增加噪声模拟低剂量胸部CT扫描的研究

目的：评估为原始图像添加噪声的模拟低剂量胸部CT的可行性、准确度及其适用范围。方法：为标准剂量下胸部常规CT、薄层高分辨CT各10例和肺结节体模的原始图像添加噪声，模拟常规CT60mA、30mA和薄层高分辨CT160mA、80mA及270mA。10mA（间隔10mA）的低剂量图像。测量主动脉弓水平和肝顶水平主动脉的噪音指数（NI），统计学方法采用配对t检验、Pearson线性相关分析。结果：体模实际扫描图像与模拟图像NI的差异有显著性（P〈0．05），呈线性相关；常规CT的NI于主动脉弓层面无差异（P〉0．05）；于肝顶层面有差异（P〈0．05），呈线性相关；薄层高分辨CT在两层面均存在差异（P〈0．05），亦存在线性相关。结论：添加噪声取得的模拟图像与实际图像噪声相当，可应用于常规胸部CT平扫的低剂量研究。

Application and research of increasing noise on simulation of the chest scan with low-dose CT

Objective： To assess the feasibility, fidelity and applicability simulation of low dose CT scan of chest. Methods： Three sets of raw conventional chest CT scan of ten patients, thin-section high resolution of adding noise on raw data for data were studied, including CT scan of ten patients, and conventional CT scan of lung nodule mannequin. Noise was added to simulate conventional CT scan with tube current of 60 mA and 30 mA, thin-section high resolution CT scan with tube current of 160 mA and 80 mA. For mannequin data, a series of noise were added to simulate tube current from 270 mA to 10mA with interval of 10 mA. Noise index （NI） on levels of aortic arch and liver was measured and compared by paired t test and Pearson correlation analysis. Results： NI was different between scan data and simulated data of mannequin （P〈0.05）, which had linear correlation; For conventional CT scan, NI was different on liver level （P〈0.05）, while not different on aortic arch level, there had linear correlation; For high resolution CT scan, NI was different on both levels （P〈0.05）,which had linear correlation too. Conclusion： For conventional CT scan, the noise between scan data and simulated data is comparable, and adding noise can be used for low dose CT study.