文献速递胸部腹部技术与安全CT

智能器官识别技术联合人工智能迭代重建算法在胸腹盆增强CT中的应用

Intelligent Organ Recognition Technology Combined With an Artificial Intelligence Iterative Reconstruction Algorithm in Contrast-enhanced CT of the Chest, Abdomen, and Pelvis.
PubMed 文献 · Acad Radiol · 2026-08-20

本研究将130例胸腹盆增强CT患者随机分为两组,A组采用常规剂量联合Karl迭代重建,B组采用智能器官识别技术联合AIIR算法。结果显示B组软组织窗最佳重建级别为AIIR 4级、肺窗为AIIR 1级,其图像质量指标均优于A组,且有效剂量及胸、腹盆部管电流分别降低36%、45%、33%。

要点

  • 胸腹盆增强CT中,AIIR算法软组织窗推荐4级、肺窗推荐1级。
  • 智能器官识别联合AIIR可显著降低辐射剂量与管电流。
  • 该技术可提升图像信噪比、对比噪声比及主观评分。

核心结论

  • 智能器官识别技术联合AIIR算法可显著改善胸腹盆增强CT图像质量并降低辐射剂量。
  • 软组织窗图像最佳重建级别为AIIR 4级,肺窗图像最佳重建级别为AIIR 1级。

研究设计与样本

  • 随机对照研究,共纳入130例患者,随机分为A组和B组。
  • A组为常规剂量组,采用Karl迭代重建算法;B组采用智能器官识别技术联合AIIR算法。
  • 比较B组不同重建级别图像质量后选出最佳级别,再比较A、B两组的图像质量、辐射剂量和平均管电流。

诊断标准与关键数值

  • B组软组织窗图像最佳重建级别为AIIR 4级,肺窗图像最佳重建级别为AIIR 1级。
  • B组最佳重建级别图像的信噪比、对比噪声比、品质因数及主观评分均值均高于A组(p < 0.001)。
  • 与A组相比,B组平均有效剂量、胸部管电流、腹盆部管电流分别降低36%、45%、33%(p < 0.001)。

临床应用与阅片要点

  • 在胸腹盆增强CT重建时,可分别选择AIIR 4级用于软组织窗、AIIR 1级用于肺窗以优化图像质量。
  • 智能器官识别技术联合AIIR有助于在降低辐射剂量的同时维持或提升图像质量。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

英文原文(摘要)· 点击展开

RATIONALE AND OBJECTIVES: To evaluate the application value of intelligent organ recognition technology combined with the artificial intelligence iterative reconstruction (AIIR) algorithm in contrast-enhanced chest-abdomen-pelvis (CAP) computed tomography (CT) and to determine the optimal reconstruction level for the AIIR algorithm. MATERIALS AND METHODS: 130 patients were randomly assigned to groups A and B. Group A was the conventional-dose group, with images reconstructed using the Karl iterative reconstruction algorithm. Group B utilized intelligent organ recognition technology, with images reconstructed using the AIIR algorithm. After comparing the image quality of group B and selecting the optimal reconstruction level for the AIIR algorithm, the image quality, radiation dose, and average tube current between groups A and B were compared.

RESULTS: The optimal reconstruction level for soft-tissue window images in group B was AIIR level 4, and for lung window images was AIIR level 1. The mean values of the signal-to-noise ratio, contrast-to-noise ratio, figure of merit, and subjective scores for the optimal reconstruction level images in group B were higher than those in group A (p < 0.001). Compared to group A, the mean effective dose, the tube current for the chest and abdominopelvic regions in group B were decreased by 36%, 45%, 33% respectively (p < 0.001).

CONCLUSION: The application of intelligent organ recognition technology, combined with the AIIR algorithm, significantly improved the image quality of contrast-enhanced CAP-CT while reducing radiation dose. AIIR level 4 should be the optimal reconstruction level for soft-tissue window images, and AIIR level 1 should be the optimal reconstruction level for lung window images.

来源:Acad Radiol,查看原文。版权归原作者所有。

引用

  1. Li HS, Wang Q, Dong YX, et al. Intelligent Organ Recognition Technology Combined With an Artificial Intelligence Iterative Reconstruction Algorithm in Contrast-enhanced CT of the Chest, Abdomen, and Pelvis. Acad Radiol. 2026. PMID: 42624732. doi:10.1016/j.acra.2026.07.070

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。