探讨人工智能（artificial intelligence，AI）质控平台在提高中孕期胎儿系统性超声检查图像质量中的应用价值。

使用“产前超声AI智慧云平台”对2022年1月至2024年12月在湖北省妇幼保健院超声诊断科行中孕期胎儿系统性超声检查的图像进行智能质控，共计11 559人次，766 233张超声图像。参与质控的医师共37人，其中高级职称3人、副高级职称7人、中级职称20人、初级职称7人。应用χ²检验对3年间的图像标准率及病例合格率进行比较，并对各年度各级医师的切面图像标准率进行对比，评估AI质控平台对于提高图像质量的价值。

2022年至2024年间3年的质控切面图像总标准率分别为70.81%、73.65%、74.10%，差异有统计学意义（χ²=931.883，P=0.000），病例合格率分别为79.51%、84.38%、90.17%，差异有统计学意义（χ²=181.479，P=0.000），病例质控缺失率为18.51%、14.47%、8.84%。每年的图像总标准率、病例总合格率及各切面病例合格率逐步提高，并趋于稳定，病例缺失率逐年降低；其中，初级医师三年图像总标准率提升了4.39%，非标准率降低了2.17%，效果最为显著。

AI质控平台可以有效改进各级医师中孕期系统性超声检查的图像质量，提升医师对标准切面的认识，增强医师进行规范化中孕期系统性超声检查的意识。

To evaluate the effectiveness of an artificial intelligence (AI)-based quality control (QC) platform in improving the image quality of mid-trimester systematic ultrasound screening.

The "Prenatal Ultrasound AI Cloud Platform" was used to perform intelligent QC of ultrasound images obtained from 11559 mid-trimester systematic ultrasound screening examinations (766 233 images) conducted between January 2022 and December 2024. A total of 37 sonographers (including 3 senior, 7 associate senior, 20 intermediate, and 7 junior practitioners) participated in the study. The χ² test was applied to compare annual standard-plane acquisition rates and case qualification rates over the 3-year period, as well as differences in standard-plane acquisition rates among operators with different professional ranks. The effectiveness of the AI-based QC platform in improving image quality was subsequently assessed.

From 2022 to 2024, the overall standard-plane compliance rates increased from 70.81% to 73.65% and 74.10%, with significant differences across the three years (χ²=931.883, P=0.000). The case qualification rates also increased from 79.51% to 84.38% and 90.17% (χ²=181.479, P=0.000). Meanwhile, the rates of missing QC-required images decreased from 18.51% to 14.47% and 8.84%, respectively. Overall, image standardization and case qualification rates improved steadily over time, whereas missing-image rates showed a continuous decline. Among all operator groups, junior sonographers demonstrated the greatest improvement, with a 4.39% increase in the overall image standardization rate and a 2.17% reduction in the non-standard image rate over the three-year period.

The AI-based QC platform effectively enhances ultrasound image quality across operators of all skill levels, improves their recognition of standard sections, and strengthens adherence to standardized mid-trimester systematic ultrasound screening.