深度学习图像重建低剂量CT尿路造影与常规CT尿路造影的前瞻性对照研究
前瞻性单中心研究纳入排泄期CT尿路造影患者,低剂量深度学习重建组与常规自适应统计迭代重建组各26例。低剂量组有效辐射剂量显著低于常规组(2.01±0.44 mSv vs 6.9±1.46 mSv,p<0.001),总体图像质量平均评分>4分且与常规组相似,两组尿路多数节段的衰减值、SNR和CNR充足且相似。结论认为深度学习重建可在降低71%辐射剂量的同时提供与常规CTU相当的图像质量。
要点
- 低剂量深度学习重建CTU的有效辐射剂量为2.01±0.44 mSv,常规迭代重建组为6.9±1.46 mSv(p<0.001)。
- 低剂量组总体图像质量平均评分>4分,与常规组相似,尿路多数节段衰减值、SNR及CNR两组充足且相似。
- 深度学习重建可在降低71%辐射剂量的同时维持与常规CTU相当的图像质量。
关键影像表现/征象
- 低剂量深度学习重建组总体图像质量良好,平均评分>4分,与常规ASIR-V组相似。
- 两组在尿路多数节段的平均CT衰减值、信噪比和对比噪声比充足且相似。
临床应用与阅片要点
- 低剂量深度学习重建CTU扫描参数为100kV及噪声指数23的自动mA调制,常规ASIR-V组为100kV及噪声指数10。
- 低剂量深度学习重建CTU在排泄期可获得与常规ASIR-V方案相似的图像质量,同时有效辐射剂量显著更低(2.01±0.44 mSv vs 6.9±1.46 mSv,p<0.001)。
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
OBJECTIVES: To compare the image quality of low-dose CT urography (LD-CTU) using deep learning image reconstruction (DLIR) with conventional CTU (C-CTU) using adaptive statistical iterative reconstruction (ASIR-V).
METHODS: This was a prospective, single-institutional study using the excretory phase CTU images for analysis. Patients were assigned to the LD-DLIR group (100kV and automatic mA modulation for noise index (NI) of 23) and C-ASIR-V group (100kV and NI of 10) according to the scan protocols in the excretory phase. Two radiologists independently assessed the overall image quality, artifacts, noise and sharpness of urinary tracts. Additionally, the mean CT attenuation, signal-to-noise ratio (SNR) and contrast-to-noise (CNR) in the urinary tracts were evaluated.
RESULTS: 26 patients each were included in the LD-DLIR group (10 males and 16 females; mean age: 57.23 years, range: 33-76 years) and C-ASIR-V group (14 males and 12 females; mean age: 60 years, range: 33-77 years). LD-DLIR group used a significantly lower effective radiation dose compared with the C-ASIR-V group (2.01 ± 0.44 mSv vs 6.9 ± 1.46 mSv, p < 0.001). LD-DLIR group showed good overall image quality with average score >4 and was similar to that of the C-ASIR-V group. Both groups had adequate and similar attenuation value, SNR and CNR in most segments of urinary tracts.
CONCLUSION: It is feasibility to provide comparable image quality while reducing 71% radiation dose in low-dose CTU with a deep learning image reconstruction algorithm compared to the conventional CTU with ASIR-V.
ADVANCES IN KNOWLEDGE: (1) CT urography with deep learning reconstruction algorithm can reduce the radiation dose by 71% while still maintaining image quality.
引用
- Cheng Y, Han Y, Li J, et al. Low-dose CT urography using deep learning image reconstruction: a prospective study for comparison with conventional CT urography. Br J Radiol. 2021. PMID: 33571034. doi:10.1259/bjr.20201291
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。