摘要
目的观察产科超声图像智能质量控制系统(IUQCS)的效能。方法以IUQCS评价64家深圳医院573名医师采集的164010幅6774胎孕中晚期胎儿声像图的质量,由2名专家(E1/E2)对其中57444幅单幅图像进行质控;对比其对切面分类及标准程度的质控效能和一致性及耗时差异。结果IUQCS质控总体标准率为81.16%,基本标准率10.10%,非标准率8.74%。IUQCS与E1/E2质控切面分类总体符合率分别为97.61%(56071/57444)及97.65%(56092/57444),一致性强(Kappa均>0.933);标准程度评价总体符合率为89.72%(50307/56071)及89.67%(50300/56092),一致性较强(Kappa均=0.658)。IUQCS每100幅质控耗时明显小于人工质控[33(29,37)s vs.705(680,730)s vs.720(696,751)s,Z均=-20.776,P均<0.001]。结论利用产科IUQCS可较为准确、高效地实现图像质控。
Objective To observe the effectiveness of obstetric intelligent ultrasonic quality control system(IUQCS).Methods IUQCS was used to evaluate the quality of 164010 prenatal ultrasonic images of 6774 fetuses in the second and the third trimester by 573 doctors from 64 hospitals in Shenzhen,while 57444 single-displayed images were extracted for quality evaluation by 2 experts(E1/E2).The efficiency of quality control and consistency of IUQCS and manual quality control on plane classification and standard degree were evaluated,and the quality control time were compared.Results The overall standard rate,substandard rate and nonstandard rate of IUQCS was 81.16%,10.10%and 8.74%,respectively.The overall coincidence rate of plane classification of IUQCS and E1/E2 was 97.61%(56071/57444)and 97.65%(56092/57444),respectively,all with strong consistency(all Kappa>0.933),and the overall coincidence rate of standard degree evaluation was 89.72%(50307/56071)and 89.67%(50300/56092),with relatively strong consistency(both Kappa=0.658).The quality control time of every 100 pieces of IUQCS was significantly shorter than that of E1/E2(33[29,37]s vs.705[680,730]s vs.720[696,751]s,both Z=-20.776,both P<0.001).Conclusion Using obstetric IUQCS could accurately and efficiently realize image quality control.
作者
谭莹
文华轩
彭桂艳
罗丹丹
温昕
江瑶
黄文兰
李胜利
TAN Ying;WEN Huaxuan;PENG Guiyan;LUO Dandan;WEN Xin;JIANG Yao;HUANG Wenlan;LI Shengli(Department of Ultrasound,Affiliated Shenzhen Maternity&Child Healthcare Hospital,Southern Medical University,Shenzhen 518028,China)
出处
《中国医学影像技术》
CSCD
北大核心
2022年第9期1361-1366,共6页
Chinese Journal of Medical Imaging Technology
基金
国家自然科学基金(81771598)。
关键词
胎儿
超声检查
人工智能
质量控制
fetus
ultrasonography
artificial intelligence
quality control
