文献速递头颈部肿瘤放疗乳腺癌放疗消化道肿瘤放疗

终身学习实现三种临床靶区自动分割

Automatic segmentation of three clinical target volumes in radiotherapy using lifelong learning.
PubMed 文献 · Radiother Oncol · 2021-01-05

研究纳入鼻咽癌、乳腺癌和直肠癌各800例患者,提出采用扩张适配器的终身学习方法逐项学习三种临床靶区分割任务。与单任务、多任务或迁移学习相比,该方法分割精度相当或更优(DSC:鼻咽癌0.86、乳腺癌0.89、直肠癌0.87),并减少约58.8%参数量、约55.6%模型大小和约25.0%训练时间。

要点

  • 终身学习逐任务学习鼻咽癌、乳腺癌、直肠癌临床靶区分割,避免灾难性遗忘。
  • 分割精度DSC分别为鼻咽癌0.86、乳腺癌0.89、直肠癌0.87。
  • 较单任务或迁移学习减少约58.8%参数量、约55.6%模型大小、约25.0%训练时间。

核心结论

  • 所提终身学习方法利用扩张适配器在学习新任务时保留既往任务知识,性能相当且所需训练数据、模型参数和训练时间更少。

研究设计与样本

  • 队列包括鼻咽癌、乳腺癌和直肠癌各800例患者,任务为三种癌症的临床靶区分割。
  • 所提终身学习网络采用扩张适配器逐项学习三个分割任务,仅微调新添加的七层扩张适配器,其余已学习层冻结。

主要发现

  • 与单任务、多任务或迁移学习相比,终身学习分割精度更优或相当,DSC为鼻咽癌0.86、乳腺癌0.89、直肠癌0.87。
  • 终身学习可避免序贯学习中的遗忘,并以更少训练数据获得良好性能。
  • 终身学习较单任务或迁移学习分别减少约58.8%参数量、约55.6%模型大小和约25.0%训练时间。

关键数值与推荐

  • 分割精度DSC:鼻咽癌0.86、乳腺癌0.89、直肠癌0.87。
  • 参数量减少约58.8%,模型大小减少约55.6%,训练时间减少约25.0%。

临床怎么用

  • 该终身学习方法可用于多癌种临床靶区自动分割模型的连续学习,减少遗忘并降低训练资源需求。

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

BACKGROUND AND PURPOSE: Convolutional neural networks (CNNs) have comparable human level performance in automatic segmentation. An important challenge that CNNs face in segmentation is catastrophic forgetting. They lose performance on tasks that were previously learned when trained on task. In this study, we propose a lifelong learning method to learn multiple segmentation tasks continuously without forgetting previous tasks. MATERIALS AND METHODS: The cohort included three tumors, 800 patients of which had nasopharyngeal cancer (NPC), 800 patients had breast cancer, and 800 patients had rectal cancer. The tasks included segmentation of the clinical target volume (CTV) of these three cancers. The proposed lifelong learning network adopted dilation adapter to learn three segmentation tasks one by one. Only the newly added dilation adapter (seven layers) was fine tuning for incoming new task, whereas all the other learned layers were frozen. RESULTS: Compared with single-task, multi-task or transfer learning, the proposed lifelong learning can achieve better or comparable segmentation accuracy with a DSC of 0.86 for NPC, 0.89 for breast cancer, and 0.87 for rectal cancer. Lifelong learning can avoid forgetting in sequential learning and yield good performance with less training data. Furthermore, it is more efficient than single-task or transfer learning, which reduced the number of parameters, size of model, and training time by ~58.8%, ~55.6%, and ~25.0%, respectively. CONCLUSION: The proposed method preserved the knowledge of previous tasks while learning a new one using a dilation adapter. It could yield comparable performance with much less training data, model parameters, and training time.

来源:Radiother Oncol,查看原文。版权归原作者所有。

引用

  1. Men K, Chen X, Yang B, et al. Automatic segmentation of three clinical target volumes in radiotherapy using lifelong learning. Radiother Oncol. 2021. PMID: 33418008. doi:10.1016/j.radonc.2020.12.034

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。