Department of Abdominal Imaging, The University of Texas MD Anderson Cancer Center, 1400 Pressler Street, Unit 1473, Houston, TX, 77030-4009, USA.
Faculty of Medicine, Alexandria University, Alexandria, Egypt.
Abdom Radiol (NY). 2023 Aug;48(8):2724-2756. doi: 10.1007/s00261-023-03966-2. Epub 2023 Jun 6.
To perform a systematic literature review and meta-analysis of the two most common commercially available deep-learning algorithms for CT.
We used PubMed, Scopus, Embase, and Web of Science to conduct systematic searches for studies assessing the most common commercially available deep-learning CT reconstruction algorithms: True Fidelity (TF) and Advanced intelligent Clear-IQ Engine (AiCE) in the abdomen of human participants since only these two algorithms currently have adequate published data for robust systematic analysis.
Forty-four articles fulfilled inclusion criteria. 32 studies evaluated TF and 12 studies assessed AiCE. DLR algorithms produced images with significantly less noise (22-57.3% less than IR) but preserved a desirable noise texture with increased contrast-to-noise ratios and improved lesion detectability on conventional CT. These improvements with DLR were similarly noted in dual-energy CT which was only assessed for a single vendor. Reported radiation reduction potential was 35.1-78.5%. Nine studies assessed observer performance with the two dedicated liver lesion studies being performed on the same vendor reconstruction (TF). These two studies indicate preserved low contrast liver lesion detection (> 5 mm) at CTDI 6.8 mGy (BMI 23.5 kg/m) to 12.2 mGy (BMI 29 kg/m). If smaller lesion detection and improved lesion characterization is needed, a CTDI of 13.6-34.9 mGy is needed in a normal weight to obese population. Mild signal loss and blurring have been reported at high DLR reconstruction strengths.
Deep learning reconstructions significantly improve image quality in CT of the abdomen. Assessment of other dose levels and clinical indications is needed. Careful choice of radiation dose levels is necessary, particularly for small liver lesion assessment.
对两种最常见的商用深度学习算法在 CT 中的应用进行系统的文献回顾和荟萃分析。
我们使用 PubMed、Scopus、Embase 和 Web of Science 对评估最常见的两种商用深度学习 CT 重建算法（True Fidelity[TF]和 Advanced intelligent Clear-IQ Engine[AiCE]）的研究进行了系统检索，这些研究的对象是人体腹部的参与者。因为目前只有这两种算法有足够的已发表数据可以进行稳健的系统分析。
有 44 篇文章符合纳入标准。其中 32 项研究评估了 TF，12 项研究评估了 AiCE。与传统 CT 相比，DLR 算法生成的图像噪声明显减少（比 IR 低 22-57.3%），但保留了理想的噪声纹理，对比度噪声比提高，病灶检测能力增强。在仅评估单个供应商的双能 CT 中也观察到了 DLR 的这些改进。报告的辐射减少潜力为 35.1-78.5%。有 9 项研究评估了这两种专用肝病灶研究的观察者性能，这两项研究都是在同一供应商的重建（TF）上进行的。这两项研究表明，在 CTDI 为 6.8 mGy（BMI 为 23.5 kg/m）至 12.2 mGy（BMI 为 29 kg/m）时，可保留低对比度肝病灶的检测（>5 mm）。如果需要检测更小的病灶和改善病灶特征，在正常体重到肥胖人群中需要 CTDI 为 13.6-34.9 mGy。在高 DLR 重建强度下，已报道有轻微的信号丢失和模糊。
深度学习重建可显著改善腹部 CT 的图像质量。需要评估其他剂量水平和临床适应证。需要谨慎选择辐射剂量水平，特别是对于小肝病灶的评估。
