使用CypixGAN框架将脑18F-FDG PET/CT图像跨模态翻译为液体衰减反转恢复图像
研究纳入143例接受PET/CT和MRI的患者,用CypixGAN(结合CycleGAN与pix2pix的L1损失函数)从18F-FDG PET和CT生成合成FLAIR图像,并与pix2pix和CycleGAN比较;CypixGAN在峰值信噪比和结构相似性指数上显著更高(P<0.001),合成FLAIR中的白质高信号与真实图像接近。
要点
- CypixGAN从脑18F-FDG PET/CT生成合成FLAIR图像,视觉质量优于pix2pix和CycleGAN。
- CypixGAN的峰值信噪比20.23±1.31、结构相似性指数0.80±0.02,均显著高于pix2pix和CycleGAN(P<0.001)。
- 合成FLAIR图像中的白质高信号与真实图像接近,表现为低绝对百分比体积差异和高Dice相似系数。
核心结论
- CypixGAN使用非配对图像仍保留空间信息,生成高质量FLAIR图像。
- 该框架可能有助于在MRI不可用时改善PET/CT的诊断性能和成本效益。
研究设计与样本
- 使用143例接受PET/CT和MRI的患者数据,其中训练79例、验证20例、测试44例。
- 采用基于生成对抗网络的深度学习框架CypixGAN,结合CycleGAN框架与pix2pix的L1损失函数。
- 分别用pix2pix、CycleGAN和CypixGAN生成合成FLAIR图像,并分割白质高信号,比较CypixGAN与其他框架的性能。
主要发现
- CypixGAN在生成合成FLAIR图像的视觉质量上优于pix2pix和CycleGAN。
- CypixGAN的峰值信噪比为20.23±1.31,显著高于pix2pix的19.35±1.43和CycleGAN的18.74±1.49(P<0.001)。
- CypixGAN的结构相似性指数为0.80±0.02,显著高于pix2pix的0.79±0.02和CycleGAN的0.78±0.02(P<0.001)。
- CypixGAN生成的合成FLAIR图像中白质高信号与真实图像接近,表现为低绝对百分比体积差异和高Dice相似系数。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
PURPOSE: PET/CT and MRI can accurately diagnose dementia but are expensive and inconvenient for patients. Therefore, we aimed to generate synthetic fluid-attenuated inversion recovery (FLAIR) images from 18 F-FDG PET and CT images of the human brain using a generative adversarial network (GAN)-based deep learning framework called the CypixGAN, which combined the CycleGAN framework with the L1 loss function of the pix2pix. PATIENTS AND METHODS: Data from 143 patients who underwent PET/CT and MRI were used for training (n = 79), validation (n = 20), and testing (n = 44) the deep learning frameworks. Synthetic FLAIR images were generated using the pix2pix, CycleGAN, and CypixGAN, and white matter hyperintensities (WMHs) were then segmented. The performance of CypixGAN was compared with that of the other frameworks. RESULTS: The CypixGAN outperformed the pix2pix and CycleGAN in generating synthetic FLAIR images with superior visual quality. Peak signal-to-noise ratio and structural similarity index (mean ± standard deviation) estimated using the CypixGAN (20.23 ± 1.31 and 0.80 ± 0.02, respectively) were significantly higher than those estimated using the pix2pix (19.35 ± 1.43 and 0.79 ± 0.02, respectively) and CycleGAN (18.74 ± 1.49 and 0.78 ± 0.02, respectively) ( P < 0.001). WMHs in synthetic FLAIR images generated using the CypixGAN closely resembled those in ground-truth images, as indicated by the low absolute percentage volume differences and high dice similarity coefficients. CONCLUSIONS: The CypixGAN generated high-quality FLAIR images owing to the preservation of spatial information despite using unpaired images. This framework may help improve diagnostic performance and cost-effectiveness of PET/CT when MRI scan is unavailable.
引用
- Lee S, Jung JH, Choi Y, et al. Cross-Modality Image Translation From Brain 18 F-FDG PET/CT Images to Fluid-Attenuated Inversion Recovery Images Using the CypixGAN Framework. Clin Nucl Med. 2024. PMID: 39325494. doi:10.1097/RLU.0000000000005441
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。