文献速递胸部儿科影像技术与安全CT

使用深度卷积神经网络去噪降低儿童计算机断层扫描(CT)辐射剂量

Radiation dose reduction in pediatric computed tomography (CT) using deep convolutional neural network denoising.
PubMed 文献 · Clin Radiol · 2024-09-21

研究对40例因「慢性咳嗽」行非增强胸部CT的儿童,用验证过的噪声插入法模拟20%剂量数据,训练深度卷积神经网络(CNN)去噪并评估30例。结果显示常规剂量联合CNN(RD+CNN,3.6±1.1)与模拟低剂量联合CNN(TD+CNN,3.4±0.9)的主观图像质量均高于常规剂量联合迭代重建(RD+IR,3.1±0.9),主观噪声伪影评分也均显著低于RD+IR(2.7±0.7),结论认为CNN去噪在降低儿童CT辐射剂量方面优于迭代重建。

要点

  • 常规剂量联合CNN去噪的图像质量评分(3.6±1.1)高于常规剂量联合迭代重建(3.1±0.9)。
  • 模拟20%剂量联合CNN去噪的图像质量评分(3.4±0.9)仍高于常规剂量联合迭代重建(3.1±0.9)。
  • 常规剂量联合CNN(3.2±0.9)与模拟低剂量联合CNN(2.9±0.6)的主观噪声伪影评分均低于常规剂量联合迭代重建(2.7±0.7)。

核心结论

  • CNN去噪作为降低儿童CT辐射剂量的手段优于迭代重建。

研究设计与样本

  • 3名获得资格证书(CAQ)认证的儿童放射科医师在4种条件下评估30例:常规剂量+迭代重建、常规剂量+CNN、模拟低剂量+迭代重建、模拟低剂量+CNN。

诊断标准与关键数值

  • 读者间信度:常规剂量+迭代重建平均κ=0.55,常规剂量+CNN平均κ=0.50,模拟低剂量+CNN平均κ=0.54,模拟低剂量+迭代重建平均κ=0.57。

临床应用与阅片要点

  • 原文未据此给出推荐。

局限性

  • 原文未报告局限性。

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

AIM: We evaluated the quality of noncontrast chest computed tomography (CT) for pediatric patients at two dose levels with and without denoising using a deep convolutional neural network (CNN).

MATERIALS AND METHODS: Forty children underwent noncontrast chest CTs for "chronic cough" using a routine dose (RD) protocol. Images were reconstructed using iterative reconstruction (IR). A validated noise insertion method was used to simulate 20% dose (TD) data for each case. A deep CNN model was trained and validated on 10 cases and then applied to the remaining 30 cases. Three certificate of qualification (CAQ)-certified pediatric radiologists evaluated 30 cases under 4 conditions: (1) RD + IR; (2) RD + CNN; (3) TD + IR; and (4) TD + CNN. Likert scales were used to score subjective image quality (1-5, 5 = excellent) and subjective noise artifact (1-4, 4 = no noise). Images were reviewed for specific findings.

RESULTS: For the 30 patients evaluated (14 female, mean age: 10.8 years, range: 0.17-17), the mean effective dose was 0.46 ± 0.21 mSv for the original RD exam, with an effective dose of 0.09 mSv for the TD exam. Both RD + CNN (3.6 ± 1.1, p < 0.001) and TD + CNN (3.4 ± 0.9, p = 0.023) had higher image quality than RD + IR (3.1 ± 0.9). Both RD + CNN (3.2 ± 0.9, p-value = <0.001) and TD + CNN (2.9 ± 0.6, p-value = 0.001) showed significantly lower subjective noise artifact scores than RD + IR (2.7 ± 0.7). There was excellent intrareader (RD + IR-RD + CNN: mean κ = 0.96, RD + IR-TD + CNN = 0.96, RD + IR-TD + IR = 0.98) and moderate inter-reader reliability (RD + IR: mean κ = 0.55, RD + CNN = 0.50, TD + CNN = 0.54, TD + IR = 0.57) on all 4 image reconstructions.

CONCLUSION: CNN denoising outperforms IR as a means of radiation dose reduction in pediatric CT.

来源:Clin Radiol,查看原文。版权归原作者所有。

引用

  1. Horst KK, Zhou Z, Hull NC, et al. Radiation dose reduction in pediatric computed tomography (CT) using deep convolutional neural network denoising. Clin Radiol. 2024. PMID: 39509751. doi:10.1016/j.crad.2024.09.011

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。