中华放射学杂志2024,Vol.58Issue(2) :172-179.DOI:10.3760/cma.j.cn112149-20231130-00433

光谱CT联合金属去伪影技术改善胸腹部 125I粒子植入患者CT图像质量的价值

The value of spectral CT combined with metal artifact reduction algorithms in improving the CT image quality for patients with 125I seeds implantation in the chest and abdomen

周宇涵 雷丽敏 王芝浩 黄文鹏 曹伟萌 董书杉 王猛 周志刚 张琳琳 王春杰
中华放射学杂志2024,Vol.58Issue(2) :172-179.DOI:10.3760/cma.j.cn112149-20231130-00433

光谱CT联合金属去伪影技术改善胸腹部 125I粒子植入患者CT图像质量的价值

The value of spectral CT combined with metal artifact reduction algorithms in improving the CT image quality for patients with 125I seeds implantation in the chest and abdomen

周宇涵 1雷丽敏 1王芝浩 1黄文鹏 2曹伟萌 1董书杉 3王猛 1周志刚 1张琳琳 王春杰
扫码查看

作者信息

  • 1. 郑州大学第一附属医院放射科,郑州 450052
  • 2. 北京大学第一医院核医学科,北京 100034
  • 3. 飞利浦临床科研部,北京 100000
  • 折叠

摘要

目的 探讨新型双层探测器光谱CT虚拟单能量图像(VMI)联合金属去伪影算法(O-MAR)在125I粒子植入术后所致不同类型伪影的去除效能及术后疗效评估中的图像质量改善价值。 方法 该研究为横断面研究。回顾性收集2022年3月至9月于郑州大学第一附属医院125I粒子植入术后行双层探测器光谱CT胸腹部联合增强扫描的患者35例。采集光谱数据分别重建为混合能量图像(CI)、VMI(50~150 keV)图像、CI+O-MAR图像、VMI(50~150 keV)+O-MAR图像,评估各组去伪影效果及图像质量改善情况。每位患者选取伪影最强的2个层面分析,共选取70层。客观指标包括伪影指数(AI)、受伪影影响软组织区域信噪比(SNR)和对比噪声比(CNR)。主观指标包括:过度矫正伪影和新伪影产生的评估、各组图像产生不同形态伪影的数目评估、受伪影影响组织的诊断及图像质量评估。多组数据间比较采用单因素方差分析。联合O-MAR组与非O-MAR组之间采用配对样本t检验比较2组间定量指标。观察者之间的一致性评估采用Kappa检验。 结果 在高/低密度伪影(ROIH/L)中,随VMI keV升高各组的AI值均呈下降趋势;在受伪影影响组织(ROIT)中,CI/VMI(70~150 keV)+O-MAR各组的SNR均优于CI/VMI组(P<0.05),CI/VMI(50~150 keV)+O-MAR各组的CNR均优于CI/VMI组(P<0.05)。伪影过度矫正及新伪影均集中分布于VMI 50 keV、VMI 70 keV组,相较于VMI(50~70 keV)组,VMI(50~70 keV)+O-MAR组中过度矫正及新伪影数目显著降低(P<0.05);在伪影形态的比较中,随VMI keV升高,高密度伪影中a型条带伪影数目逐渐减少,在VMI 150 keV+O-MAR组高密度伪影中最低为3;e型极少或无伪影数目增多,在VMI 150 keV+O-MAR组高密度伪影中最高为23,且各形态高密度伪影总数随VMI keV升高呈减少趋势。随VMI keV升高VMI+O-MAR各组高密度伪影诊断及图像质量评分均高于低密度伪影(P<0.05)。 结论 VMI联合O-MAR可显著提高125I粒子植入术后CT随访的客观及主观图像质量,病灶显示能力及诊断信心。此外,VMI+O-MAR对高密度伪影的矫正效果更为显著。 Objective To investigate the value of the virtual monoenergetic image (VMI) obtained by a new dual-layer detector spectral CT combined with metal artifact reduction algorithms(O-MAR) in reduction of different types of artifacts caused by 125I seeds implantation and in improvement of the post-operative CT image quality. Methods This was a cross-sectional study. Thirty-five patients who underwent dual-layer detector spectral CT scanning of the chest and abdomen after 125I seeds implantation were retrospectively included at the First Affiliated Hospital of Zhengzhou University from March to September 2022. The spectral data were collected and reconstructed into conventional CT image (CI), VMI image (50-150 keV, 20 keV/level), CI+O-MAR image, and VMI+O-MAR image (50-150 keV, 20 keV/level). The artifacts′ removal effects and image quality improvement in each group were evaluated. Two slices with the strongest artifacts were selected for analysis for each patient, resulting in a total of 70 slices. Objective indicators including artifact index (AI), signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of soft tissue regions affected by artifacts were measured and calculated. Subjective indicators including the overcorrected artifacts and new artifacts, the different forms of artifacts, the diagnosis of artifacts, and the image quality were assessed. One-way analysis of variance was used for comparisons among multiple groups. Paired t test was used to compare the quantitative indicators between the combined O-MAR group and the non-O-MAR group. Kappa statistics was used to evaluate the consistency between observers. Results In high/low-density artifacts (ROIH/L), the AI values in all groups showed decrease with increasing VMI keV. In artifact-affected tissue (ROIT), SNR of the CI/VMI (70-150 keV)+O-MAR group were significantly higher than those of the CI/VMI group (P<0.05), CNR of the CI/VMI(50-150 keV)+O-MAR group were significantly higher than those of the CI/VMI group (P<0.05). Both overcorrection and new artifacts mainly presented in VMI 50 keV and VMI 70 keV groups Compared with VMI (50-70 keV) group, significantly less numbers of overcorrection and new artifacts were found in VMI (50-70 keV)+O-MAR group (P<0.05) regarding the comparison of artifact types, with the VMI keV increasing, the number of a-type banded artifacts gradually decreased on images with high-density artifacts, reaching a minimum of 3 in the VMI 150 keV+O-MAR group while the number of e-type artifacts with little or no artifacts increased, with the highest number of 23 in the VMI 150 keV+O-MAR group. The total number of high-density artifacts in each type decreased with increasing VMI keV. As VMI keV increased, the diagnostic and image quality scores of high-density artifacts in each group were significantly higher than those of low-density artifacts in the VMI+O-MAR group (P<0.05). Conclusions VMI combined with O-MAR can significantly improve the objective and subjective image quality of follow-up CT imaging after 125I seed implantation, enhancing lesion visibility and diagnostic confidence. Additionally, VMI+O-MAR showed more pronounced correction effect on high-density artifacts.

