Study on the Effect of Individual Factors on the Radiation Dose in Chest CT

Objective To explore the inuence of patient’s individual factors on the radiation dose in chest computed tomography (CT) scan. Methods on the clinical chest CT scan scheme and the scanning conditions were Basic data of 103 patients who underwent chest CT scanning, including gender, age, height, weight and diseases, were prospectively collected, and the dose length product (DLP) of each patient was Multivariate regression analysis was made on the obtained data.

Mean that gradually increase in age was related with 2.19 mGy•cm increase in the DLP value, 1 kg increase in weight was associated with 5.54 mGy•cm increase in the DLP value.

Conclusion
For chest CT, age and weight are the major impact individual factors of radiation dose. This model has shown obvious clinical signi cance and can provide solid theoretical basis for clinical application in reducing the radiation dose in chest CT.

Background
With the rapid development of medical science and technology, X-ray computed tomography (CT) has been widely used in clinical practice and the low-dose chest CT has become one of the means of lung cancer screening [1][2], but its accompanying radiation problem was also closely watched [3], so how to practice the as low as reasonably achievable (ALARA) principle and reduce the radiation dose has Page 4/18 become the hot topic in the eld [4]. In this study, the in uence of individual factors on radiation dose was studied.

Research data
The data of 103 patients were collected, including gender, age, height, weight, dose length product (DLP), and the prevalence of 20 basic diseases such as interstitial changes, infection, in ammation, proliferation, ber, pneumatocele, lung nodules, pleural effusion, lung cancer, mediastinal lymph nodes, atelectasis, pulmonary bullae, emphysema, breast carcinoma, calci ed lesion, pulmonary congestion, pleural thickness, bronchiectasis, mosaic perfusion, lung destroyed and so on. There were 51 males and 52 females, aged 27-88 years, 58 ± 14.2 years on average. The baseline data of all patients are summarized in Table 1. For continuous variables, the description method is determined according to whether the data obeys normal distribution or not, if obeys the normal distribution, use the available mean ± standard deviation (SD); if not obey the normal distribution, use the median (1/4 quartile-3/4 quartile). For the classi cation variables such as basic diseases, the number of patients and the percentage of the total number of patients are used to describe the data. Image quality assessment ≥ 4 is classi ed as high-quality image, ≤ 3 is classi ed as inferior image (Fig. 1). There was no correlation between image quality and chest CT radiation dose (r=-0.083, P = 0.407 > 0.05). Furthermore, Image quality had no effect on chest CT radiation dose (Table 2). Correlation Index Of Dlp The correlation analysis of DLP with individual factors and underlying diseases are shown in Table 3.
Because DLP does not obey normal distribution, Mann-Whitney U test was used to analyze the differences and effects of basic diseases such as gender and interstitial changes on DLP. There was a signi cant difference between DLP and gender (P = 0.000013 < 0.01), and calci ed lesion had effect on DLP (P = 0.01 < 0.05).

Multiple Regression Analysis
By observing the scatter plots of individual factors (age, height, weight) and DLP of patients, it can be seen that the radiation dose has a good linear relationship with height and weight, and an approximate linear relationship with age ( Fig. 2).
According to the previous correlation analysis, there was a signi cant correlation between DLP and sex, height, weight, age and calci cation. Furthermore, multivariate stepwise regression analysis was used to investigate the in uence of ve signi cantly correlated indexes on DLP. According to the signi cant results in Table 4, Only weight and age had signi cant effects on DLP, while other variables had effects on DLP but had no statistical signi cance. Next, the tting degree of the regression model will be illustrated by the multiple correlation coe cient R value, variation degree variable R 2 and analysis of variance table. Regression model statistics show that the value of index R is 0.779 indicates the regression model has medium-high correlation, the value of adjusted R 2 is 0.599 indicates the regression model has high in uence intensity, and the Durbin-Watson test value is 2.058, which further indicates that the observed indicators are related and independent (Table 5). Variance analysis of regression model shows that the regression model has statistical signi cance, F(2,100) = 77.128, P = 5.45E-21 < 0.01, suggesting that there is a linear correlation between dependent variables and independent variables. It also further shows that the effects of the two indicators included in the model on DLP are statistically signi cant (Table 6).

