ISSN 1004-4140
CN 11-3017/P
鲁新亮, 陶建华, 马文涛, 等. 基于不同重建算法的容积再现成像在诊断鼻区线性骨折中的差异研究[J]. CT理论与应用研究(中英文), 2024, 33(5): 609-618. DOI: 10.15953/j.ctta.2023.212.
引用本文: 鲁新亮, 陶建华, 马文涛, 等. 基于不同重建算法的容积再现成像在诊断鼻区线性骨折中的差异研究[J]. CT理论与应用研究(中英文), 2024, 33(5): 609-618. DOI: 10.15953/j.ctta.2023.212.
LU X L, TAO J H, MA W T, et al. Differences in Volume Rendering Imaging Based on Different Algorithms in Assisting Detection of Linear Fracture of Nasal Bone Area[J]. CT Theory and Applications, 2024, 33(5): 609-618. DOI: 10.15953/j.ctta.2023.212. (in Chinese).
Citation: LU X L, TAO J H, MA W T, et al. Differences in Volume Rendering Imaging Based on Different Algorithms in Assisting Detection of Linear Fracture of Nasal Bone Area[J]. CT Theory and Applications, 2024, 33(5): 609-618. DOI: 10.15953/j.ctta.2023.212. (in Chinese).

基于不同重建算法的容积再现成像在诊断鼻区线性骨折中的差异研究

Differences in Volume Rendering Imaging Based on Different Algorithms in Assisting Detection of Linear Fracture of Nasal Bone Area

  • 摘要: 目的:探索最佳的容积再现成像(VR)重建算法,以提高对鼻区线性骨折的诊断效能。方法:回顾性纳入2022年8月至2023年8月的成人鼻骨CT影像资料,随机选取100例鼻区线性骨折和35例无骨折患者,分别行平滑算法(Smooth)、标准算法(Standard)、锐利算法(Sharp)、骨算法(Bone)的VR后处理。两名放射医师以双盲法对VR有无骨折、鼻骨孔显示及图像质量进行评分。采用相同协议对CT质控模体进行扫描,并测量不同重建算法模体图像的噪声功率谱(NPS),任务传递函数(TTF)和可检测性指数( d')。结果:医师对鼻区线性骨折的诊断效能在VR_Standard、VR_Sharp和VR_Bone之间存在差异,VR_Sharp的鼻骨孔显示评分高于VR_Standard,并且VR_Sharp的图像质量评分高于VR_Standard和VR_Bone;随着重建算法锐利程度的增加,噪声量和空间分辨力逐渐增加;Standard组、Sharp组和Bone组的NPS峰值和TTF50%分别为(225.85 HU2·mm2,0.42),(416.67 HU2·mm2,0.53)和(1 888.20 HU2·mm2,0.8)。当待测目标直径为1 mm时,Sharp组的d' 值最高。结论:VR_Sharp对鼻区线性骨折的诊断效能最佳,能更好的发挥VR在辅助诊断中的价值。

     

    Abstract: Objective: To explore the optimal reconstruction algorithm for volume rendering imaging (VR), improving the diagnostic efficacy of linear fractures of nasal bone area. Methods: Adult CT images of the nasal bone from August 2022 to August 2023 were retrospectively included, and 100 patients with linear fracture and 35 patients without fracture in the nasal region were randomly selected and underwent post-processing of VR with Smooth, Standard, Sharp, and Bone algorithms, respectively. Two radiologists scored the VR with and without fracture, the display of the nasal foramen, and the image quality in a double-blind method. The CT phantom was used for measuring the noise power spectrum (NPS), task transfer function (TTF) and detectability index (d') of the CT images of different reconstruction algorithms using the same scanning protocol. Results: The diagnostic efficacy for linear nasal fractures varied between VR_Standard, VR_Sharp, and VR_Bone, with higher scores for the display of the nasal foramen in VR_Sharp than in VR_Standard and higher image quality scores in VR_Sharp than in VR_Standard and VR_Bone. As the sharpness of the reconstruction algorithm increased, the amount of noise and spatial resolution gradually increased. The NPSpeak and TTF50% for the Standard, Sharp, and Bone groups were (225.85 HU2·mm2, 0.42), (416.67 HU2·mm2, 0.53), and (1888.20 HU2·mm2, 0.8), respectively. The Sharp group had the highest d' value when the diameter of the target to be measured was 1 mm. Conclusion: VR_Sharp has the best diagnostic efficacy for linear fractures in the nasal region, which better utilizes the value of VR in aiding diagnosis.

     

/

返回文章
返回