离体牙长度的锥形束CT测量研究

Measurement of length of human teeth in vitro using cone beam CT

目的:研究牙齿锥形束CT影像的垂直向放大误差及减小该误差的方法。方法:使用Galileo锥形束CT对32颗离体牙进行扫描,通过Galaxis3D软件对CBCT扫描数据进行重建,得到离体牙牙尖、根尖与扫描视野(FOV)中心所在的断层影像,并在该断层影像内测得离体牙的影像长度以及FOV中心至牙尖与根尖连线中点的距离(R值)。将R值代入锥形束CT影像放大率校正方程,获得离体牙的校正长度;使用游标卡尺测得离体牙的实际长度;使用SPSS 16.0软件对数据进行统计学分析。结果:离体牙的CBCT影像长度大于实际长度且差异显著(P<0.05),两者差值d1的均数为0.15±0.067m m,95%置信区间为0.13-0.21m m,平均误差百分比为0.74%±0.33%;离体牙的校正长度与实际长度之间差异不显著(P>0.05),差值d2的均数为0.02±0.074m m,95%置信区间为-0.01-0.05m m,平均误差百分比为0.10%±0.36%;放大率M1与M2之间的差异不显著(P>0.05)。结论:牙齿在锥形束CT的影像长度大于实际长度,平均放大率为7.4‰,应用CBCT影像放大率校正方程可使牙齿影像长度平均放大率降低到1.0‰。

Objective： To investigate the accuracy of CBCT imaging on measurement of length of human teeth in vitro.Methods： The model was established by scanning 32 teeth in vitro by Galileo cone beam CT and further analyzing the data through Galaxis3 D software. By using the sectional images including the cusp, root apex and FOV center, R-value, the distance from FOV center to the midpoint of the cusp and root was measured and analyzed by the regression equation to calculate the magnification and corrected value. A slide caliper was used to measure the distance from the tip to the root of experimental teeth as actual value. Paired t-test was conducted with SPSS16.0. Results： CBCT measurements of experimental teeth were larger than actual value and the difference was significant（P〈0.05）. The mean and standard deviation of difference was 0.15±0.067 mm. 95% confidence interval of difference was from 0.13 mm to 0.21 mm. The average percentage error was 0.74% ±0.33%. There was no significant difference between corrected value and actual value（P〉0.05）. The mean and standard deviation of the difference was 0.02±0.074 mm. 95% confidence interval of difference was from-0.01 mm to 0.05 mm. The average percentage error was 0.10% ±0.36%. There was no significant difference between magnification estimated by R-value and actual magnification（P〉0.05）. Conclusion： There was a vertical amplifier error on CBCT images of experimental teeth, the average percentage of vertical amplifier error was 7.4‰. R-value and the corrected equation may reduce the error from 7.4‰ to 1.0‰.