人工智能技术评估早孕期胎盘体积应用价值分析  

Analysis of the Application Value of Artificial Intelligence Technology to Evaluate Placental Volume in Early Pregnancy

作　　者：张芸 丁霞 李刚 龚俊丹 丁文萍[2] 徐成[3] Zhang Yun;Ding Xia;Li Gang;Gong Jundan;Ding Wenpin;Xu Cheng(Maternal and Child Health Hospital of Qiaokou,Wuhan,Hubei,430030,China)

机构地区：[1]武汉市硚口区妇幼保健院,湖北武汉430030 [2]武汉市妇女儿童医疗保健中心,湖北武汉430000 [3]武汉工程大学,湖北武汉430205

出　　处：《中国初级卫生保健》2024年第3期56-59,共4页Chinese Primary Health Care

基　　金：武汉市医学科研项目(WX21Q66)。

摘　　要：目的:将三维超声影像学方法测量的胎盘体积与人工智能方法测算的胎盘体积进行对比,探讨人工智能在测算胎盘体积中的应用价值。方法:以妊娠11~13+6周孕妇胎盘图像为研究对象,依据胎盘形状不同将孕妇分成圆形胎盘组和非圆形胎盘组,用三维容积探头采集图像后,由超声医师手动勾画胎盘平面并测量胎盘体积。同时,人工智能技术对三维超声图像进行学习后,利用Unet作为主干网络,对胎盘图像进行分割并计算出胎盘体积。比较两种方法在测算胎盘体积准确性及耗时方面的差异。结果:两种方法测量的圆形胎盘组及非圆形胎盘组的胎盘体积比较,差异均无统计学意义(P>0.05),一致性强;而测量时间方面,人工智能方法测算圆形胎盘组及非圆形胎盘组胎盘体积的用时分别为(22.78±6.90) s和(31.17±6.10) s,均小于三维超声勾画用时,两组比较,差异有统计学意义(t值分别为20.79和19.54,P值均<0.01)。结论:人工智能方法测算胎盘体积与三维超声勾画测算的胎盘体积无明显差异,而人工智能工作效率显著优于医生的人工勾画,可作为孕早期测算胎盘体积的高效工具,为临床早期识别及筛查胎儿生长受限提供参考依据。OBJECTIVE To investigate the application of artificial intelligence(AI)in measuring placental volume by comparing placental volume measured by three-dimensional ultrasonography with placental volume automatically identified and calculated by artificial intelligence(AI).METHODS The placental images of pregnant women at 11~13+6 weeks of gestation were selected as the research object,and were divided into round placenta group and non-round placenta group according to different placenta shapes.After collecting the images with three-dimensional volume probe,the placental plane was manually delineated and the placental volume was measured by sonographers.At the same time,after learning the three-dimensional ultrasound images by artificial intelligence technology,Unet was used as the backbone network to segment the placental images to calculate the placental volume.The two methods were compared for differences in the ability to measure placental volume and time consumption.RESULTS There was no significant difference in the volume of placenta measured by the two methods between the round placenta group and the non-round placenta group(P>0.05).In terms of measurement time,the time taken by the artificial intelligence method to measure the placental volume of the round placenta group and the non-round placenta group was(22.78±6.90)s and(31.17±6.10)s,respectively,which were smaller than the time taken by the three-dimensional ultrasound imaging method,and the difference between the two groups was statistically significant(t=20.79,19.54,P<0.01).CONCLUSION There is no significant difference between the measurement of placental volume by artificial intelligence method and that by three-dimensional ultrasound,and the work efficiency of artificial intelligence is significantly better than that of doctors'manual sketch,which can be used as an efficient tool to measure placental volume in early pregnancy and provide a reference for early clinical identification and screening of fetal growth restriction.

关 键 词：胎儿生长受限 早期妊娠 人工智能 胎盘体积 超声影像 

分 类 号：R714[医药卫生—妇产科学]