonably achievable (ALARA), meaning the radiation application should be justified and optimized. Although the application of CT scan to detect and follow-up Covid-19 is justified, the radiation dose delivered to patients in some imaging centers is not optimized, as some centers do not reduce the radiation dose to the least possible for them, such as H2 in this study. H2 was equipped with a 128-slice spiral scanner with iterative reconstruction; however, the radiation dose they delivered to patients was even more than a 4 slice helical scanner with filtered back projection reconstruction while they could at least reduce the dose to 2 mSv.

This might be the result of the fact that the dose limits do not apply to patient dosimetry or medical use. Therefore, there is no obligation to reduce the patient's radiation doses to the least possible in a pandemic.

In the introduction section, it was mentioned that European Commission (Comission, 2014) defined a DRL of 400 mGy cm (270–700 mGy cm). In a review article, Paulo et al. (2020) reported the DRL for chest CT scans by DLPs between 170 and 1400 mGy cm (Paulo et al., 2020). Considering Table 4 of this study, 75th percentile of DRLs (79.25–437.80 mGy cm) shows that the values are in the range of these DRLs. However, in pandemics such as Covid-19, the CT scan is used for screening or follow-up, and it is recommended to use a low or ultra-low-dose, single-phase, fast-scan protocol (Zhou et al., 2021); that the DLPs can be significantly lower. In this study, the least 75th percentile of DLPs for a multidetector spiral CT scanner equipped with IR was 79.25 mGy cm which is much smaller than DRL suggested by Paulo et al. or European Commission, and the least 75th percentile of DLPs for a 4 slice, helical, CT scanner and filtered back projection reconstruction was 147.05 mGy cm which is again smaller than the lower boundary of DRL that were suggested in the mentioned studies.

When there is no defined DRL in a pandemic situation, some centers use routine protocols such as H2, and others try to reduce the dose; however, they might not reduce it to the lowest possible for their scanner type. For instance, the Philips and Siemens scanners (i.e., H2 and H4, respectively) were spiral and equipped with iterative reconstruction; however, their mean dose was significantly different (9.51 vs. 1.96 mSv). G.E. bright speed scanners (i.e., H1 and H3) were helical and not equipped with an iterative reconstruction method; however, their mean dose is also significantly different (6.99 vs. 3.88 mSv). Regarding the facilities of each scanner type, if H2 and H1 had used the same protocol as H4 and H3, respectively, they would be more successful in reducing the patient's radiation dose.

Our country has 31 provinces; if we could gather the DLP of chest CT for Covid-19 patients, we would show the great diversity in imaging protocols in different cities in a single country; in this way, the need to define boundaries to limit the radiation dose of patients in medical studies became more prominent. Although it was a multicenter study, the information was gathered from 4 hospitals in two provinces. This can be considered as a limitation of the present study. Even though patient size affects the DLP, which is used to estimate effective dose. However, Hu et al. (2021) declare that Size-specific dose estimatation of adult in chest CT scans is important. In this study, patient size was not considered directly in effective dose calculation, which can be another limitation of the study.

The ideal is that in the pandemic situation, when especial imaging such as lung CT scan is so popular, the international regularized organizations such as the international atomic energy agency (IAEA) or ICRP provide a DRL range, and the atomic energy agency of each country supervise the DRLs.

5. Conclusion

This study showed that in a pandemic situation that sequential CT imaging might apply to patients to reduce the patient's dose to the lowest reasonable achievable; some centers altered their CT imaging parameters. Although this protocol alteration reduced image quality a little, the chest CT imaging diagnostic value was still the same. In this study, each of the patient-friendly centers had changed their protocols according to their facilities to reduce the cumulative dose of patients compared to normal conditions, but there was no pre-determined standard limit for this purpose, so the vacuum of the existence of DRL is very felt. It is suggested that international regularized organizations provide DRLs for CT scans in pandemic situations or any other situation that sequential CT imaging might need.

Ethical Approval code

This research was approved by the Faculty Research Ethics Committee of Shiraz University of Medical Sciences with the Ethical Approval code of I.R.SUMS.REC.1399.835.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Dr. Chris Chantler

Data availability

No data was used for the research described in the article.