Institute of Physics "Gleb Wataghin", University of Campinas, Campinas, Brazil.
Institute of Physics "Gleb Wataghin", University of Campinas, Campinas, Brazil.
Phys Med. 2021 Mar;83:264-277. doi: 10.1016/j.ejmp.2021.03.007. Epub 2021 May 10.
Breast dosimetry in mammography is an important aspect of radioprotection since women are exposed periodically to ionizing radiation due to breast cancer screening programs. Mean glandular dose (MGD) is the standard quantity employed for the establishment of dose reference levels in retrospective population studies. However, MGD calculations requires breast glandularity estimation. This work proposes a deep learning framework for volume glandular fraction (VGF) estimations based on mammography images, which in turn are converted to glandularity values for MGD calculations.
208 virtual breast phantoms were generated and compressed computationally. The mammography images were obtained with Monte Carlo simulations (MC-GPU code) and a ray-tracing algorithm was employed for labeling the training data. The architectures of the neural networks are based on the XNet and multilayer perceptron, adapted for each task. The network predictions were compared with the ground truth using the coefficient of determination (r).
The results have shown a good agreement for inner breast segmentation (r = 0.999), breast volume prediction (r = 0.982) and VGF prediction (r = 0.935). Moreover, the DgN coefficients using the predicted VGF for the virtual population differ on average 1.3% from the ground truth values. Afterwards with the obtained DgN coefficients, the MGD values were estimated from exposure factors extracted from the DICOM header of a clinical cohort, with median(75 percentile) values of 1.91(2.45) mGy.
We successfully implemented a deep learning framework for VGF and MGD calculations for virtual breast phantoms.
由于乳腺癌筛查计划，女性会定期接受电离辐射，因此乳房剂量测定是放射防护的一个重要方面。乳腺平均剂量（MGD）是用于在回顾性人群研究中建立剂量参考水平的标准量。然而，MGD 的计算需要估计乳腺的腺体含量。本研究提出了一种基于乳房 X 光照片的体积腺体分数（VGF）估计的深度学习框架，该框架可进一步转换为腺体含量，用于 MGD 计算。
生成了 208 个虚拟乳房体模并进行了计算机压缩。乳房 X 光图像是通过蒙特卡罗模拟（MC-GPU 代码）获得的，并采用射线追踪算法对训练数据进行标记。神经网络的架构基于 XNet 和多层感知机，针对每个任务进行了调整。使用决定系数（r）比较网络预测值与真实值。
结果表明，内乳分割（r=0.999）、乳房体积预测（r=0.982）和 VGF 预测（r=0.935）的结果具有很好的一致性。此外，使用预测的 VGF 对虚拟人群进行的 DgN 系数平均差异为 1.3%。然后，使用获得的 DgN 系数，从临床队列的 DICOM 标头中提取的曝光因子估算 MGD 值，中位数（75 百分位数）值为 1.91（2.45）mGy。
我们成功地为虚拟乳房体模实现了 VGF 和 MGD 计算的深度学习框架。
