文献速递神经核医学

不真实数据增强可提高基于深度学习的多巴胺转运体SPECT分类对不同站点和不同相机间变异的鲁棒性

Unrealistic Data Augmentation Improves the Robustness of Deep Learning-Based Classification of Dopamine Transporter SPECT Against Variability Between Sites and Between Cameras.
PubMed 文献 · J Nucl Med · 2024-09-03

该研究提出使用强不真实数据增强(基于高斯模糊和加性噪声)训练卷积神经网络(CNN),用于自动分类多巴胺转运体SPECT图像,并在两个独立数据集上验证其相对于无增强和nnU-Net增强的鲁棒性。结果显示,强不真实增强的CNN在较低和较高空间分辨率数据集上的总体准确率分别为0.989(95%CI 0.978-0.996)和0.975(95%CI 0.960-0.986),均优于无增强和nnU-Net增强(McNemar P<0.001)。结论认为强不真实数据增强可改善CNN分类对未见采集设置的泛化能力。

要点

  • 强不真实数据增强训练的CNN在较低分辨率数据集准确率为0.989(95%CI 0.978-0.996)。
  • 强不真实数据增强训练的CNN在较高分辨率数据集准确率为0.975(95%CI 0.960-0.986)。
  • 强不真实增强的准确率优于无增强和nnU-Net增强(McNemar P<0.001)。

核心结论

  • 强不真实数据增强使基于CNN的123I-2β-甲酯基-3β-(4-碘苯基)-N-(3-氟丙基)去甲托烷SPECT图像分类对未见采集设置具有更好的泛化能力。

研究设计与样本

  • 研究在包含1,100例123I-2β-甲酯基-3β-(4-碘苯基)-N-(3-氟丙基)去甲托烷SPECT图像的均质数据集上训练CNN,采用基于高斯模糊和加性噪声的强不真实数据增强。
  • 将强不真实数据增强与无增强及基于强度的nnU-Net增强在两个独立数据集上进行比较,两个数据集的空间分辨率分别为较低(n=645)和显著较高(n=640)。

主要发现

  • 强不真实增强训练的CNN在较低分辨率独立测试数据集上总体准确率为0.989(95%CI 0.978-0.996),无增强为0.960(95%CI 0.942-0.974),nnU-Net增强为0.972(95%CI 0.956-0.983)。
  • 强不真实增强训练的CNN在较高分辨率独立测试数据集上总体准确率为0.975(95%CI 0.960-0.986),无增强为0.953(95%CI 0.934-0.968),nnU-Net增强为0.950(95%CI 0.930-0.966)。
  • 强不真实增强与无增强及nnU-Net增强的所有比较McNemar P<0.001。

临床怎么用

  • 作者认为强不真实数据增强可改善CNN分类对未见采集设置的泛化能力,并假设该策略可迁移至其他核医学成像应用。

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

We propose strongly unrealistic data augmentation to improve the robustness of convolutional neural networks (CNNs) for automatic classification of dopamine transporter SPECT against the variability between sites and between cameras. Methods: A CNN was trained on a homogeneous dataset comprising 1,100 123I-labeled 2β-carbomethoxy-3β-(4-iodophenyl)-N-(3-fluoropropyl)nortropane SPECT images using strongly unrealistic data augmentation based on gaussian blurring and additive noise. Strongly unrealistic data augmentation was compared with no augmentation and intensity-based nnU-Net augmentation on 2 independent datasets with lower (n = 645) and considerably higher (n = 640) spatial resolution. Results: The CNN trained with strongly unrealistic augmentation achieved an overall accuracy of 0.989 (95% CI, 0.978-0.996) and 0.975 (95% CI, 0.960-0.986) in the independent test datasets, which was better than that without (0.960, 95% CI, 0.942-0.974; 0.953, 95% CI, 0.934-0.968) and with nnU-Net augmentation (0.972, 95% CI, 0.956-0.983; 0.950, 95% CI, 0.930-0.966) (all McNemar P < 0.001). Conclusion: Strongly unrealistic data augmentation results in better generalization of CNN-based classification of 123I-labeled 2β-carbomethoxy-3β-(4-iodophenyl)-N-(3-fluoropropyl)nortropane SPECT images to unseen acquisition settings. We hypothesize that this can be transferred to other nuclear imaging applications.

来源:J Nucl Med,查看原文。版权归原作者所有。

引用

  1. Buddenkotte T, Buchert R. Unrealistic Data Augmentation Improves the Robustness of Deep Learning-Based Classification of Dopamine Transporter SPECT Against Variability Between Sites and Between Cameras. J Nucl Med. 2024. PMID: 39054285. doi:10.2967/jnumed.124.267570

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。