儿童胸部CT辐射剂量指标比较及扫描前剂量预算回归模型的建立

目的 比较容积CT剂量指数（CTDIVOL）与基于水当量直径（DW）的体型特异性剂量估算值（SSDEDW）在辐射剂量衡量上的差异，分析CTDIVOL、SSDEDW、剂量长度乘积（DLP）与儿童年龄（Age）、胸围（CC）的相关性,建立快速计算CTDIVOL、SSDEDW、DLP的回归模型。方法 回顾性分析2021年1月—12月我院748例儿童胸部影像学资料，根据年龄分为<5岁、5~≤10岁及10~≤15岁3个年龄组。手动测量最中间横断面图像的兴趣区面积（AROI）及CT值（CTROI），计算基于DW的SSDEDW，使用t检验比较不同年龄段CTDIVOL与SSDEDW间差异。Pearson分析CTDIVOL、SSDEDW、DLP与Age、CC间相关性，建立基于Age、CC计算CTDIVOL、SSDEDW、DLP的回归模型并对其准确性进行分析。结果 所有患儿CTDIVOL为（3.05±0.72）mGy；SSDEDW为（5.69±0.87）mGy；CTDIVOL较SSDEDW低46.40%，二者间比较差异有统计学意义（t=-241.95，P＜0.001）。基于Age的回归模型：CTDIVOL：y=1.78exp(0.06x)(x=Age，R2=0.64)；SSDEDW：y=4.16exp(0.03x)(x=Age，R2=0.51)；DLP：y=40.10exp(0.09x)(x=Age，R2=0.81)；基于CC的回归模型：CTDIVOL：y=-0.76+0.06x(x=CC，R2=0.84)；SSDEDW：y=0.06x+1.99(x=CC，R2=0.55)；DLP：y=-84.26+2.96x(x=CC，R2=0.81)。基于变量Age、CC计算CTDIVOL其误差分别为（0.13±0.10）、（0.09±0.06）；基于变量Age、CC计算SSDEDW其误差分别为（0.10±0.06）、（0.09±0.07）；基于变量Age、CC计算DLP其误差分别为（0.13±0.09）、（0.13±0.10）；基于CC计算的辐射剂量指标其误差值较Age小。结论 在儿童胸部CT辐射剂量衡量上，CTDIVOL较SSDEDW低估其辐射剂量，使用SSDEDW更为精确;使用年龄、胸围作为水当量直径的替代指征来预计算患儿的辐射剂量指标，可以在检查前快速预算其辐射，对于辐射剂量更好的控制具有重要作用

Comparisons of chest CT radiation dose indexes in children and establishment of the regression models for prescaned radiation dose calculation

Objective To compare the difference between the volume CT dose index (CTDIVOL) and the size-specific dose estimate (SSDEDW) based on the water equivalent diameter (DW) in the measurement of radiation dose, and analyze the correlations of CTDIVOL, SSDEDW and dose length production（DLP） with age and, chest circumference（CC） and establish a regression model for rapid calculation of CTDIVOL, SSDEDW and DLP. Methods The imaging data of 748 children of chest were retrospectively analyzed. According to age, they were divided into three groups: <5 years old, 5~≤10 years old and 10~≤15 years old. We manually measured the area of interest (AROI) and CT value (CTROI) on the representative central cross-sectional image, calculated the SSDEDW based on DW, and compared the differences between CTDIVOL and SSDEDW at different age groups using t-test. Pearson analysis was performed to test the correlations of CTDI VOL, SSDEDW and DLP with age and CC. Consequently, we established regression models based on age and CC to calculate CTDIVOL, SSDEDW and DLP, and analyzed the accuracy of the models. Results CTDIVOL of all children was (3.05±0.72) mGy, SSDEDW was (5.69±0.87) mGy, CTDIVOL was 46.40% lower than SSDEDW, and the difference was statistically significant (t=-241.95, P<0.001). Age-based regression models were developed as follows: CTDIVOL: y=1.78exp(0.06x) (x=Age, R2=0.64); SSDEDW: y=4.16exp(0.03x) (x=Age，R2=0.51); DLP: y=40.10exp(0.09x) (x=Age，R2=0.81). Regression models based on CC were also established: CTDIVOL: y=-0.76+0.06x (x=CC,R2=0.84); SSDEDW: y=0.06x+1.99 (x=CC，R2= 0.55); DLP: y=-84.26+2.96x (x=CC，R2=0.81). The errors of CTDIVOL calculated based on the variables including age and CC were (0.13±0.10) and (0.09±0.06), respectively, and there was no statistically significant difference between them (t=11.92, P<0.001); the errors of SSDEDW calculated based on age and CC were (0.10±0.06) and (0.09±0.07), respectively, and there was no significant difference between them (t=-17.41, P<0.001); the errors of DLP calculated based on age and CC were (0.13±0.09) and (0.13±0.10), respectively, and there was also no statistically significant difference between them (t=-7.07, P<0.001); the error of radiation dose index calculated based on CC is smaller than on age. Conclusion In measuring the radiation dose of chest CT in children, CTDIVOL could underestimated its radiation dose compared with SSDEDW indicating SSDEDW could be more accurate. Using age and CC as the substitute indications of DW to calculate the radiation dose index of children before CT scan can quickly calculate their radiation before examination, which plays an important role in better control of radiation dose