CSIRO Health and Biosecurity, Australian E-Health Research Centre, Brisbane, Australia; School of Engineering and Built Environment, Griffith University, Brisbane, Australia.
CSIRO Health and Biosecurity, Australian E-Health Research Centre, Brisbane, Australia.
Comput Methods Programs Biomed. 2021 Aug;207:106127. doi: 10.1016/j.cmpb.2021.106127. Epub 2021 May 5.
Cerebral microbleeds (CMB) are important biomarkers of cerebrovascular diseases and cognitive dysfunctions. Susceptibility weighted imaging (SWI) is a common MRI sequence where CMB appear as small hypointense blobs. The prevalence of CMB in the population and in each scan is low, resulting in tedious and time-consuming visual assessment. Automated detection methods would be of value but are challenged by the CMB low prevalence, the presence of mimics such as blood vessels, and the difficulty to obtain sufficient ground truth for training and testing. In this paper, synthetic CMB (sCMB) generation using an analytical model is proposed for training and testing machine learning methods. The main aim is creating perfect synthetic ground truth as similar as reals, in high number, with a high diversity of shape, volume, intensity, and location to improve training of supervised methods.
sCMB were modelled with a random Gaussian shape and added to healthy brain locations. We compared training on our synthetic data to standard augmentation techniques. We performed a validation experiment using sCMB and report result for whole brain detection using a 10-fold cross validation design with an ensemble of 10 neural networks.
Performance was close to state of the art (~9 false positives per scan), when random forest was trained on synthetic only and tested on real lesion. Other experiments showed that top detection performance could be achieved when training on synthetic CMB only. Our dataset is made available, including a version with 37,000 synthetic lesions, that could be used for benchmarking and training.
Our proposed synthetic microbleeds model is a powerful data augmentation approach for CMB classification with and should be considered for training automated lesion detection system from MRI SWI.
脑微出血（CMB）是脑血管疾病和认知功能障碍的重要生物标志物。磁敏感加权成像（SWI）是一种常见的 MRI 序列，其中 CMB 表现为小的低信号斑点。CMB 在人群和每次扫描中的患病率都很低，导致视觉评估繁琐且耗时。自动化检测方法将具有价值，但由于 CMB 的低患病率、存在类似血管的模拟物以及难以获得足够的训练和测试真实数据，因此受到挑战。本文提出了一种基于分析模型的合成 CMB（sCMB）生成方法，用于训练和测试机器学习方法。主要目的是创建与真实数据尽可能相似的大量具有高多样性的形状、体积、强度和位置的完美合成真实数据，以提高监督方法的训练效果。
使用随机高斯形状对 sCMB 进行建模，并将其添加到健康脑区。我们比较了在我们的合成数据上进行训练与标准增强技术的效果。我们使用 sCMB 进行了验证实验，并报告了使用 10 折交叉验证设计和 10 个神经网络的集成进行全脑检测的结果。
当随机森林仅在合成数据上进行训练并在真实病变上进行测试时，性能接近最新水平（每个扫描约有 9 个假阳性）。其他实验表明，仅在合成 CMB 上进行训练时，可以获得最佳的检测性能。我们的数据集包括一个包含 37,000 个合成病变的版本，可用于基准测试和训练。
我们提出的合成微出血模型是一种用于 CMB 分类的强大数据增强方法，应该考虑用于从 MRI SWI 训练自动病变检测系统。
