[论文解读] Robust Automatic Whole Brain Extraction on Magnetic Resonance Imaging of Brain Tumor Patients using Dense-Vnet

Sara Ranjbar, Kyle Singleton+6 more|arXiv (Cornell University)|Jun 4, 2020

Brain Tumor Detection and Classification参考文献 44被引用 6

一句话总结

该论文提出DeepBrain，一种基于深度学习的全脑提取方法，采用Dense-Vnet架构，可准确分割脑肿瘤患者T1Gd和FLAIR MRI扫描中的脑组织。该模型在自动生成的概率掩码上进行训练，对受肿瘤影响的MRI图像实现了94.5%的Dice分数、96.4%的敏感度和98.5%的特异性，表明其在仅使用50组训练样本的情况下仍具备强大鲁棒性与数据效率。

ABSTRACT

Whole brain extraction, also known as skull stripping, is a process in neuroimaging in which non-brain tissue such as skull, eyeballs, skin, etc. are removed from neuroimages. Skull striping is a preliminary step in presurgical planning, cortical reconstruction, and automatic tumor segmentation. Despite a plethora of skull stripping approaches in the literature, few are sufficiently accurate for processing pathology-presenting MRIs, especially MRIs with brain tumors. In this work we propose a deep learning approach for skull striping common MRI sequences in oncology such as T1-weighted with gadolinium contrast (T1Gd) and T2-weighted fluid attenuated inversion recovery (FLAIR) in patients with brain tumors. We automatically created gray matter, white matter, and CSF probability masks using SPM12 software and merged the masks into one for a final whole-brain mask for model training. Dice agreement, sensitivity, and specificity of the model (referred herein as DeepBrain) was tested against manual brain masks. To assess data efficiency, we retrained our models using progressively fewer training data examples and calculated average dice scores on the test set for the models trained in each round. Further, we tested our model against MRI of healthy brains from the LBP40A dataset. Overall, DeepBrain yielded an average dice score of 94.5%, sensitivity of 96.4%, and specificity of 98.5% on brain tumor data. For healthy brains, model performance improved to a dice score of 96.2%, sensitivity of 96.6% and specificity of 99.2%. The data efficiency experiment showed that, for this specific task, comparable levels of accuracy could have been achieved with as few as 50 training samples. In conclusion, this study demonstrated that a deep learning model trained on minimally processed automatically-generated labels can generate more accurate brain masks on MRI of brain tumor patients within seconds.

研究动机与目标

• 开发一种针对脑肿瘤患者MRI扫描的鲁棒、自动化全脑提取方法，以应对传统方法因病理组织结构扭曲而失效的问题。
• 解决现有头骨剥离技术在T1Gd和FLAIR序列中应用时准确率有限的问题，尤其针对脑肿瘤患者。
• 评估在最小化处理、自动生成的标签而非人工分割上训练的深度学习模型的性能。
• 通过确定实现高性能所需的最少训练样本数量，评估数据效率。

提出的方法

• 该模型名为DeepBrain，是基于Dense-Vnet架构的3D U-Net变体，专为3D医学图像分割设计。
• 使用SPM12软件生成灰质、白质和脑脊液的概率图，并将其合并为单个全脑掩码用于训练。
• 网络在脑肿瘤患者的T1Gd和FLAIR MRI序列上端到端训练，使用这些自动生成的标签。
• 通过Dice相似性系数、敏感度和特异性与人工勾画的脑掩码对比，评估模型性能。
• 通过在逐步缩小的训练数据子集上重新训练模型，开展数据效率实验，以评估性能阈值。
• 还使用LBP40A数据集中的健康脑MRI对模型进行了评估，以比较在病理与非病理条件下的性能差异。

实验结果

研究问题

• RQ1 在自动生成的概率掩码上训练的深度学习模型，能否在脑肿瘤MRI扫描中实现高精度的全脑提取？
• RQ2 所提出的模型在脑肿瘤患者T1Gd和FLAIR序列上的性能与人工分割基准相比如何？
• RQ3 实现肿瘤影响MRI中全脑提取的稳健性能，所需的最少训练样本数量是多少？
• RQ4 模型在脑肿瘤与健康脑MRI扫描中的性能表现有何差异？
• RQ5 当模型在有限数据上进行训练时，其精度保持程度如何，从而体现其数据效率？

主要发现

• DeepBrain模型在脑肿瘤MRI扫描中实现了94.5%的平均Dice分数，表明与人工脑掩码具有高度重叠。
• 在受肿瘤影响的MRI图像上，记录到96.4%的敏感度和98.5%的特异性，表明其对非脑结构具有强检测与排除能力。
• 在LBP40A数据集的健康脑MRI上，模型性能提升至96.2%的Dice分数，敏感度为96.6%，特异性为99.2%。
• 数据效率实验表明，仅使用50组训练样本即可实现相近性能，表明模型能从有限数据中实现良好泛化。
• 模型在数秒内即可生成准确的脑掩码，适用于临床工作流程，如术前规划和肿瘤分割。
• 通过SPM12自动生成标签进行训练，实现了高性能，且无需耗时的人工分割。