Abstract

Objective To investigate the value of the virtual monoenergetic image (VMI) obtained by a new dual-layer detector spectral CT combined with metal artifact reduction algorithms(O-MAR) in reduction of different types of artifacts caused by 125I seeds implantation and in improvement of the post-operative CT image quality. Methods This was a cross-sectional study. Thirty-five patients who underwent dual-layer detector spectral CT scanning of the chest and abdomen after 125I seeds implantation were retrospectively included at the First Affiliated Hospital of Zhengzhou University from March to September 2022. The spectral data were collected and reconstructed into conventional CT image (CI), VMI image (50-150 keV, 20 keV/level), CI+O-MAR image, and VMI+O-MAR image (50-150 keV, 20 keV/level). The artifacts′ removal effects and image quality improvement in each group were evaluated. Two slices with the strongest artifacts were selected for analysis for each patient, resulting in a total of 70 slices. Objective indicators including artifact index (AI), signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of soft tissue regions affected by artifacts were measured and calculated. Subjective indicators including the overcorrected artifacts and new artifacts, the different forms of artifacts, the diagnosis of artifacts, and the image quality were assessed. One-way analysis of variance was used for comparisons among multiple groups. Paired t test was used to compare the quantitative indicators between the combined O-MAR group and the non-O-MAR group. Kappa statistics was used to evaluate the consistency between observers. Results In high/low-density artifacts (ROIH/L), the AI values in all groups showed decrease with increasing VMI keV. In artifact-affected tissue (ROIT), SNR of the CI/VMI (70-150 keV)+O-MAR group were significantly higher than those of the CI/VMI group (P<0.05), CNR of the CI/VMI(50-150 keV)+O-MAR group were significantly higher than those of the CI/VMI group (P<0.05). Both overcorrection and new artifacts mainly presented in VMI 50 keV and VMI 70 keV groups Compared with VMI (50-70 keV) group, significantly less numbers of overcorrection and new artifacts were found in VMI (50-70 keV)+O-MAR group (P<0.05) regarding the comparison of artifact types, with the VMI keV increasing, the number of a-type banded artifacts gradually decreased on images with high-density artifacts, reaching a minimum of 3 in the VMI 150 keV+O-MAR group while the number of e-type artifacts with little or no artifacts increased, with the highest number of 23 in the VMI 150 keV+O-MAR group. The total number of high-density artifacts in each type decreased with increasing VMI keV. As VMI keV increased, the diagnostic and image quality scores of high-density artifacts in each group were significantly higher than those of low-density artifacts in the VMI+O-MAR group (P<0.05). Conclusions VMI combined with O-MAR can significantly improve the objective and subjective image quality of follow-up CT imaging after 125I seed implantation, enhancing lesion visibility and diagnostic confidence. Additionally, VMI+O-MAR showed more pronounced correction effect on high-density artifacts.

关键词

体层摄影术,X线计算机/125I粒子植入/虚拟单能级成像/金属去伪影技术/图像质量评估

Key words

Tomography, X-ray computed/125I seeds implantation/Virtual monochromatic imaging/Orthopedic metal artifact reduction algorithms/Image quality assessment

引用本文复制引用

基金项目

北京市希思科临床肿瘤研究基金(Y-2019AZMS-0505)

出版年

2024
中华放射学杂志
中华医学会

中华放射学杂志

CSTPCDCSCD北大核心
影响因子:1.756
ISSN:1005-1201
参考文献量24
段落导航相关论文