文献速递核素与放射性配体治疗放射性药物与辐射防护

深度神经网络在挑战性散射条件下用于快速准确估计定量SPECT/CT散射

A deep neural network for fast and accurate scatter estimation in quantitative SPECT/CT under challenging scatter conditions.
PubMed 文献 · Eur J Nucl Med Mol Imaging · 2020-05-15

研究训练深度卷积神经网络(DCNN)从90Y轫致辐射SPECT发射投影和CT衰减投影估计散射,在模拟球体模型、肝脏模型和90Y放射性栓塞患者中与蒙特卡罗(MC)散射估计比较。在10%噪声水平下DCNN-SC的平均球体对比恢复率为55%,MC-SC为52%;患者病灶-肝脏摄取比DCNN-SC与MC-SC一致性CCC=0.98,与90Y PET一致性CCC=0.96。DCNN生成全部128个投影的散射估计约需40秒,MC散射模型使用12个处理器约需80分钟。

要点

  • DCNN-SC与MC-SC的对比恢复率-噪声及肺插入残留误差-噪声曲线相似。
  • 10%噪声水平下平均球体对比恢复率:NO-SC 24%、MC-SC 52%、DCNN-SC 55%、TRUE-SC 67%。
  • 患者病灶-肝脏摄取比DCNN-SC与MC-SC一致性CCC=0.98,与90Y PET一致性CCC=0.96,与NO-SC一致性CCC=0.59。

核心结论

  • 在包括患者研究的多种90Y测试数据中,深度学习与基于MC的散射估计重建图像在剂量学和安全相关指标上表现相当。
  • 该方法可通过更换训练数据推广至其他核素,因速度比MC快数个数量级且保持高精度,适合实时临床使用。

研究设计与样本

  • 训练13层深度卷积神经网络(DCNN)从90Y轫致辐射SPECT发射投影和CT衰减投影分别估计散射投影,训练标签为体模中MC生成的「真实」散射投影。
  • 测试数据集包括带肺插入的模拟球体模型、肝脏模型测量和90Y放射性栓塞后患者。
  • 比较OS-EM SPECT重建:无散射校正(NO-SC)、真实散射(TRUE-SC,仅模拟数据可用)、DCNN估计散射(DCNN-SC)和先前开发的MC散射模型(MC-SC),并在可用时与90Y PET比较。

主要发现

  • 在10%相同噪声水平下(多次实现),平均球体对比恢复率(CR)为NO-SC 24%、MC-SC 52%、DCNN-SC 55%、TRUE-SC 67%。
  • 肝脏模型中肝脏插入的平均CR为NO-SC 32%、MC-SC 73%、DCNN-SC 65%;低浓度肝外插入的平均对比噪声比(可见性指数)分别为2、19、61。
  • 患者中病灶-肝脏摄取比DCNN-SC中位数4.8(范围0.02-13.8)与MC-SC中位数4.0(范围0.13-12.1;CCC=0.98)及90Y PET中位数4.9(范围0.02-11.2;CCC=0.96)高度一致,与NO-SC中位数2.8(范围0.3-7.2;CCC=0.59)一致性差。
  • 训练后的DCNN在台式计算机单i5处理器上生成患者扫描全部128个视图的散射估计约需40秒,MC散射模型使用12个处理器约需80分钟。

关键数值与推荐

  • 10%噪声水平下平均球体CR:NO-SC 24%、MC-SC 52%、DCNN-SC 55%、TRUE-SC 67%。
  • 肝脏模型肝脏插入平均CR:NO-SC 32%、MC-SC 73%、DCNN-SC 65%。
  • 低浓度肝外插入平均对比噪声比(可见性指数):NO-SC 2、MC-SC 19、DCNN-SC 61。
  • 患者病灶-肝脏摄取比一致性:DCNN-SC与MC-SC CCC=0.98;DCNN-SC与90Y PET CCC=0.96;DCNN-SC与NO-SC CCC=0.59。
  • 计算时间:DCNN约40秒(单i5处理器)对比MC散射模型约80分钟(12个处理器)。

临床怎么用

  • 原文结论指出该方法因高速(比MC快数个数量级)且保持高精度,适合实时临床使用。
  • 原文结论指出该方法可通过更换训练数据推广至其他核素。

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)。

原文

PURPOSE: A major challenge for accurate quantitative SPECT imaging of some radionuclides is the inadequacy of simple energy window-based scatter estimation methods, widely available on clinic systems. A deep learning approach for SPECT/CT scatter estimation is investigated as an alternative to computationally expensive Monte Carlo (MC) methods for challenging SPECT radionuclides, such as 90Y. METHODS: A deep convolutional neural network (DCNN) was trained to separately estimate each scatter projection from the measured 90Y bremsstrahlung SPECT emission projection and CT attenuation projection that form the network inputs. The 13-layer deep architecture consisted of separate paths for the emission and attenuation projection that are concatenated before the final convolution steps. The training label consisted of MC-generated "true" scatter projections in phantoms (MC is needed only for training) with the mean square difference relative to the model output serving as the loss function. The test data set included a simulated sphere phantom with a lung insert, measurements of a liver phantom, and patients after 90Y radioembolization. OS-EM SPECT reconstruction without scatter correction (NO-SC), with the true scatter (TRUE-SC) (available for simulated data only), with the DCNN estimated scatter (DCNN-SC), and with a previously developed MC scatter model (MC-SC) were compared, including with 90Y PET when available. RESULTS: The contrast recovery (CR) vs. noise and lung insert residual error vs. noise curves for images reconstructed with DCNN-SC and MC-SC estimates were similar. At the same noise level of 10% (across multiple realizations), the average sphere CR was 24%, 52%, 55%, and 67% for NO-SC, MC-SC, DCNN-SC, and TRUE-SC, respectively. For the liver phantom, the average CR for liver inserts were 32%, 73%, and 65% for NO-SC, MC-SC, and DCNN-SC, respectively while the corresponding values for average contrast-to-noise ratio (visibility index) in low-concentration extra-hepatic inserts were 2, 19, and 61, respectively. In patients, there was high concordance between lesion-to-liver uptake ratios for SPECT reconstruction with DCNN-SC (median 4.8, range 0.02-13.8) compared with MC-SC (median 4.0, range 0.13-12.1; CCC = 0.98) and with 90Y PET (median 4.9, range 0.02-11.2; CCC = 0.96) while the concordance with NO-SC was poor (median 2.8, range 0.3-7.2; CCC = 0.59). The trained DCNN took ~ 40 s (using a single i5 processor on a desktop computer) to generate the scatter estimates for all 128 views in a patient scan, compared to ~ 80 min for the MC scatter model using 12 processors. CONCLUSIONS: For diverse 90Y test data that included patient studies, we demonstrated comparable performance between images reconstructed with deep learning and MC-based scatter estimates using metrics relevant for dosimetry and for safety. This approach that can be generalized to other radionuclides by changing the training data is well suited for real-time clinical use because of the high speed, orders of magnitude faster than MC, while maintaining high accuracy.

来源:Eur J Nucl Med Mol Imaging,查看原文。版权归原作者所有。

引用

  1. Xiang H, Lim H, Fessler JA, et al. A deep neural network for fast and accurate scatter estimation in quantitative SPECT/CT under challenging scatter conditions. Eur J Nucl Med Mol Imaging. 2020. PMID: 32415551. doi:10.1007/s00259-020-04840-9

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。