Yang Ching-Ching
Department of Medical Imaging and Radiological Sciences, Tzu-Chi University of Science and Technology, Hualien, Taiwan.
PLoS One. 2016 Mar 7;11(3):e0149904. doi: 10.1371/journal.pone.0149904. eCollection 2016.
Scatter is a very important artifact causing factor in dental cone-beam CT (CBCT), which has a major influence on the detectability of details within images. This work aimed to improve the image quality of dental CBCT through scatter correction.
Scatter was estimated in the projection domain from the low frequency component of the difference between the raw CBCT projection and the projection obtained by extrapolating the model fitted to the raw projections acquired with 2 different sizes of axial field-of-view (FOV). The function for curve fitting was optimized by using Monte Carlo simulation. To validate the proposed method, an anthropomorphic phantom and a water-filled cylindrical phantom with rod inserts simulating different tissue materials were scanned using 120 kVp, 5 mA and 9-second scanning time covering an axial FOV of 4 cm and 13 cm. The detectability of the CT image was evaluated by calculating the contrast-to-noise ratio (CNR).
Beam hardening and cupping artifacts were observed in CBCT images without scatter correction, especially in those acquired with 13 cm FOV. These artifacts were reduced in CBCT images corrected by the proposed method, demonstrating its efficacy on scatter correction. After scatter correction, the image quality of CBCT was improved in terms of target detectability which was quantified as the CNR for rod inserts in the cylindrical phantom.
Hopefully the calculations performed in this work can provide a route to reach a high level of diagnostic image quality for CBCT imaging used in oral and maxillofacial structures whilst ensuring patient dose as low as reasonably achievable, which may ultimately make CBCT scan a reliable and safe tool in clinical practice.
散射是牙科锥形束CT（CBCT）中一个非常重要的伪影成因，对图像中细节的可检测性有重大影响。这项工作旨在通过散射校正提高牙科CBCT的图像质量。
从原始CBCT投影与通过外推拟合到使用两种不同轴向视野（FOV）大小获取的原始投影的模型所获得的投影之间的差异的低频分量，在投影域中估计散射。通过蒙特卡罗模拟优化曲线拟合函数。为了验证所提出的方法，使用120 kVp、5 mA和9秒的扫描时间，对覆盖4 cm和13 cm轴向FOV的拟人化体模和带有模拟不同组织材料的棒状插入物的充水圆柱形体模进行扫描。通过计算对比度噪声比（CNR）评估CT图像的可检测性。
在未进行散射校正的CBCT图像中观察到束硬化和杯状伪影，特别是在使用13 cm FOV获取的图像中。通过所提出的方法校正后的CBCT图像中这些伪影减少，证明了其在散射校正方面的有效性。散射校正后，CBCT的图像质量在目标可检测性方面得到改善，目标可检测性量化为圆柱形体模中棒状插入物的CNR。
希望这项工作中进行的计算能够为在口腔和颌面结构中使用的CBCT成像达到高水平的诊断图像质量提供一条途径，同时确保患者剂量尽可能低，这最终可能使CBCT扫描成为临床实践中可靠且安全的工具。
