Synthetic CT Generation of the Pelvis in Patients With Cervical Cancer: A Single Input Approach Using Generative Adversarial Network.
Synthetic CT Generation of the Pelvis in Patients With Cervical Cancer: A Single Input Approach Using Generative Adversarial Network.
复制标题
DOI：
10.1109/access.2021.3049781
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
Muzic RF
中科院分区：
文献类型：
--
作者：
Baydoun A;Xu KE;Heo JU;Yang H;Zhou F;Bethell LA;Fredman ET;Ellis RJ;Podder TK;Traughber MS;Paspulati RM;Qian P;Traughber BJ;Muzic RF
多模态成像构成了精确医学的基础，特别是在肿瘤学中，需要可靠和快速的成像技术以确保充分的诊断和治疗。在宫颈癌中，精确肿瘤学需要采集18F标记的2-氟-2-脱氧-D-葡萄糖（FDG）正电子发射断层扫描（PET）、磁共振（MR）和计算机断层扫描（CT）图像。此后，图像被共同配准以导出FDG-PET衰减校正和放射治疗规划所需的电子密度属性。然而，这种传统的方法受到MR-CT配准缺陷的影响，增加了治疗费用，并增加了患者的辐射暴露。为了克服这些缺点，我们提出了一个新的框架，跨模态图像合成，我们应用于MR-CT图像翻译宫颈癌的诊断和治疗。该框架基于条件生成对抗网络（cGAN），并说明了一种新颖的策略，可以简单但有效地解决深度学习中消失梯度与特征提取的范式。其贡献总结如下：1）该方法-称为sU-cGAN-首次使用编码器/解码器深度为2的浅U网（sU-Net）作为生成器; 2）sU-cGAN的输入是用于放射诊断的相同MR序列，即T2加权的Turbo自旋回波单次激发（TSE-SSH）MR图像; 3）尽管训练数据有限，且采用单输入通道方法，sU-cGAN仍优于其他最先进的深度学习方法，并实现了精确的合成CT（sCT）生成。总之，建议的框架应在临床环境中进一步研究。此外，sU-Net模型值得在其他计算机视觉任务中探索。
Multi-modality imaging constitutes a foundation of precision medicine, especially in oncology where reliable and rapid imaging techniques are needed in order to insure adequate diagnosis and treatment. In cervical cancer, precision oncology requires the acquisition of 18F-labeled 2-fluoro-2-deoxy-D-glucose (FDG) positron emission tomography (PET), magnetic resonance (MR), and computed tomography (CT) images. Thereafter, images are co-registered to derive electron density attributes required for FDG-PET attenuation correction and radiation therapy planning. Nevertheless, this traditional approach is subject to MR-CT registration defects, expands treatment expenses, and increases the patient’s radiation exposure. To overcome these disadvantages, we propose a new framework for cross-modality image synthesis which we apply on MR-CT image translation for cervical cancer diagnosis and treatment. The framework is based on a conditional generative adversarial network (cGAN) and illustrates a novel tactic that addresses, simplistically but efficiently, the paradigm of vanishing gradient vs. feature extraction in deep learning. Its contributions are summarized as follows: 1) The approach –termed sU-cGAN-uses, for the first time, a shallow U-Net (sU-Net) with an encoder/decoder depth of 2 as generator; 2) sU-cGAN’s input is the same MR sequence that is used for radiological diagnosis, i.e. T2-weighted, Turbo Spin Echo Single Shot (TSE-SSH) MR images; 3) Despite limited training data and a single input channel approach, sU-cGAN outperforms other state of the art deep learning methods and enables accurate synthetic CT (sCT) generation. In conclusion, the suggested framework should be studied further in the clinical settings. Moreover, the sU-Net model is worth exploring in other computer vision tasks.
登录
查看更多内容
影响因子：
3.5
作者：
Fetty, Lukas;Loefstedf, Tommy;Kuess, Peter
通讯作者：
Kuess, Peter
DOI：
10.1016/j.engappai.2018.11.013
发表时间：
2019-02-01
影响因子：
8
作者：
Ben-Cohen, Avi;Klang, Eyal;Greenspan, Hayit
通讯作者：
Greenspan, Hayit
DOI：
10.2967/jnumed.109.064824
发表时间：
2010-01
期刊：
Journal of nuclear medicine : official publication, Society of Nuclear Medicine
影响因子：
--
作者：
Fang YH;Asthana P;Salinas C;Huang HM;Muzic RF Jr
通讯作者：
Muzic RF Jr
影响因子：
19.7
作者：
Herold, Christian J.;Lewin, Jonathan S.;Hricak, Hedvig
通讯作者：
Hricak, Hedvig
影响因子：
3.8
作者：
Baldwin, Lesley N.;Wachowicz, Keith;Fallone, B. Gino
通讯作者：
Fallone, B. Gino