| 曹源,刘娜,宁宁,等.深度学习重建算法在"三低"下肢CT血管成像中的应用价值[J].中华放射医学与防护杂志,2026,46(3):294-300.Cao Yuan,Liu Na,Ning Ning,et al.Application value of deep learning reconstruction algorithm in lower extremity CT angiography with "triple-low" protocol[J].Chin J Radiol Med Prot,2026,46(3):294-300 |
| 深度学习重建算法在"三低"下肢CT血管成像中的应用价值 |
| Application value of deep learning reconstruction algorithm in lower extremity CT angiography with "triple-low" protocol |
| 投稿时间：2025-06-01 |
| DOI：10.3760/cma.j.cn112271-20250601-00186 |
| 中文关键词: 下肢 CT血管成像 深度学习 低剂量 |
| 英文关键词:Lower extremities Computed tomography angiography (CTA) Deep learning Low dose |
| 基金项目:大连市科技创新基金(2024JJ13PT066) |
|
| 摘要点击次数: 3065 |
| 全文下载次数: 54 |
| 中文摘要: |
| 目的 评估深度学习重建算法(DLR)在低辐射剂量、低对比剂流速和低对比剂用量条件下("三低")对下肢CT血管成像(CTA)图像质量的改善效果。方法 该研究为横断面研究。前瞻性纳入2024年3月至11月在大连大学附属中山医院放射科行下肢CTA检查的60例患者,并根据区组随机化分为对照组和试验组,各30例。对照组采用管电压100 kV,对比剂100~120 ml,注射流速4.0~4.5 ml/s以及混合迭代重建算法(HIR);试验组采用管电压80 kV,对比剂75~85 ml,注射流速3.0~3.5 ml/s,并分别进行HIR和DLR重建。定量评估包括腹主动脉、股总动脉、股浅动脉、腘动脉、胫前动脉的CT值、噪声(SD)、信噪比(SNR)和对比噪声比(CNR)。由两名医师采用5分量表对图像质量进行主观评价。客观图像质量指标和主观评分的比较分别采用单因素方差分析。结果 在试验组中,DLR图像各条血管的CT值、SNR和CNR均高于HIR图像(t=-6.08~-0.14, P<0.05);除腘动脉和胫前动脉外,试验组-DLR其余血管的SNR和CNR均高于对照组(t=-4.78~-2.19, P<0.05)。试验组-DLR较对照组在腹主动脉、股浅动脉、腘动脉和胫前动脉处的CT值差异均无统计学意义(P>0.05),而SD值均显著低于对照组和试验组-HIR(t =3.43~12.53, P<0.05)。3组图像的主观评分差异具有统计学意义(t =-6.01~3.27, P＜0.05),试验组-DLR得分显著优于对照组和试验组-HIR。试验组的CTDIvol、DLP和碘负荷较对照组均显著降低(t=2.36、2.19、14.29, P＜0.05)。结论 深度学习重建算法能够在"三低"方案下获得诊断质量良好的下肢CTA图像,显著降低辐射剂量与对比剂用量,从而提升患者检查的安全性。 |
| 英文摘要: |
| Objective To evaluate the improvement effect of deep learning reconstruction (DLR) algorithm on image quality of lower extremity CT angiography (CTA) with a "triple-low" protocol (low radiation dose, low contrast agent flow rate and low contrast agent dosage). Methods This study was a cross-sectional study. Sixty patients who underwent lower-extremity CTA examinations in the Department of Radiology of Zhongshan Hospital Affiliated to Dalian University, from March to November 2024 were prospectively included and randomly divided into a control group (routine-dose, n = 30) and an experimental group (low-dose, n = 30). The control group uesd tube voltage of 100 kV, contrast agent of 100-120 ml, injection flow rate of 4.0-4.5 ml/s, and hybrid iterative reconstruction (HIR) algrithm. The experimental group used tube voltage of 80 kV, contrast agent of 75-85 ml, injection flow rate of 3.0-3.5 ml/s, and performed both HIR and DLR reconstructions. Quantitative assessment included CT values, noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) of the abdominal aorta, common femoral artery, superficial femoral artery, popliteal artery, and anterior tibial artery. Two physicians evaluated the image quality subjectively using a 5-point scale. Inter-group comparison was conducted using One-way ANOVA for both objective and subjective image quality. Results In the experimental group, the CT values, SNR, and CNR in theDLR images were all higher than those in the HIR images (t = -6.08 to -0.14, P < 0.05). Except for the popliteal artery and anterior tibial arteries, the SNR and CNR of the remaining vessels in the DLR images were higher than those in the control group (t = -4.78 to -2.19, P<0.05). There were no statistically significant differences in CT vlaues of the abdominal aorta, superficial femoral artery, popliteal artery and anterior tibial artery between DLR images of the experimental group and images of the control-group (P > 0.05). while the SD values were significantly lower than those of the control group and the HIR images of experimental group (t = 3.43-12.53, P < 0.05). The differences in subjective image quality scores among the three groups were statistically significant (t = -6.01-3.27, P < 0.05), and the score of the experimental group-DLR was significantly higher than that of the control group and the experimental group-HIR. The CTDIvol, DLP and iodine volume of the experimental group were significantly lower than those of the control group (t = 2.36, 2.19, 14.29, P < 0.05). Conclusion With a triple-low protocol, DLR algorithm can generate high quality lower-extremity CTA images, significantly reducing radiation dose and contrast agent usage, thereby enhangcing patient safety. |
| HTML 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|