64排螺旋CT卷积核的选择在横断面及容积再现重组中对鼻区骨细微结构显示的价值

The value of selective MSCT convolution kernels for the nasal bone microstructure display in cross-sectional and reorganization of VR images

目的探讨64排螺旋CT卷积核的合理选择在横断面及容积再现(VR)重组图像中对鼻区骨细微结构显示的价值。方法对60例患者鼻区扫描图像分别选择卷积核为H50s,H60s及H70s进行横断面重建后再行VR重组。结果卷积核H50s时,横断面薄层图像空间分辨率低,显示对比差;VR重组图像表面过于平滑,对鼻区骨细微结构的显示明显不足。共显示鼻颌缝112条(其中76条清晰显示,36条隐约可见),鼻额缝56条(37条可见,19条隐约可见),鼻间缝38条(20条可见,18条隐约可见),鼻骨孔29个,其他孔及筛沟不可见;卷积核H60s时,横断面薄层图像空间分辨率较高,显示对比良好,VR重组图像有一定的颗粒感,对于解剖结构的显示率高,可清晰显示鼻颌缝119条,鼻额缝59条,鼻间缝59条,鼻骨孔81个,鼻骨其他孔17个,筛沟50条;卷积核H70s时,横断面薄层图像空间分辨率更高,显示对比佳,图像清晰锐利,VR图像表面粗糙,颗粒感明显增强,部分解剖结构显示不清,共显示鼻颌缝显示117条,鼻额缝显示58条(其中38条可见,20条隐约可见),鼻间缝显示57条(其中37条可见,20条隐约可见),鼻骨孔显示79个,鼻骨其他孔显示15个,筛沟显示47条(其中30条可见,17条隐约可见)。结论卷积核的合理选择在横断面及VR重组图像中对于鼻区骨细微结构的显示具有重要的作用。

Objective To investigate the value of selective MSCT convolution kernels for the nasal bone microstructure display in cross-sectional and reorganization of VR images. Methods Sixty cases of patients with nasal area scan images were selected for the convolution kernel H50s, H60s and HT0s cross-sectional reconstruction respectively, and then reline VR restructuring. Results The convolution kernel was H50s, the spatial resolution of transverse thin section images was low, and the display contrast was poor; the surface of the VR reconstructed image was too smooth, and the nasal bone microstructures were obviously insufficient. Total nasal maxillary suture 112 were displayed （of which 76 showed clearly, 36 vaguely visible ）, 56 nasal frontal suture （37 visible, 19 vaguely visible）, 38 nasal suture （20 visible, 18 vaguely visible）, 29 nasal bone hole, the other nasal holes and ethmoidal sulcuses were invisible. Convolution kernel was H60s, the spatial resolution of transverse thin section images was high, and the display contrast was good; VR reorganization images had a certain grainy, display rate of anatomical structure was high . The nasal maxillary suture 119 were displayed, 59 nasal frontal suture, 59 nasal suture, 81 nasal hole, the other 17 nasal hole, 50 ethmoidal sulcus. Convolution kernel was H70s, the spatial resolution of transverse thin section images was higher, and the display contrast was better, the images were clear and sharp, the surface of the VR images was rough, the graininess was obviously enhanced, some anatomical structure was unclear. The nasal maxillary suture 117 were displayed, 58 nasal frontal suture （38 visible, 20 vaguely visible）, 57 nasal suture （37 visible, 20 vaguely visible）, 79 nasal hole, 15 the other nasal hole, 47 ethmoidal sulcus （30 visible, 17 vaguely visible）. Conclusion The rational selection of the convolution kernel has an important value for the display of the nasal region in transverse section and the VR image.