人工智能质量控制平台对基层医院妇产科产前超声诊断水平的提升作用研究OA
Effect of AI-based quality control platform on improving diagnostic level of prenatal ultrasound of department of obstetrics and gynecology in primary hospitals
目的:探讨基于人工智能(AI)的质量控制平台在提升基层医疗机构产前超声标准切面获取质量与一致性的应用价值.方法:采用单中心回顾性研究,选取2021年1月至2025年2月在上海市松江区妇幼保健院进行常规产前超声检查的200例孕妇超声图像数据,将2021年1月至2022年12月的100例孕妇产前超声检查采用图像传统人工质量控制模式,2023年3月至2025年2月的100例孕妇产前超声检查图像应用AI质量控制模式.比较两种质量控制模式中孕妇不同孕期的超声标准切面达标率和关键诊断切面图像达标率.结果:采用AI质量控制模式的孕妇孕早期、孕中期Ⅰ级、孕中期Ⅱ级和孕中期Ⅲ级超声标准切面达标率分别为88%(88/100)、91%(91/100)、84%(84/100)和79%(79/100),均高于人工质量控制模式,差异均有统计学意义(x2=9.12、8.53、7.78、10.04,P<0.05).采用AI质量控制模式的孕妇超声检查关键诊断切面中的丘脑水平、四腔心、宫颈长轴、股骨长轴及颈后透明层厚度(NT)图像达标率均高于人工质量控制模式,差异均有统计学意义(x2=9.01、8.95、8.12、7.89、11.24,P<0.05).结论:AI质量控制平台应用能够有效提升基层医院产前超声标准切面获取能力与图像质量,提高产前超声标准切面整体及各关键诊断切面的图像达标率.
Objective:To explore application value of artificial intelligence(AI)-based quality control platform in enhancing the obtained quality and consistency of standard cross-section of prenatal ultrasound in healthcare institutions of grassroots.Methods:A single center retrospective study was adopted in this study.The image data of 200 pregnant women,who underwent conventionally prenatal ultrasound examination at the Songjiang District Maternal and Child Health Hospital of Shanghai between January 2021 and February 2025,were selected.The prenatal ultrasound examination of 100 pregnant women during January 2021 and December 2022 adopted conventionally manual quality control(QC)mode,and other 100 pregnant women during March 2023 and February 2025 adopted AI-based QC mode to conduct prenatal ultrasound examination.The compliance rate of standard cross-section of ultrasound,and the compliance rate of images in critically diagnostic cross-section in different stages of pregnancy between two kinds of QC modes were compared.Results:The compliance rates of standard cross-section of ultrasound of adopting AI-based QC mode were respectively 88%(88/100),91%(91/100),84%(84/100),and 79%(79/100),all of which were higher than those of manual QC mode,and the differences were significant(x2=9.12,8.53,7.78,10.04,P<0.05).The compliance rate of images about the thalamic level,four-chamber heart,cervix long-axis,femur long-axis,and NT in critically diagnostic cross-section of ultrasound examination of pregnant women in adopting AI-based QC mode were all higher than those of manual QC mode,and the differences were significant(x2=9.01,8.95,8.12,7.89,11.24,P<0.05).Conclusion:The application of the AI quality control platform can effectively improve the ability of primary hospitals to obtain standard prenatal ultrasound sections and enhance image quality,thereby increasing the overall compliance rate of standard prenatal ultrasound sections,including key diagnostic sections.
吴凤妹;高珊;徐婷婷;鲁凤
上海市松江区妇幼保健院影像超声科 上海 201600上海市松江区妇幼保健院影像超声科 上海 201600上海市松江区妇幼保健院影像超声科 上海 201600上海市松江区妇幼保健院产科 上海 201600
医药卫生
人工智能(AI)质量控制产前超声标准切面基层医疗
Artificial intelligence(AI)Quality controlPrenatal ultrasoundStandardized cross-sectionPrimary medicine
《中国医学装备》 2026 (2)
94-97,4
上海市松江区科技攻关项目(2024SJKJGG14) Science and Technology Research Project of Songjiang District,Shanghai(2024SJKJGG14)
评论