基于深度学习的后处理降噪改善四分之一辐射剂量冠状动脉CT血管成像
回顾性纳入2022—2023年221例回顾性心电门控CCTA患者,其中40例双期无运动伪影者用于分析;四分之一剂量图像经外部数据集训练的残差密集网络降噪后,噪声从37 HU降至18 HU(全剂量22 HU),CAD-RADS一致性从中等(0.60)提升至优秀(0.82),诊断显著狭窄的AUC从0.93升至0.97(P=0.032)。
要点
- 四分之一剂量CCTA经DLNR降噪后噪声由37 HU降至18 HU,接近全剂量22 HU(P<0.001)。
- DLNR使四分之一剂量与全剂量CAD-RADS一致性由中等(κ=0.60)提升至优秀(κ=0.82)。
- 以全剂量为参考,降噪后诊断显著狭窄AUC 0.97优于原始四分之一剂量0.93(P=0.032)。
核心结论
- DLNR可显著改善四分之一剂量CCTA的图像质量、CAD-RADS评估一致性及以全剂量为参考的显著狭窄诊断效能。
研究设计与样本
- 回顾性研究,2022—2023年回顾性心电门控CCTA患者221例,最终分析40例(年龄71±7岁,男性24例),采用剂量调制使一个期相全剂量、另一期相四分之一剂量扫描,仅纳入双期冠状动脉无运动伪影者,迭代重建后由外部数据集训练的残差密集网络对四分之一剂量图像降噪。
关键影像表现/征象
- 四分之一剂量CCTA图像噪声37 HU,经DLNR降至18 HU,全剂量图像为22 HU(P<0.001)。
- DLNR降噪后图像与全剂量图像的CAD-RADS一致性由中等(κ=0.60,95%CI 0.41-0.78)提升至优秀(κ=0.82,95%CI 0.66-0.94)。
诊断标准与关键数值
- 以全剂量CCTA为参考,降噪后图像诊断显著狭窄AUC为0.97(95%CI 0.95-1.00),原始四分之一剂量图像AUC为0.93(95%CI 0.89-0.98),差异P=0.032。
- CAD-RADS一致性以Cohen kappa评估:原始四分之一剂量κ=0.60(95%CI 0.41-0.78),降噪后κ=0.82(95%CI 0.66-0.94)。
局限性
- 研究为回顾性设计,仅纳入双期冠状动脉无运动伪影的患者,最终分析样本为40例。
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
PURPOSE: To evaluate the impact of deep learning-based post-hoc noise reduction (DLNR) on image quality, coronary artery disease reporting and data system (CAD-RADS) assessment, and diagnostic performance in quarter-dose versus full-dose coronary CT angiography (CCTA) on external datasets.
MATERIALS AND METHODS: We retrospectively reviewed 221 patients who underwent retrospective electrocardiogram-gated CCTA in 2022-2023. Using dose modulation, either mid-diastole or end-systole was scanned at full dose depending on heart rates, and the other phase at quarter dose. Only patients with motion-free coronaries in both phases were included. Images were acquired using iterative reconstruction, and a residual dense network trained on external datasets denoised the quarter-dose images. Image quality was assessed by comparing noise levels using Tukey's test. Two radiologists independently assessed CAD-RADS, with agreement to full-dose images evaluated by Cohen's kappa. Diagnostic performance for significant stenosis referencing full-dose images was compared between quarter-dose and denoised images by the area under the receiver operating characteristic curve (AUC) using the DeLong test.
RESULTS: Among 40 cases (age, 71 ± 7 years; 24 males), DLNR reduced noise from 37 to 18 HU (P < 0.001) in quarter-dose CCTA (full-dose images: 22 HU), and improved CAD-RADS agreement from moderate (0.60 [95 % CI: 0.41-0.78]) to excellent (0.82 [95 % CI: 0.66-0.94]). Denoised images demonstrated a superior AUC (0.97 [95 % CI: 0.95-1.00]) for diagnosing significant stenosis compared with original quarter-dose images (0.93 [95 % CI: 0.89-0.98]; P = 0.032).
CONCLUSION: DLNR for quarter-dose CCTA significantly improved image quality, CAD-RADS agreement, and diagnostic performance for detecting significant stenosis referencing full-dose images.
引用
- Morikawa T, Nishii T, Tanabe Y, et al. Deep learning-based post-hoc noise reduction improves quarter-radiation-dose coronary CT angiography. Eur J Radiol. 2025. PMID: 40505606. doi:10.1016/j.ejrad.2025.112232
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。