摘要
目的为解决目前有创血糖监测仪操作繁琐、过程痛苦、伤口有感染风险等问题,设计一种基于近红外技术的妊娠期无创血糖仪。方法采用支持向量机、支持向量机回归与卷积神经网络融合算法,通过双波长光电容积脉搏波信号特征(如波形幅值、斜率、谐波分量)提取血糖关联参数,构建血糖预测模型。临床试验中,同步采集50名孕妇的无创数据与传统有创数据(全血与指尖血),通过t检验分析评估其性能,根据人工智能算法模型计算得到被测者的血糖值。结果通过临床验证发现该无创血糖仪检测31~42岁孕妇的无创-有创血糖值组间差异P值为0.222、0.655以及体重在70~100 kg孕妇组中无创-有创血糖值组间差异P值为0.113、0.607,表明无创与传统血糖测量方法差异无统计学意义,证明了无创血糖仪测量血糖的可行性与准确性。结论妊娠期近红外无创血糖仪在部分孕妇群体中血糖检测结果准确,但适用范围较小,仍存在提升空间。
Objective To solve the problems such as cumbersome operation,painful process and risk of wound infection in current invasive blood glucose monitors,design a non-invasive blood glucose monitor for pregnancy based on near-infrared technology.Methods The fusion algorithm of support vector machine,support vector machine regression and convolutional neural network was adopted.The blood glucose correlation parameters were extracted through the characteristics of dual-wavelength photoelectric volumetric pulse wave signals(such as waveform amplitude,slope,harmonic components)to construct a blood glucose prediction model.In the clinical trial,non-invasive data and traditional invasive data(whole blood and fingertip blood)of 50 pregnant women were collected simultaneously.The performance was evaluated through t-test analysis,and the blood glucose values of the subjects were calculated based on the artificial intelligence algorithm model.Results Through clinical verification,it was found that the P values of the differences between the non-invasive and invasive blood glucose levels of pregnant women aged 31-42 years detected by this non-invasive blood glucose monitor were 0.222 and 0.655,and the P values of the differences between the non-invasive and invasive blood glucose levels of pregnant women with a weight of 70-100 kg were 0.113 and 0.607,indicating that there was no statistically significant difference between the non-invasive and traditional blood glucose measurement methods,proving the feasibility and accuracy of non-invasive blood glucose monitor in measuring blood glucose.Conclusion The near-infrared non-invasive blood glucose monitor during pregnancy has accurate blood glucose detection results in some pregnant women.However,its scope of application is relatively limited and there is still room for improvement.
作者
路鹤晴
程立瑞
张晓峰
李树晗
朱志刚
LU Heqing;CHENG Lirui;ZHANG Xiaofeng;LI Shuhan;ZHU Zhigang(Research Department of Integration of Medicine and Engineering,Obstetrics and Gynecology Hospital of Tongji University,Shanghai 201204,China;School of Health Engineering and Science,University of Shanghai for Science and Technology,Shanghai 200093,China)
出处
《中国医疗设备》
2025年第6期23-28,54,共7页
China Medical Devices
基金
上海市科技创新行动计划项目(21S31902100)
上海申康医院发展中心管理研究项目(2024SKMR-27)
国家卫健委医管所2022医学工程研究项目(2022MEB108)。
关键词
无创
血糖监测
妊娠期糖尿病
光电容积技术
脉搏波信号
临床测试
non-invasive
blood glucose monitors
gestational diabetes mellitus
photoelectric volumetric technology
pulse wave signal
clinical test
