文献速递技术与安全PETCT

基于Transformer深度学习模型实现超快速多示踪剂全身PET成像

Ultrafast Multi-tracer Total-body PET Imaging Using a Transformer-Based Deep Learning Model.
PubMed 文献 · Acad Radiol · 2025-08-30

与原始超快速扫描相比,两种模型均显著提升主观图像质量和病灶可检测性,并改善客观图像质量指标。

要点

  • 两种模型均显著提升多示踪剂超快速全身PET的主观图像质量与病灶可检测性。
  • [18F]FDG数据集3秒采集PSNR:模型1为48.169±6.121,模型2为48.123±6.103,均高于超快速扫描的44.092±7.508(p<0.001)。
  • 模型1与模型2预测的PET图像在图像质量和病灶可检测性方面表现相当。

核心结论

  • 所提出的模型1和模型2能有效提升超快速采集多示踪剂全身PET扫描的图像质量。
  • 预测的PET图像在图像质量和病灶可检测性方面表现出相当的性能。

研究设计与样本

  • 回顾性分析单中心临床uEXPLORER PET/CT数据集,包括[18F]FDG(N=50)、[18F]FAPI(N=45)和[68Ga]FAPI(N=60)研究。
  • 每位患者均行标准300秒采集,并通过列表模式数据截断生成3、6、15、30和40秒超快速扫描PET图像。
  • 开发两种3D SwinUNETR-V2架构变体:模型1(仅PET输入)和模型2(PET+CT融合输入),在三个数据集上采用5折交叉验证进行训练和测试。

诊断标准与关键数值

  • 与原始超快速扫描相比,模型1和模型2显著提升多示踪剂PET图像的主观图像质量和病灶可检测性。
  • 模型1和模型2改善了客观图像质量指标。
  • [18F]FDG数据集3秒采集PSNR:模型1为48.169±6.121,模型2为48.123±6.103,超快速扫描为44.092±7.508,p<0.001。
  • [18F]FDG数据集6秒采集PSNR:模型1为48.997±5.960,模型2为48.461±5.897,超快速扫描为46.503±7.190,p<0.001。
  • [18F]FDG数据集15秒采集PSNR:模型1为50.310±5.674,模型2为50.042±5.734,超快速扫描为49.331±6.732,p<0.001。

临床应用与阅片要点

  • 原文未据此给出推荐。

局限性

  • 原文未报告局限性。

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)。

英文原文(摘要)· 点击展开

BACKGROUND: Reducing PET scan acquisition time to minimize motion-related artifacts and improving patient comfort is always demanding. This study proposes a deep-learning framework for synthesizing diagnostic-quality PET images from ultrafast scans in multi-tracer total-body PET imaging.

METHODS: A retrospective analysis was conducted on clinical uEXPLORER PET/CT datasets from a single institution, including [18F]FDG (N=50), [18F]FAPI (N=45) and [68Ga]FAPI (N=60) studies. Standard 300-s acquisitions were performed for each patient, with ultrafast scan PET images (3, 6, 15, 30, and 40 s) generated through list-mode data truncation. We developed two variants of a 3D SwinUNETR-V2 architecture: Model 1 (PET-only input) and Model 2 (PET+CT fusion input). The proposed methodology was trained and tested on all three datasets using 5-fold cross-validation.

RESULTS: The proposed Model 1 and Model 2 significantly enhanced subjective image quality and lesion detectability in multi-tracer PET images compared to the original ultrafast scans. Model 1 and Model 2 also improved objective image quality metrics. For the [¹⁸F]FDG datasets, both approaches improved peak signal-to-noise ratio (PSNR) metrics across ultra-short acquisitions: 3 s: 48.169±6.121 (Model 1) vs. 48.123±6.103 (Model 2) vs. 44.092±7.508 (ultrafast), p < 0.001; 6 s: 48.997±5.960 vs. 48.461±5.897 vs. 46.503±7.190, p < 0.001; 15 s: 50.310±5.674 vs. 50.042±5.734 vs. 49.331±6.732, p < 0.001.

CONCLUSION: The proposed Model 1 and Model 2 effectively enhance image quality of multi-tracer total-body PET scans with ultrafast acquisition times. The predicted PET images demonstrate comparable performance in terms of image quality and lesion detectability.

来源:Acad Radiol,查看原文。版权归原作者所有。

引用

  1. Sun H, Sanaat A, Yi W, et al. Ultrafast Multi-tracer Total-body PET Imaging Using a Transformer-Based Deep Learning Model. Acad Radiol. 2025. PMID: 40885633. doi:10.1016/j.acra.2025.08.015

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。