Discussion
Since the clinical application of spiral CT, the radiation dose of CT is high, accounting for about 1/2 of the total medical radiation exposure [5]. Therefore, low-dose CT scanning has been the focus of clinical and scienti c research [6]. Previous studies have explored the factors affecting the radiation dose of patients from the aspects of scanning methods and reconstruction methods, but have not discussed the factors affecting the radiation dose from the factors of the patients themselves.
In the early stage, the scanning radiation dose was reduced mainly by reducing the tube current, tube voltage, reducing the scanning time and applying automatic tube current modulation technology (auto mA) [7]. Some researchers [8][9][10][11][12] believe that the image quality can meet the diagnostic requirements when the dose is 40mAs. Zhang [8][9] reduces radiation dose through tube voltage, detector width and pitch combined with organ dose modulation technology. Song [10] found that the dose of some models of spiral CT is inversely proportional to the pitch when the tube current is constant, while some models adopt the technology of automatically adjusting tube current, and the radiation dose remains unchanged when the pitch changes. Of course, the matching of detector and pitch not only affects the volume coverage and image quality, but also affects the radiation dose to a certain extent [11][12].
In terms of reconstruction method, excessive reduction of radiation dose will lead to insu cient data acquisition and increased noise in ltered back projection (FBP) reconstruction, which will result in image quality not meeting the needs of diagnosis [13]. Model-based iterative reconstruction (MBIR) is an iterative algorithm based on original data, and a new generation of model-based iterative reconstruction (the new version of MBIR, MBIRn), which has unique advantages in reducing X-ray radiation dose and improving image quality [14]. Iterative reconstruction (IR) is to improve image quality and reduce image noise and artifacts through many iterations at lower radiation dose [15]. Adaptive statistical iterative reconstruction (ASIR) is based on the system statistical model, and its clinical application value has been con rmed [16][17]. when the iterative intensity of conventional chest CT plain scan ASIR is set at 40%-60%, the image quality noise and contrast noise are higher, the lung window and mediastinal window score is the highest [18], while the image quality is guaranteed, the radiation dose is signi cantly reduced [19].
In 2015, the World Health Organization determined that the body mass index (BMI) of normal adults was 18.5-24.9 kg/m2. BMI ≥ 25.0 kg/m2 was obese, and BMI < 18.5 kg/m2 was emaciated. Qin [20] conducted a low-dose chest scan study using 256 iCT according to this grouping standard, which only revealed that this grouping was bene cial to reducing radiation dose. Other domestic researchers [21][22][23] also studied the reduction of radiation dose based on body mass, but did not explain the relationship between body mass and radiation dose. Saade et al. [24] divides people according to their different body weight into four grades, ≤ 60, 60-80, 81-100 and ≥ 101 kg, to study the effect of body weight and radiation dose. It was found that there was a certain relationship between body weight and radiation dose. As the body weight of Chinese population is generally lower than that of European and American people [25], Chen et al. [26] divides people with a body weight of 60 kg into two groups: ≤60 kg and > 60 kg. Different scanning parameters are selected respectively, and it is also found that body weight can affect the dose of CT radiation. This study shows that body weight not only affects the radiation dose, but also reveals that there is a positive correlation between individual radiation dose and body weight, the slope is 5.54, that is, each increase in body weight of 1 kg, DLP will increase 5.54 mGy•cm (regression equation ). Although there was a signi cant correlation between height and DLP (P = 0.00002 < 0.01, see Table 3), the multiple regression analysis showed that it had no signi cant effect on DLP (P = 0.917 > 0.05).
In radiological dosimetry, the difference of individual response to ionizing radiation is the key factor affecting the accurate estimation of biological dose. Studies have shown that individual differences in gene expression are mainly related to gender, age, smoking, lifestyle and in ammatory response [27][28][29].
