38 K. Mitsui et al.: Radioprotection 2025, 60( 1), 37 – 43
et al., 2010; Kawamoto et al., 2020; Sekkat et al., 2024). For radiation dose management in paediatric brain CT, age grouping has been recommended to assess the appropriateness of radiation dose( ICRP, 2017, European Commission, 2018). It has been demonstrated that the CTDIvol( Inoue et al., 2022b) and SSDE( Fujii et al., 2023) increase with age and weight in paediatric brain CT performed with AEC and that the dose indices correlate slightly better with weight than age. Furthermore, the CTDIvol( Inoue et al., 2022c) and SSDE( Fujii et al., 2023) have been shown to be highly correlated with the WED in paediatric brain CT, suggesting that AEC achieves appropriate dose modulation according to the head size. Although the WED is not necessarily readily available, body weight is a convenient indicator of the patient size and is suitable for routine radiation dose management. In adults, the radiation dose indices in CT of the body have been reported to correlate with body weight( Israel et al., 2010; Iball and Tout, 2014; Inoue et al., 2016, 2022d) and the WED( Fujii et al., 2017; Boere et al., 2018; Klosterkemper et al., 2019), and body weight is considered in radiation dose management.
Adult head sizes do not differ greatly among patients, and the patient size is not usually considered in radiation dose management for adult brain CT. Even in a report that determined the diagnostic reference level for groups divided according to the ED, the patient size was not considered for brain CT( Moghadam et al., 2021). However, four of eight facilities included in the study used AEC, which indicates that the head size was considered in clinical practice. In this study, we investigated the relationship between the patient size( head size or body size) and radiation dose indices in adult brain CT performed using AEC to assess the significance of considering the patient size for radiation dose management.
2 Materials and methods
2.1 Subjects
Patients who underwent brain CT for clinical indications according to our standard protocol were retrospectively analysed. A 64-detector-row CT scanner( Optima CT 660 Discovery Edition; GE Healthcare, Milwaukee, WI, USA) was used. Two hundred adult patients( 100 females and 100 males), aged 20 years or older, were consecutively enrolled. The exclusion criteria were no body weight record, weight larger than 100 kg, and mean WED longer than 20 cm. The upper limits of weight and WED were set to exclude outliers in terms of body size or head size. When CTscans were repeated in one patient, data of the first scan for the patient were used for analysis. Patient age, height, weight and body mass index( BMI)( means ± SDs) were 65.4 ± 16.5 years, 159.3 ± 9.7 cm, 56.4 ± 13.2 kg and 22.1 ± 4.1 kg / m 2, respectively. This study was approved by the Kitasato University Medical Ethics Organization( B22-166), and need for informed consent was waived due to its retrospective design.
2.2 Imaging procedures
The head of the patient was placed on the head holder, and the posteroanterior and lateral scout images were obtained in that order. Axial CT images parallel to the orbitomeatal line were acquired covering the lower margin of the posterior fossa and the top of the brain. Inbuilt AEC software( Auto mA and
Smart mA) was used to modulate the tube current, with a noise index of 3.3, a maximum current of 350 mA and the minimum current of 50 mA. The software adjusts the tube current so that the noise in the CT image reconstructed by filtered backprojection equals the noise index. Lateral scout images were used to estimate the attenuation strength and modulate the tube current. Additionally, organ dose modulation was applied over the orbit to reduce radiation dose to the eye lens. This function selectively reduces radiation exposure from the anterior direction and does not change posterior exposure( Gandhi et al., 2015; Dixon et al., 2016). Other imaging parameters were axial mode, a tube voltage of 120 kV, a rotation time of 1 s, a beam width of 20 mm, a slice thickness of 5 mm and a slice increment of 5 mm. Adaptive statistical iterative reconstruction( ASiR, 50 % IR blending) was used for image reconstruction.
2.3 Data analysis
The CTDIvol and DLP automatically calculated by the CT scanner were recorded. The ED and WED were determined based on CT images using the dose management system Radimetrics( version 3.4, Bayer Medical Care Inc., Indianola, PA, USA), and the values averaged over the entire imaging range were used for analysis. The ED calculated by Radimetrics is not the geometric mean of the anteroposterior and lateral diameters but the diameter of a circle with the same area as the image cross-section. The f factor was calculated as f = 1.9852 exp( �0.0486 WED), and the SSDE was obtained by multiplying the f factor by the CTDIvol( AAPM 2019). The dose indices( CTDIvol, SSDE and DLP) were plotted against the head size indices( ED and WED), followed by linear regression. The relationships of the CTDIvol and SSDE with body size indices( body weight, body height and BMI) were also assessed.
2.4 Statistical analysis
Correlations were examined using linear regression. The significance of the correlation was examined using a test of no correlation. The Hittner, May and Silver’ s modification of Dunn and Clark’ s z with common items was used to compare correlation coefficients. A parallel slopes test was performed using analysis of covariance( ANCOVA) to compare the slopes of the regression lines. In ANCOVA, parallel slopes were tested by adding an interaction term to the model for the factor by the covariate. Analysis was performed using R software( version 4.0.3, R Foundation for Statistical Computing, Vienna, Austria).
3 Results
The WED, ED, CTDIvol, SSDE and DLP were 16.6 ± 0.7 cm, 15.3 ± 0.6 cm, 38.6 ± 3.8 mGy, 34.1 ± 2.3 mGy and 582 ± 74 mGy ∙ cm, respectively. The CTDIvol and DLP were lower than the Japanese diagnostic reference levels( CTDIvol, 77 mGy; DLP, 1350 mGy ∙ cm)( Kanda et al., 2021).
The CTDIvol showed high positive linear correlations with the ED and WED( Fig. 1, p < 0.001 for both). The correlation between the CTDIvol and WED was higher than that between the CTDIvol and ED( p < 0.001).