[论文解读] A deep learning classification scheme based on augmented-enhanced features to segment organs at risk on the optic region in brain cancer patients

José Dolz, Nicolas Reyns+5 more|arXiv (Cornell University)|Mar 30, 2017

Brain Tumor Detection and Classification参考文献 11被引用 8

一句话总结

本研究提出一种基于增强特征（AE-FV）的深度学习分类方法，可自动分割脑癌患者MRI中的危及器官（OARs）——视神经、视交叉、垂体腺及垂体柄。通过将经典特征、上下文特征与纹理特征整合至堆叠去噪自编码器框架中，该方法在平均Dice分数上达到0.79–0.83，优于传统特征与支持向量机（SVM）方法，且具有更快、更一致的结果表现。

ABSTRACT

Radiation therapy has emerged as one of the preferred techniques to treat brain cancer patients. During treatment, a very high dose of radiation is delivered to a very narrow area. Prescribed radiation therapy for brain cancer requires precisely defining the target treatment area, as well as delineating vital brain structures which must be spared from radiotoxicity. Nevertheless, delineation task is usually still manually performed, which is inefficient and operator-dependent. Several attempts of automatizing this process have reported. however, marginal results when analyzing organs in the optic region. In this work we present a deep learning classification scheme based on augmented-enhanced features to automatically segment organs at risk (OARs) in the optic region -optic nerves, optic chiasm, pituitary gland and pituitary stalk-. Fifteen MR images with various types of brain tumors were retrospectively collected to undergo manual and automatic segmentation. Mean Dice Similarity coefficients around 0.80 were reported. Incorporation of proposed features yielded to improvements on the segmentation. Compared with support vector machines, our method achieved better performance with less variation on the results, as well as a considerably reduction on the classification time. Performance of the proposed approach was also evaluated with respect to manual contours. In this case, results obtained from the automatic contours mostly lie on the variability of the observers, showing no significant differences with respect to them. These results suggest therefore that the proposed system is more accurate than other presented approaches, up to date, to segment these structures. The speed, reproducibility, and robustness of the process make the proposed deep learning-based classification system a valuable tool for assisting in the delineation task of small OARs in brain cancer.

研究动机与目标

• 为解决在脑癌放疗过程中，视区小而复杂的危及器官（OARs）手动勾画耗时、不一致的问题。
• 提升对小尺寸OARs（视神经、视交叉、垂体腺及垂体柄）自动分割的准确性，因以往方法在该任务中表现有限。
• 将手工设计特征（上下文、纹理）与深度学习表征相结合，以增强在低对比度、异质性脑结构中的特征判别能力。
• 开发一种快速、稳健且可重复的分割系统，适用于临床放疗计划。
• 与人工勾画结果对比评估性能，并证明其等效或优于现有最先进方法。

提出的方法

• 采用堆叠去噪自编码器（SDA）从MRI图像块中学习分层、具有判别力的特征，以基于特征工程设计的架构替代标准卷积神经网络（CNNs）。
• 通过结合经典特征（强度、位置）与上下文、纹理特征，构建增强型特征向量（AE-FV），以提升表征能力。
• 将AE-FV输入深度神经网络进行分类，使模型能够捕捉低对比度区域中的细微解剖模式。
• 在15例回顾性脑MRI扫描中进行模型训练与验证，使用人工分割作为真实标签。
• 采用Dice相似系数（DSC）、Hausdorff距离与表面距离指标评估性能，并与人工轮廓进行统计比较。
• 该方法避免依赖多模态影像融合（如CT-MR融合），从而实现更快处理速度并提升临床集成可行性。

实验结果

研究问题

• RQ1 将手工设计特征（上下文、纹理）与深度学习结合，能否提升视区小而复杂OARs的分割准确性？
• RQ2 所提出的基于AE-FV的深度学习模型在分割性能与处理时间方面，相较于传统机器学习方法（如SVM）表现如何？
• RQ3 该方法在多大程度上实现与人工勾画相当的结果，特别是在观察者变异方面？
• RQ4 该方法在分割视神经、视交叉与垂体结构方面，是否优于现有的基于图谱或模型的方法？
• RQ5 该方法能否在无需多模态影像融合（如CT与MRI）的情况下实现高精度分割？

主要发现

• 所提出的深度学习方法在视神经上获得0.79的平均Dice相似系数（DSC），视交叉为0.83，垂体腺为0.76，垂体柄为0.77。
• 与仅使用经典特征相比，引入增强特征（AE-FV）显著提升了分割性能。
• 该方法在准确率与处理时间方面均优于支持向量机（SVM），且各次运行结果变异更小。
• 在大多数情况下，自动分割结果与人工轮廓在统计上无显著差异，表明其具备临床可行性。
• 该方法表现出良好的鲁棒性与速度，分类时间显著短于以往基于图谱的方法（例如<20分钟 vs. 类似研究中约20分钟）。
• 即使在表现最差的案例中，DSC值仍保持在0.55–0.60以上，表明自动勾画结果仍为人工参考的可靠近似。