The results of Li [30] showed that the relative expression level of mRNA of most genes in peripheral blood of female irradiated in vitro was higher than that of male, indicating that the radio sensitivity of women was higher than that of men. This study also showed that there was a very signi cant correlation between gender and DLP (P = 0.000003 < 0.01 (Table 3)), and there was a very signi cant difference between gender and DLP (P = 0.000013 < 0.01). However, in the regression analysis, the effect of gender on DLP was not statistically signi cant (P = 0.177 > 0.05). The main reason may be that there was a strong correlation between sex and height (r = 0.416, P = 0.00003 < 0.01). The expression of FDXR gene is most affected by age. When Kajimura et al. [31] conducted cluster analysis on the number of dicentric cells and the expression level of related genes mRNA in 229 Japanese primary explosive population, it was found that both of them decreased with age, suggesting that the age of the recipients may affect the changes of radiation-induced gene expression, but its molecular mechanism remains to be further studied. The results show that the number of bicentric cells and the mRNA expression level of related genes decrease with the increase of age, suggesting that the age of the recipients may affect the changes of gene expression induced by radiation, but its molecular mechanism remains to be further studied. This study shows that there is not only a signi cant correlation between age and DLP (P = 0.016 < 0.05 (Table 3)), but also a main in uencing factor in the establishment of regression equation (P = 8.08E < 0.01 (Table 4)). It also reveals that there is a positive correlation between individual radiation dose and age, the slope is 2.19, that is, gradually increase in age was related with 2.19mGy•cm increase in the DLP value (regression equation ).
In general, the radiation dose assessment parameters of CT scanning are volume CT dose index (CTD Ivol ) and DLP [32][33][34]. The effective radiation dose (ED) of the patient can be calculated by the formula: ED = DLP × k (chest conversion coe cient, k = 0.014 mSv•mGy − 1 •cm − 1 ) [35]. Their consistency is good [36]. CTD Ivol will underestimate the radiation dose received by the patient, and the smaller the BMI, the greater the underestimated dose value [37]. In 2014, the American Society of Medical physicists proposed the method of body type-speci c dose estimation (size speci c dose estimation, SSDE) [38], which uses the water equivalent diameter (Dw) to estimate SSDE. SSDE refers to the estimated CT dose accepted by the patient after body size correction, which is based on the volume CT dose index CTD Ivol displayed on the CT interface. Compared with CTD Ivol and DLP assessment methods, SSDE is relatively accurate, but there is still a certain gap between the estimated value of SSDE and the true value of radiation exposure of patients. Based on this, this paper chooses the DLP value which is easy to obtain directly instead of SSDE value as the index of radiation dose, which may be the de ciency of this study.

Conclusions
Through the analysis of the in uence of individual factors on chest CT radiation dose, the regression equation of individual factors affecting chest radiation dose is established: DLP=-39.45 + 5.54*Wight + 2.19*Age. The regression equation establishes a theoretical model for reducing the radiation dose of chest CT. According to the model, the radiation dose is reduced based on the individual factors and combined with a variety of scanning parameters, which will be the focus of the next research. It will also bring important guiding signi cance to clinical work.

Materials And Methods
General Information

Declarations
Ethics approval and consent to participate Human subjects Experimental protocols were approval by the local ethics committee of the Traditional Chinese Medicine Hospital. Written informed consent was obtained from the patient before screening.

Consent for publication
Not applicable.

Availability of data and materials
The datasets supporting the conclusions of this article are included within the article.

Competing interests
The authors declare that there is no con ict of interests regarding the publication of this paper.   Scatter plot of individual factors and radiation dose (DLP). a Scatter plot of the age and the DLP. b Scatter plot of the height and the DLP, which has a good linear relationship. c Scatter plot of the weight and the DLP, which has a good linear relationship.