Development of 3D patient-based super-resolution digital breast phantoms using machine learning.
Development of 3D patient-based super-resolution digital breast phantoms using machine learning.
复制标题
DOI：
10.1088/1361-6560/aae78d
复制
发表时间：
2018-11-12
影响因子：
3.5
通讯作者：
Sechopoulos I
中科院分区：
文献类型：
--
作者：
Caballo M;Fedon C;Brombal L;Mann R;Longo R;Sechopoulos I
数字幻影是优化和评估x射线成像系统的重要工具，应该理想地反映人体解剖的三维结构及其潜在的变异性。此外，他们需要在足够高的空间分辨率下包括一个很好的细节水平，以准确地模拟人体解剖结构的连续性。提出了一种提高基于患者的数字乳房图像空间分辨率的方法，该方法可用于乳房成像的计算机模拟。给定有限分辨率的乳房断层成像图像，所提出的方法不仅可以简单地通过超采样，还可以通过生成额外的随机腺体细节来解释腺体边缘和链，以补偿由于所使用的成像系统的有限空间分辨率而在原始图像中可能未检测到的那些细节，从而随意生成幻象并增加其分辨率。提出的算法使用监督学习来预测由于分辨率有限而导致的腺体损失，然后通过学习低分辨率和高分辨率图像之间的映射来真实地恢复这种损失。在7个体素维度（60μm-480μm）重构的高分辨率同步加速器图像（检测器像素尺寸为60μm）上对其进行训练，并应用于临床乳腺CT系统（检测器像素尺寸为194μm）获取的患者图像（体素尺寸为68μm）生成超分辨率幻影。通过同步加速器（相对预测误差0.010±0.004，恢复精度0.95±0.04）和临床图像（平均腺体误差为194μm: 0.15%±0.12%）对所建立方法的适用性进行了评估。最后，乳房放射科医生通过盲目比较原始图像和幻象图像来评估发展的幻象的真实性，导致无法区分真实图像和幻象图像。总之，所提出的方法可以从乳房断层成像患者图像中生成超分辨率的幻影，可用于未来的计算机模拟，以优化新的乳房成像技术。
Digital phantoms are important tools for optimization and evaluation of x-ray imaging systems, and should ideally reflect the three-dimensional structure of human anatomy and its potential variability. In addition, they need to include a good level of detail at a high enough spatial resolution to accurately model the continuous nature of the human anatomy. A pipeline to increase the spatial resolution of patient-based digital breast phantoms that can be used for computer simulations of breast imaging is proposed. Given a tomographic breast image of finite resolution, the proposed methods can generate a phantom and increase its resolution at will, not only simply through super-sampling, but also by generating additional random glandular details to account for glandular edges and strands to compensate for those that may have not been detected in the original image due to the limited spatial resolution of the imaging system used. The proposed algorithms use supervised learning to predict the loss in glandularity due to limited resolution, and then to realistically recover this loss by learning the mapping between low and high resolution images. They were trained on high-resolution synchrotron images (detector pixel size 60μm) reconstructed at seven voxel dimensions (60μm-480μm), and applied to patient images acquired with a clinical breast CT system (detector pixel size 194μm) to generate super-resolution phantoms (voxel sizes 68μm). Several evaluations were made to assess the appropriateness of the developed methods, both with the synchrotron (relative prediction error 0.010±0.004, recovering accuracy 0.95±0.04), and with the clinical images (average glandularity error at 194μm: 0.15%±0.12%). Finally, a breast radiologist assessed the realism of the developed phantoms by blindly comparing original and phantom images, resulting in not being able to distinguish the real from the phantom images. In conclusion, the proposed method can generate super-resolution phantoms from tomographic breast patient images that can be used for future computer simulations for optimization of new breast imaging technologies.
登录
查看更多内容
影响因子：
2.5
作者：
Brombal, Luca;Donato, Sandro;Golosio, Bruno
通讯作者：
Golosio, Bruno
影响因子：
3.8
作者：
Caballo M;Boone JM;Mann R;Sechopoulos I
通讯作者：
Sechopoulos I
影响因子：
3.8
作者：
Caballo M;Mann R;Sechopoulos I
通讯作者：
Sechopoulos I
影响因子：
3.8
作者：
Pokrajac, David D.;Maidment, Andrew D. A.;Bakic, Predrag R.
通讯作者：
Bakic, Predrag R.
DOI：
10.1016/j.ejmp.2016.04.011
发表时间：
2016-05-01
影响因子：
3.4
作者：
Sarno, A.;Mettivier, G.;Russo, P.
通讯作者：
Russo, P.