Department of Engineering and Applied Physics, University of Science and Technology of China, Hefei, Anhui, China.
Cancer Center, Sichuan Academy of Medical Sciences · Sichuan Provincial People's Hospital, Chengdu, Sichuan, China.
J Appl Clin Med Phys. 2022 Nov;23(11):e13775. doi: 10.1002/acm2.13775. Epub 2022 Sep 28.
The purpose of this work is to develop and evaluate a novel cycle-contrastive unpaired translation network (cycleCUT) for synthetic computed tomography (sCT) generation from T1-weighted magnetic resonance images (MRI).
The cycleCUT proposed in this work integrated the contrastive learning module from contrastive unpaired translation network (CUT) into the cycle-consistent generative adversarial network (cycleGAN) framework to effectively achieve unsupervised CT synthesis from MRI. The diagnostic MRI and radiotherapy planning CT images of 24 brain cancer patients were obtained and reshuffled to train the network. For comparison, the traditional cycleGAN and CUT were also implemented. The sCT images were then imported into a treatment planning system to verify their feasibility for radiotherapy planning. The mean absolute error (MAE), peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM) between the sCT and the corresponding real CT images were calculated. Gamma analysis between sCT- and CT-based dose distributions was also conducted.
Quantitative evaluation of an independent test set of six patients showed that the average MAE was 69.62 ± 5.68 Hounsfield Units (HU) for the proposed cycleCUT, significantly (p-value < 0.05) lower than that for cycleGAN (77.02 ± 6.00 HU) and CUT (78.05 ± 8.29). The average PSNR was 28.73 ± 0.46 decibels (dB) for cycleCUT, significantly larger than that for cycleGAN (27.96 ± 0.49 dB) and CUT (27.95 ± 0.69 dB). The average SSIM for cycleCUT (0.918 ± 0.012) was also significantly higher than that for cycleGAN (0.906 ± 0.012) and CUT (0.903 ± 0.015). Regarding gamma analysis, cycleCUT achieved the highest passing rate (97.95 ± 1.24% at the 2%/2 mm criteria and 10% dose threshold) but was not significantly different from the others.
The proposed cycleCUT could be effectively trained using unaligned image data, and could generate better sCT images than cycleGAN and CUT in terms of HU number accuracy and fine structural details.
本研究旨在开发并评估一种新颖的循环对比无配对翻译网络（cycleCUT），以实现基于 T1 加权磁共振成像（MRI）的合成计算机断层扫描（sCT）生成。
本研究中提出的 cycleCUT 将对比无配对翻译网络（CUT）的对比学习模块集成到循环一致生成对抗网络（cycleGAN）框架中，以有效地实现无监督 CT 从 MRI 合成。从 24 例脑癌患者的诊断 MRI 和放射治疗计划 CT 图像中获取并重新排列以训练网络。为了比较，还实现了传统的 cycleGAN 和 CUT。然后将 sCT 图像导入治疗计划系统，以验证其用于放射治疗计划的可行性。计算 sCT 与相应真实 CT 图像之间的平均绝对误差（MAE）、峰值信噪比（PSNR）和结构相似性指数（SSIM）。还对 sCT-和 CT 剂量分布之间的伽马分析进行了比较。
对 6 例患者的独立测试集进行的定量评估表明，对于所提出的 cycleCUT，平均 MAE 为 69.62±5.68 亨氏单位（HU），显著低于 cycleGAN（77.02±6.00 HU）和 CUT（78.05±8.29 HU）。cycleCUT 的平均 PSNR 为 28.73±0.46 分贝（dB），明显大于 cycleGAN（27.96±0.49 dB）和 CUT（27.95±0.69 dB）。cycleCUT 的平均 SSIM（0.918±0.012）也明显高于 cycleGAN（0.906±0.012）和 CUT（0.903±0.015）。关于伽马分析，cycleCUT 达到了最高通过率（在 2%/2mm 标准和 10%剂量阈值下为 97.95±1.24%），但与其他方法没有显著差异。
所提出的 cycleCUT 可以使用未对齐的图像数据进行有效训练，并且在 HU 数量准确性和精细结构细节方面可以比 cycleGAN 和 CUT 生成更好的 sCT 图像。
